Exploiting the aggregation propensity of beta-lactamases to design inhibitors that induce enzyme misfolding

The study addresses the escalating problem of bacterial resistance to beta-lactam antibiotics driven by beta-lactamases, particularly class A ESBLs (e.g., TEM and SHV) in Gram-negative pathogens. Traditional active-site beta-lactamase inhibitors (e.g., tazobactam, clavulanate) can impose selective pressure that accelerates active-site evolution and may paradoxically act as pharmacological chaperones that stabilize and enhance folding of the enzymes. The authors hypothesize that exploiting conserved aggregation-prone regions (APRs) outside the active site to induce selective aggregation and misfolding of beta-lactamases can inactivate these resistance enzymes and restore susceptibility to beta-lactam antibiotics. The purpose is to design and test synthetic peptides (Pept-Ins) targeting APRs in TEM and SHV beta-lactamases, evaluate their specificity and efficacy in vitro and in vivo, and assess generalizability to structurally distinct beta-lactamases such as NDM-1.

Background highlights include: (1) Early emergence and global spread of beta-lactamase-mediated resistance, with ESBLs extending resistance to later-generation cephalosporins and monobactams. (2) Beta-lactamases are grouped into Ambler classes A, B, C, D; class A employs serine-mediated hydrolysis, class B uses zinc. (3) Active-site inhibitor strategy and its drawbacks: mechanism-based inhibitors (tazobactam, clavulanate, sulbactam) foster selective pressure and can act as pharmacological chaperones. Clinical examples of chaperones include DGJ for alpha-galactosidase in Fabry disease and tafamidis for transthyretin amyloidosis, where active-site binding can stabilize folding. (4) APRs are common in globular proteins, often unique in the proteome, and drive amyloid-like aggregation; seeding is sequence-specific and can be induced by synthetic APR peptides. Prior work demonstrated feasibility of targeted aggregation in bacteria, plants, and mammalian cells, enabling selective functional knock-down via sequence-specific aggregation.

• Bioinformatic prediction of APRs: TANGO used to identify APRs in TEM-1; 7 candidate APRs identified (one in signal peptide, six in the globular domain). Comparison to SCOPe-derived dataset (9017 single-domain proteins) to contextualize APR counts.
• Cloning and expression: TEM-1 and SHV-11 fused to GFP (linker) and expressed in E. coli under arabinose promoter; imaging by structured illumination microscopy (SIM) to observe inclusion bodies. Recombinant TEM-1 and SHV-11 (His-tagged) expressed in E. coli BL21 and purified (pET system); SEC-MALS used to confirm monomeric state.
• Biophysics of stability and aggregation: Thermal denaturation monitored via intrinsic fluorescence (BaryCentric Mean) to determine Tm, and Right-Angle Light Scattering to monitor aggregation and Tagg. Tested effects of active-site inhibitors (tazobactam, clavulanate, vaborbactam, avibactam) across concentrations; correlation analysis of inhibitor concentration vs scattering at 60°C.
• Evolutionary and structural analysis: TEM variants from BLDB were compiled; mutations mapped and classified (extended-spectrum, inhibitor resistance, stability, other). FoldX (using PDB 1XPB) predicted stability effects; mutations mapped to structure and compared to APR locations (TANGO). Key clinical variants (TEM-10, TEM-30, TEM-52, TEM-15, TEM-155) recombinantly produced; aggregation profiles assessed ± tazobactam.
• Pept-In design and synthesis: Tandem APR repeats flanked by arginines (solubility/uptake) separated by a proline linker. APR lengths normalized to 7 residues, avoiding aggregation breakers. Solid-phase peptide synthesis (Genscript and in-house), HPLC purification ≥90% purity; N-terminal acetylation and C-terminal amidation. Design of peptides TEM1.0–TEM7.0 targeting TEM APRs (Table 1), highlighting TEM3.0 (APR3: RLTAFLHNRRPRLTAFLHNRR). Cross-reactive designs TEM3.1–3.3 combining TEM and SHV APR variants.
• Antibacterial assays: MIC determination by broth microdilution (EUCAST). Checkerboard synergy assays with beta-lactams (penicillin G, ampicillin) to compute FICI (synergy if ≤0.5). Tested multiple clinical E. coli isolates carrying TEM/SHV and other beta-lactamases.
• Aggregation assays: DLS for hydrodynamic radius; Thioflavin T and pFTAA fluorescence kinetics; TEM electron microscopy. Aggregation triggers included LPS, polyphosphate (PolyP), and crowding (PEG). Peptide-induced aggregation of purified TEM measured by pFTAA; enzymatic activity loss assessed by nitrocefin assay.
• Cellular aggregation and specificity: SIM imaging of FITC-labeled TEM3.2 in E. coli UZ_TEM104 and controls; inclusion bodies stained with pFTAA/HS169. Western blot of inclusion body fractions probed with anti-TEM or anti-SHV to confirm target aggregation with various peptides; mass spectrometry proteomics confirmed bands. FACS measured PI (viability) and pFTAA (aggregation) after treatments.
• Safety and mammalian off-target assessment: In silico proteome scans (human/mouse) for exact APR matches (none found). Hemolysis assay on human erythrocytes; CellTiter Blue cytotoxicity on multiple human cell lines and primary cells; microscopy of HeLa–E. coli co-cultures treated with FITC-TEM3.2 to assess selectivity of uptake.
• In vivo mouse UTI model: Female C57BL/6JAX, urethral inoculation with UPEC E. coli (blaTEM-1). Treatment at 60 and 120 min post-infection with oral ampicillin (30 mg/kg) plus IV/IP/SC FITC-TEM3.2 (10 mg/kg) or tazobactam (10 mg/kg). CFU quantified in bladder, kidney, ureter at 24 h; FACS of recovered bacteria assessed peptide uptake and aggregation. Tolerability studies included dose escalation (2–20 mg/kg) via multiple routes and hematology after daily IV (5 mg/kg).
• Generalization to NDM-1: APR identification with WALTZ; design of NDM1-1 (RTAQILNWRRPRTAQILNWRR) and NDM1-2 (RLAAALMLRRPRAQILNWIRR) targeting T101–W107 helix. Checkerboard FICI with NDM-5 strains vs controls (TEM-1 strain, off-target peptides BGAL and toxic P33).

• TEM-1 harbors multiple APRs (7 predicted by TANGO), a high count relative to proteins of similar size. Both TEM-1 and SHV-11 form inclusion bodies in cells when overexpressed.
• Biophysical stability/aggregation: TEM-1 is marginally stable (Tm ≈ 43.3°C) vs SHV-11 (Tm ≈ 68.8°C). Aggregation onset temperatures are low and similar (Tagg ≈ 44.0°C for TEM-1; 45.6°C for SHV-11). SEC-MALS confirmed monomeric starting material.
• Active-site inhibitors act as pharmacological chaperones: Tazobactam, clavulanate, vaborbactam, and avibactam reduce aggregation of TEM-1 in a dose-responsive manner; correlations between inhibitor concentration and reduced light scattering are significant (P < 0.0001, < 0.0004, < 0.04, and < 0.006, respectively).
• Evolutionary constraint of APRs: Clinical mutations that expand spectrum/resistance cluster near the active site and are often destabilizing (e.g., G238S, R164H). Compensatory stabilizing mutations exist (e.g., M182T). Mutations largely avoid APRs, which remain invariant across variants, supporting APR-targeting as robust to mutational escape.
• Pept-Ins restore beta-lactam susceptibility: In E. coli UZ_TEM104 (TEM+), peptides were non-toxic up to 100 µg/mL. TEM3.0 (APR3) at 50 µg/mL reduced penicillin G MIC 8-fold (from 1600 to 200 µg/mL). Tazobactam at 50 µg/mL reduced MIC 4-fold (to 400 µg/mL). Combined TEM3.0+tazobactam (25+25 µg/mL) further reduced MIC to 50 µg/mL.
• Sequence specificity and cross-reactivity: TEM3.0 synergized with penicillin only in TEM strains; SHV-matched TEM3.1 synergized only in SHV strains. Mixed APR peptides TEM3.2 and TEM3.3 synergized in both TEM and SHV strains (FICI ≤ 0.5), demonstrating sequence-dependent targeting. No synergy with an unrelated antibiotic (kanamycin) or with control peptides (BGAL or APR-proline mutants).
• Broad activity across clinical isolates: In 16 additional E. coli clinical isolates, 30 µg/mL TEM3.2 sensitized 87% (14/16) to ampicillin (MIC decreased from >64 µg/mL to ≤1 µg/mL, often ≤0.06 µg/mL). Presence of CTX-M or OXA did not abrogate effect, whereas presence of NDM metallo-β-lactamases prevented sensitization.
• Mode of action confirmed: TEM3.2 forms amyloid-like aggregates upon triggers (PolyP), with ThT/pFTAA positivity and fibrils by EM. Peptide induces aggregation of purified TEM-1 (pFTAA kinetics) and loss of nitrocefin-hydrolysis activity. In cells, FITC-TEM3.2 localizes to inclusion bodies co-staining with pFTAA; inclusion body fractions show enriched TEM or SHV after matched-peptide treatments by Western blot; no aggregation of unrelated fluorescent proteins.
• FACS evidence: TEM3.2 alone shifts bacterial population to pFTAA+ (aggregation) without PI uptake; combined with penicillin G yields dual pFTAA+ and PI+ (cell death), indicating aggregation-mediated re-sensitization.
• Safety and selectivity: No exact APR matches found in human/mouse proteomes. TEM3.2 was non-hemolytic and non-cytotoxic in multiple human cell lines and primary cells; FITC-TEM3.2 preferentially accumulates in bacteria in co-culture imaging.
• In vivo efficacy: In a mouse UTI model (ampicillin plus TEM3.2 via IV/IP/SC), bacterial loads reduced by up to ~2–2.8 log10 CFU in kidney and ureter, slightly outperforming tazobactam (best ~1.8 log10 reduction). FACS of bacteria from treated mice confirmed peptide uptake and aggregation.
• Generalization to NDM-1: Two peptides (NDM1-1 and NDM1-2) targeting an APR in NDM-1 synergized with penicillin in NDM-5 strains (FICI ≤ 0.5) but not in TEM-1 strains; control peptides showed no synergy, supporting target specificity.

The findings demonstrate that targeting conserved aggregation-prone regions outside the active site can selectively induce misfolding and aggregation of beta-lactamases, inactivating these resistance determinants and restoring susceptibility to beta-lactam antibiotics. This strategy circumvents the drawbacks of active-site inhibitors, which can stabilize enzyme folding and accelerate active-site evolution. The invariance of APRs across clinical TEM variants and the sequence specificity observed in synergy and aggregation assays support the robustness and selectivity of the approach. The restoration of antibiotic sensitivity across diverse clinical isolates and proof-of-concept efficacy in a mouse UTI model underscore translational potential. The approach is adaptable to distinct beta-lactamases, as shown for NDM-1, suggesting broader applicability. While the aggregation induced by Pept-Ins is not intrinsically bactericidal, it creates a vulnerability that antibiotics can exploit, aligning with an adjuvant strategy to rejuvenate existing beta-lactams.

This work introduces Pept-Ins—synthetic peptides that exploit conserved APRs to induce selective aggregation and inactivation of beta-lactamases—as a strategy to overcome beta-lactam resistance. The authors show that conventional active-site inhibitors act as pharmacological chaperones that stabilize beta-lactamases, whereas APR-targeting peptides counteract resistance by misfolding the enzymes. Pept-Ins targeting TEM/SHV restore susceptibility to beta-lactams across clinical isolates and are effective in a mouse UTI model. The concept generalizes to structurally unrelated NDM-1, indicating broad potential. Future work should optimize peptide pharmacokinetics and delivery, expand coverage to additional beta-lactamases, assess resistance liabilities, and progress toward clinical evaluation of peptide–antibiotic combinations.

• Peptides are not yet optimized for drug-like properties (stability, PK/PD, delivery), though initial tolerability was favorable.
• Target specificity implies different peptides are needed per beta-lactamase type; mixed-resistance strains with unrelated enzymes (e.g., NDM) were not sensitized by TEM/SHV-targeting peptides.
• Efficacy was demonstrated primarily with penicillin/ampicillin; performance with other beta-lactams and in different infection models remains to be established.
• In vivo results derive from a single species and infection model (murine UTI); broader validation across pathogens, sites, and dosing regimens is needed.
• While no exact APR matches were found in mammalian proteomes, comprehensive off-target assessments in complex in vivo settings and long-term exposure studies are pending.