Design of 8-mer peptides that block *Clostridioides difficile* toxin A in intestinal cells

Clostridioides difficile is a Gram-positive, spore-forming pathogen and a leading cause of antibiotic-associated diarrhea and colitis. Its pathogenicity is primarily driven by two large glucosylating toxins, TcdA and TcdB, which bind epithelial cell surfaces, translocate into the cytosol, and glucosylate Rho-family GTPases, disrupting cellular functions. Current treatments (antibiotics, monoclonal antibody Bezlotoxumab against TcdB, fecal microbiota transplant) have limitations, including recurrence, cost, delivery burdens, and safety risks. Short peptides are attractive as specific, cost-effective inhibitors. Building on prior work that identified a 10-mer peptide (NPA) that binds TcdA glucosyltransferase domain (GTD) and neutralizes TcdA in small intestinal (but not colon) cells, this study asks whether engineered, shorter peptides can bind the TcdA GTD catalytic site with high affinity and neutralize TcdA toxicity in human colon epithelial cells. The purpose is to design and validate 8-mer peptides that block TcdA activity in colon epithelium and to demonstrate a design-simulate-screen-test pipeline.

Standard care with metronidazole or vancomycin leads to ~20% recurrence and disrupts the microbiome, facilitating C. difficile colonization. Bezlotoxumab (anti-TcdB) reduces recurrence but is costly and requires IV administration. Fecal microbiota transplant remains investigational and carries risks (e.g., pathogen transmission, antibiotic resistance). Prior peptide discovery via phage display identified RP (EGWHAHTGGG) inhibiting TcdA GTD. Computational work previously identified NPA (DYWFQRHGHR) binding TcdA GTD and neutralizing TcdA in jejunum but not colon cells, with a hypothesis that small intestinal proteases may cleave to more active fragments, whereas colon epithelium lacks such proteases. The PepBD algorithm has been used to design peptide binders to diverse targets (tRNA, proteins including IgG domains and SARS-CoV-2 RBD), and a bead-based microfluidic screening platform has enabled rapid assessment of peptide binding specificity in other systems (Cas9, VCAM-1, IgG Fab).

- Computational determination of optimal peptide length: Atomistic MD simulations of RP and NPA bound to TcdA GTD were used to decompose residue-wise interaction energies and simulate 6-, 7-, and 8-mer fragments. 8-mers exhibited favorable binding free energies (e.g., 8-mer NPA DYWFQRHG ΔGbinding = −6.35 kcal/mol), guiding an 8-mer design focus. - Peptide design with PepBD: Using the 8-mer NPA:TcdA GTD complex as the reference, PepBD performed Monte Carlo sequence and conformation moves under three composition constraints (Cases 1–3) to optimize a score (Iscore) combining binding energy and bound-state peptide stability. Multiple random seeds were used per case to diversify search paths. - MD simulations and binding free energy estimation: Selected low-Iscore peptides were simulated (three 100 ns runs each; AMBER18, ff14SB, explicit TIP3P water) and analyzed by MM/GBSA with variable dielectric to compute ΔGbinding from equilibrated trajectory segments. Select complexes were simulated to 500 ns to check stability. - Bead-based dual-fluorescence screening: Peptides (SA1–SA7, 8-mer NPA, 8-mer RP) were synthesized on ChemMatrix Aminomethyl resin via Fmoc chemistry with a GSG linker and displayed on beads. A microfluidic imaging platform quantified binding of Alexa Fluor 594-labeled TcdA (red halo intensity) and exclusion of UDP-Glucose-Fluorescein (green) at the TcdA halo region as a proxy for GTD active site occupancy/selectivity. Approximately 30 beads per peptide were imaged; statistics used two-tailed t-tests versus controls. - Functional assay in primary human colon epithelial monolayers: Primary human colonic epithelial stem cells (descending colon) were expanded and differentiated into absorptive lineage on Transwells. Peptides were pre-incubated apically (1 mM, 2 h) before addition of TcdA (30 pM). Transepithelial electrical resistance (TEER) was monitored over time to assess barrier disruption and peptide-mediated protection. - Surface plasmon resonance (SPR): SA1 was modified (C-terminal Lys-amide) and grafted to a mixed ethylene glycol thiol SAM on gold via EDC/NHS chemistry. Equilibrium and kinetic binding of TcdA to immobilized SA1 were measured on a KSV SPR 200 system in Tris/NaCl/MgCl2 running buffer. Equilibrium responses were fit to a Langmuir isotherm to obtain KD and Qmax; dynamic injections provided adsorption (ka) and desorption (kd) rate constants.

- 8-mer peptides are optimal among tested fragments for engaging the TcdA GTD catalytic site: 8-mer NPA fragment DYWFQRHG showed ΔGbinding = −6.35 kcal/mol; 8-mer RP fragment EGWHAHTG showed ΔGbinding = −5.79 kcal/mol. - PepBD-designed peptides: Seven candidates (SA1–SA7) were identified. Computed binding free energies (kcal/mol; mean ± SE) included: SA1 (EFWWRRHN) −15.94 ± 0.40; SA2 −13.19 ± 0.39; SA3 −11.76 ± 0.42; SA4 −11.01 ± 0.49; SA5 −6.16 ± 0.45; SA6 −12.54 ± 0.49; SA7 −9.56 ± 0.47, outperforming the 8-mer NPA reference (−6.35 ± 0.51). Key SA1 interacting residues with GTD were Trp3, Trp4, Arg5, Arg6, His7, Asn8. - Bead-based screening: All peptides bound TcdA; peptides from Cases 1–2 (SA1, SA2, SA3, SA4, SA5) outperformed Case 3 (SA6, SA7). SA1 had the highest mean red halo fluorescence (strongest TcdA binding). Exclusion of UDP-Glucose-Fluorescein at the TcdA binding region was significant for most peptides (p < 0.005); SA1 showed a significant reduction in green fluorescence at the halo versus control (p = 0.0011), indicating selective GTD active-site engagement. - Functional neutralization in human colon epithelium: In differentiated human colon monolayers exposed to 30 pM TcdA, SA1 pre-treatment (1 mM, 2 h) preserved barrier function with ~79% protection at 12 h post-exposure (TEER), with one-way ANOVA p < 0.005 and multiple comparisons p < 0.05 from 12–20 h. - SPR binding parameters: SA1 bound TcdA with KD = 56.1 ± 29.8 nM and Qmax = 12.0 ± 2.2 nmol/m². The average adsorption rate constant ka = 7.0 ± 2.5 × 10^−5 nmol^−1 s^−1, consistent with rapid recognition for a peptide inhibitor. These affinities and kinetics support competitive inhibition against UDP-Glucose (TcdA KM ~ 4.5 μM; cellular UDP-Glucose ~ 92 μM).

The study demonstrates a generalized computational–experimental pipeline to design short peptides that bind the TcdA GTD catalytic site and neutralize toxin activity in primary human colon epithelium. Simulations guided selection of 8-mers, and PepBD efficiently identified sequences with diverse physicochemical properties that improved binding energetics over a prior reference peptide. SA1 showed strong computed binding, selective GTD active-site engagement in bead assays, potent functional protection of epithelial barrier integrity in a clinically relevant colon model, and mid-nanomolar affinity to TcdA by SPR with rapid association. Residue-level analysis highlights contributions from aromatic (Trp) and basic (Arg, His) residues, informing future designs. Collectively, these findings support the hypothesis that rationally designed 8-mer peptides can outcompete the UDP-glucose substrate at the catalytic site and block TcdA-mediated cytotoxicity in colonic epithelial cells.

This work identifies and validates an 8-mer peptide, SA1 (EFWWRRHN), that binds TcdA with mid-nanomolar affinity and protects primary human colon epithelial monolayers from TcdA-induced barrier loss. The integrated PepBD design, MD simulation, bead-based screening, primary cell functional assay, and SPR characterization pipeline provides a robust framework for developing peptide inhibitors against toxin catalytic domains. These results suggest SA1 is a promising lead for therapeutic development against C. difficile infection. Future directions include evaluating in vivo efficacy and safety, optimizing peptide stability and delivery to the colon, expanding designs to target TcdB or combined toxins, and further structure–activity studies guided by residue-level interactions.

- Experimental validation focused on in vitro systems: bead-based binding, primary human colon epithelial monolayers (single-donor source noted), and SPR with immobilized peptide; no in vivo efficacy or pharmacokinetic data were presented. - The SPR dataset largely comprised single measurements per condition (with limited replicates at select concentrations), which may affect precision of kinetic parameter estimates. - The study targeted TcdA; effects on TcdB were not assessed here. - Bead-based UDP-Glucose-Fluorescein exclusion serves as a proxy for GTD active-site occupancy and may not fully capture competitive dynamics in cellular contexts.