Inhibition of Clostridium difficile TcdA and TcdB toxins with transition state analogues

Clostridium difficile causes life-threatening diarrhea and is the leading cause of healthcare-associated bacterial infections in the United States. TcdA and TcdB bacterial toxins are primary determinants of disease pathogenesis and are attractive therapeutic targets. TcdA and TcdB contain domains that use UDP-glucose to glucosylate and inactivate host Rho GTPases, resulting in cytoskeletal changes causing cell rounding and loss of intestinal integrity. Transition state analysis revealed glucocationic character for the TcdA and TcdB transition states. We identified transition state analogue inhibitors and characterized them by kinetic, thermodynamic and structural analysis. Iminosugars, isofagomine and noeuromycin mimic the transition state and inhibit both TcdA and TcdB by forming ternary complexes with Tcd and UDP, a product of the TcdA- and TcdB-catalyzed reactions. Both iminosugars prevent TcdA- and TcdB-induced cytotoxicity in cultured mammalian cells by preventing glucosylation of Rho GTPases. Iminosugar transition state analogues of the Tcd toxins show potential as therapeutics for C. difficile pathology.


Synthesis of isofagomine
Isofagomine was synthesized as reported (Rybczynski, P.; Tretyakov, A.; Fuerst, D.; Sheth, K. New method for preparing isofagomine and it's derivatives, US Patent 2010/0160638) with some modifications. Nitrile 1 (prepared from D-arabinose) was reduced with BH3.SMe2 to the amine which was isolated as it's N-Boc derivative 2. This was then treated with aq HCl followed by hydrogenolysis affording isofagomine 3.
A solution of 1 (0.507 g, 2.03 mmol) in dry THF (10 mL) was stirred at RT under argon and borane dimethyl sulphide complex (1.03 mL, 0.822 g, 10.2 mmol) was added, then the solution was heated under reflux for 16 h. Methanol (8 mL) was carefully added and the solution was heated under reflux for 3 h and then concentrated to dryness. Methanol was added to the residue and the mixture concentrated to dryness three times. Then methanol (20 mL) was added followed by 10% aq sodium carbonate solution (5 mL). Ditert-butyl dicarbonate 1.35 g, 6.10 mmol) was added and the mixture was stirred for 1 h and then concentrated to dryness. Chloroform (30 mL) was added and the mixture was filtered through celite and the solids washed with chloroform. Evaporation of the filtrates gave a solid. This was triturated with hexanes, filtered and dried to give 2 (0.62 g, 1.75 mmol, 86%) with NMR spectra consistent with that reported 1 .
To a solution of 2 (9.80 g, 27.7 mmol) in THF (100 mL) was added 5M aq HCl (75 mL) and the solution was stirred at RT for 2 h and then concentrated to dryness. A solution of the residue in methanol was treated with Amberlyst A26 (OH-) resin until the solution pH was > 7, and then filtered and concentrated to dryness. The residue in THF (100 mL) and methanol (100 mL) was stirred under hydrogen in the presence of 10% Pd/C for 16 h. After removal of the solids and solvent and chromatography of the residue DCM/MeOH/aq NH3 5:4:1) afforded isofagomine 3 (2.80 g, 19.0 mmol, 68%) which could be converted to the tartrate salt as reported 1 .

Supplementary Figure 1. Measurement of forward commitment (Cf) for UDP-glucose hydrolysis with
TcdB-GTD. TcdB-GTD was mixed with [6′′-14 C]-UDP-glucose and incubated for 5 seconds to form 3.4 µM glucose. At that time the reaction was chased with a large excess of unlabeled UDP-glucose and the synthesis of additional glucose after the chase was monitored to 65 sec. After extrapolation to the origin [glucose] = 3.29 ± 0.04 µM; using eq 8, 9 and 10 the Cf determined was 0.064 ± 0.01. Data was plotted using GraphPad Prism 8. Data represents the mean of each time point and error bars represent ± SEM. Source data provided as a Source Data file.

Supplementary Figure 2. Kinetic analysis of Inhibitor binding. a
TcdB-GTD glucosyltransferase activity was measured in the presence of UDP and varying concentrations of UDP-glucose. Lineweaver Burk curves are shown for 6-3 H UDP-glucose. Squares represent 40 µM UDP, diamonds represent 20 µM UDP, triangles represent 10 µM UDP and circles represent 0 µM UDP. Data points are the mean of experimental data and lines represent the global fit to equation 13 for competitive inhibition as described in methods. b TcdA-GTD glucosyltransferase activity was measured in the presence of isofagomine and varying concentrations of UDP-glucose. Squares represent 4 µM isofagomine, diamonds represent 2 µM isofagomine, inverted triangles represent 1 µM isofagomine and squares represent 0 µM isofagomine. Data points are the mean of experimental data and lines represent the global fit to equation 12 for uncompetitive inhibition as described in methods. Source data provided as a Source Data file. Vero cells were pre-treated with varying concentrations of isofagomine or noeuromycin for 20 minutes before addition of 1 nM TcdA or 1 pM TcdB. Cells were incubated with TcdB or TcdA for 1.5 and 2 hr respectively before sample imaging. Rounded cells were quantified and expressed as a percentage of total cells. The IC50 for each iminosugar was calculated by fitting the data to a non-linear regression model of Log[inhibitor] vs response. A representative plot from 1 experiment is shown on the left, with representative image of 12.5 µM iminosugar shown on the right (Scale bar represents 10µm). Data points represent the mean % cell rounding from 4 images of each well ± SEM. IC50 values at the top of each graph represent the mean IC50 ± SEM of 3 biological replicates (n=3). a TcdB + isofagomine. b TcdB + noeuromycin. c TcdA + isofagomine. d TcdA + noeuromycin. Source data provided as a Source Data file.

Supplementary Figure 8. IMR90 cells treated with TcdA and TcdB and isofagomine or noeuromycin.
Representative images of human IMR90 cells from 3 experiments (n=3) . Scale bar represents 10µm. Cells were pre-treated with either 100 µM isofagomine (Isof) or noeuromycin (Noe) for 30 min before treatment with buffer, 1 nM TcdA or 0.1 nM TcdB and images were acquired after 6 hr. a TcdA treatment. b Treatment with isofagomine, then TcdA. c Treatment with noeuromycin, then TcdA. d TcdB treatment. e Treatment with isofagomine, then TcdB. f Treatment with noeuromycin, then TcdB. g Untreated control cells. Images show cell rounding with addition of TcdA or B versus no toxin control, and addition of inhibitors prevents cell rounding. and < I(hkl) > is its mean intensity. F. c I is the integrated intensity and σ(I) is its estimated standard deviation. G. d Rwork = (Σhkl|Fo-Fc|)/ΣhklFo where Fo and Fc are the observed and calculated structure factors. H. e Rfree is calculated as for Rwork but from a randomly selected subset of the data (5%), which were excluded from the refinement calculation.