Development of a natural product optimization strategy for inhibitors against MraY, a promising antibacterial target

Antimicrobial resistance (AMR) is a global crisis demanding new antibacterial agents effective against resistant pathogens such as methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant enterococci (VRE). MraY, a universally conserved and essential bacterial membrane enzyme that forms lipid I during peptidoglycan biosynthesis, is a promising antibacterial target. Several nucleoside natural products (e.g., tunicamycins, muraymycins, mureidomycins, capuramycin) inhibit MraY and share a uridine moiety that binds the uridine pocket, with other moieties engaging various hotspots on MraY. However, rational design remains difficult due to MraY’s conformational dynamics and the need to balance target inhibition with bacterial membrane permeability and cellular accumulation. The complex, polar structures of these antibiotics and the multi-step syntheses required for analogues hinder rapid optimization. The authors aim to develop a platform that simplifies and accelerates the structural optimization of MraY inhibitory natural products by enabling comprehensive, rapid, in situ synthesis and evaluation of analogue libraries to identify potent, drug-like antibacterial leads and to generalize this strategy to other complex natural product scaffolds.

Prior work established MraY as essential and structurally characterized inhibitor binding modes, with the uridine moiety anchoring in the uridine pocket and accessory motifs engaging hotspots HS1–HS6. Natural product classes including tunicamycin, muraymycin, mureidomycin, and capuramycin exhibit varying spectra and SAR features. Previous in situ screening used amide coupling or CuAAC to assemble libraries but left cytotoxic reagents/by-products, limiting cell-based assays. Hydrazone formation (aldehyde/ketone with hydrazine, producing only water) is a bioorthogonal ligation suited to in situ cell assays. Accessory motifs strongly influence both MraY inhibition and bacterial accumulation (e.g., long lipophilic chains aid muraymycin uptake; aromatic substituents effective for capuramycin; acyl chain structure affects off-target GPT inhibition and cytotoxicity for tunicamycin/caprazamycin). Structural studies of MraY-inhibitor complexes and SAR work on nucleoside antibiotics informed the selection of core aldehydes and diverse hydrazine fragments.

Design and library construction: Natural products were split into a core fragment (uridine-containing, expected to drive target binding) and an accessory fragment (modulating affinity, selectivity, and disposition). Seven aldehyde core fragments across four classes (tunicamycin, muraymycin, 3′-hydroxymureidomycin, capuramycin) were prepared based on prior syntheses, with conjugated aldehydes to stabilize hydrazones. Ninety-eight hydrazine accessory fragments were synthesized/prepared to maximize chemotype diversity: benzoyl-type (BZ), phenylacetyl-type (PA), acyl-type (AC), N-acyl aminoacyl (AA), and lipid amino acid (LA) series spanning various alkyl chain lengths and amino acid side chains (Ala, Phe, Ser, Gln, Glu, Lys, Arg). In situ hydrazone assembly: In 96-well plates, 10 mM DMSO stock solutions of aldehyde cores (15 µL) and hydrazines (16 µL) were mixed (~1:1 stoichiometry) and shaken 30 min at room temperature without additives. DMSO was removed under vacuum overnight, and residues dissolved in 30 µL DMSO to yield 5 mM hydrazone solutions. LC–MS showed most hydrazones formed at ≥80% yield; some were partial or unreactive depending on hydrazine structure. Biochemical assay (MraY inhibition): Using dansylated lipid I formation assay (UDP-MurNAc-dansylpentapeptide 10 µM; undecaprenyl phosphate (C55-P) 50 µM; assay buffer 50 mM Tris-HCl pH 7.6, 50 mM KCl, 25 mM MgCl2, 0.2% Triton X-100, 8% glycerol). Final library assay concentrations were set so that aldehyde cores alone gave <20% inhibition: MRY 500 nM; TUN 5 µM; MRD 5 µM; CAP 50 µM. Fluorescence monitored after 3 h at 25 °C (Ex 355 nm, Em 535 nm). Hydrazine fragments alone showed up to 58% inhibition at 200 µM. Antibacterial screening: MICs (microdilution, CLSI M100) against ESKAPE panel and E. coli ATCC 25922 at three concentrations (0.5, 5, 50 µM, n=1 for library screening). Strains: Enterococcus faecium ATCC 35667, Staphylococcus aureus ATCC 29213, Pseudomonas aeruginosa ATCC 27853, Klebsiella pneumoniae ATCC 13883, Enterobacter cloacae ATCC 13047, Acinetobacter baumannii ATCC 19606, Escherichia coli ATCC 25922. Hit resynthesis and validation: Selected hydrazones from the MRY sub-library (negatives: MRYP-BZ2, MRYP-AA63, MRYP-LA89; positives: MRYP-LA80, MRYP-LA92, MRYP-LA98) were resynthesized (MRYP-CHO + acyl hydrazide in 0.1% TFA/MeOH, RT 24 h; purified via C18 SPE) and characterized by 1H NMR and LC–MS. IC50 against MraYSA determined (n=3) and MICs measured (broth microdilution). Stable analogue design and synthesis: To overcome hydrazone hydrolysis, amide (1–4) and anilide (5–8) analogues based on MRYP-LA92 were synthesized with Lys or Arg residues and varied linkers, using routes similar to MRY-type cores. MraY inhibition (IC50) and MICs were determined as above. HepG2 cytotoxicity measured via WST-8 assay (24 h exposure). Antibacterial properties: Time-kill assays for analogue 2 against S. aureus ATCC 29213 and E. coli ATCC 25922 at 0.25×, 1×, 4× MIC; resistance emergence over 30 serial passages at 0.5× MIC compared with vancomycin and rifampicin; TEM on S. aureus treated with 2 vs controls; in vivo efficacy in neutropenic mouse thigh infection model (S. aureus ATCC 29213), single subcutaneous dose (5 or 30 mg/kg) vs comparators (levofloxacin, linezolid). Colony counts quantified after 24 h. Structural studies: Cryo-EM structures of MraYAA in complex with analogues 2 and 3 using NB7 nanobody. Data collected on Titan Krios/K3; processed in cryoSPARC and RELION with C2 symmetry, yielding 2.88 Å (2) and 2.70 Å (3) reconstructions. Models built/refined using reference PDB 5CKR, Phenix, Coot; densities for ligands validated. Coordinates deposited: PDB 9B70 (2) and 9B71 (3).

- Library assembly: A 686-compound hydrazone build-up library (7 aldehyde cores × 98 hydrazines, with some unreactive pairs) was synthesized in situ with mostly ≥80% conversion by LC–MS. - Enzymatic inhibition trends: MRY-based analogues were the most potent overall. LA-type hydrazines (lipidated amino acids) enhanced MraY inhibition in TUN and MRY sub-libraries; effective residues differed (TUN: Ala, Phe, Ser; MRY: Lys, Arg). BZ-type contributed in MRY; CAP sub-library saw BZ-type effects; MRD only LA97 enhanced inhibition. Core aldehydes alone were 100–1000× less potent than parent natural products (except capuramycin cores). - Antibacterial screening: In TUN and MRY sub-libraries, LA-type hydrazones showed moderate/strong activity at 0.5–5 µM, including some activity against Gram-negatives for MRY analogues. MRD and CAP sub-libraries exhibited little antibacterial activity. - Validated hits (MRY sub-library): Resynthesized hydrazones showed IC50 (MraYSA) of 2.4–7.8 nM for MRYP-BZ2, -LA80, -LA92, -LA98; weaker for MRYP-AA63 (67 nM) and MRYP-LA89 (120 nM). Despite strong enzyme inhibition, MRYP-BZ2 lacked antibacterial potency, highlighting the importance of lipophilic chains for cell activity. MRYP-LA92 and MRYP-LA98 (basic Lys) had broad-spectrum activity; MRYP-LA98 MICs 1–16 µg/mL across ESKAPE panel. - Stable analogues (1–8): Retained potent MraY inhibition (IC50 1.7–6.0 nM). Strong activity against Gram-positives (MICs: S. aureus 0.5–1 µg/mL; E. faecium 0.25–1 µg/mL) and notable Gram-negative activity (K. pneumoniae 4–8; A. baumannii 2–4; E. coli 4–8 µg/mL). Activity against P. aeruginosa varied widely (8–128 µg/mL). HepG2 cytotoxicity moderate (IC50 12–25 µM). - Bactericidal activity and resistance: Analogue 2 was bactericidal, reducing S. aureus by ~2 log10 cfu/mL at 6 h (4× MIC) and rapidly decreasing E. coli counts (<1 log10 at 3 h, 4× MIC). Resistance emergence for S. aureus under 0.5× MIC exposure rose only 2-fold over 30 days, contrasting with rifampicin and vancomycin controls. - In vivo efficacy: In a neutropenic mouse thigh infection model (S. aureus ATCC 29213), analogue 2 significantly reduced bacterial burden at 30 mg/kg, comparable to standard comparators; multiple analogues (2–6) showed efficacy signals. - Mechanism/structural insights: Cryo-EM structures revealed analogues 2 and 3 bind MraY engaging uridine and HS1 similar to muraymycin but uniquely occupy HS6 (TM4/5 groove) instead of HS2. Basic Lys/Arg interacts with HS6, orienting the aliphatic tail toward a hydrophobic groove (F180, G184, N187, A188, V296, T299, V302, I303), suggesting a distinct binding mode. TEM suggested membrane perturbation upon treatment with 2, indicating possible dual action (membrane interaction plus MraY inhibition). - Generality: The build-up library strategy applied to tubulin-binding natural products (epothilone B, paclitaxel, vinblastine) generated a 588-analogue library (6 cores × 98 hydrazines), yielding analogues with enhanced tubulin polymerization activity and cell cytotoxicity in HCT-116, demonstrating versatility.

The study addresses the challenge of rapidly optimizing complex MraY-inhibitory natural products by implementing an in situ build-up library approach based on hydrazone ligation. This chemoselective, high-yielding reaction allowed direct enzymatic and whole-cell evaluations without purification, enabling efficient SAR elucidation across multiple scaffolds. Results show that long lipophilic chains combined with basic residues (Lys/Arg) are critical for translating potent enzymatic inhibition into antibacterial activity, especially for polar muraymycin-derived cores where cellular accumulation is limiting. Converting labile hydrazones to stable amide/anilide analogues preserved MraY potency and delivered broad-spectrum antibacterial activity, including against multidrug-resistant clinical isolates, with low propensity for resistance development. Cryo-EM revealed a unique binding profile for analogues 2 and 3, engaging HS6 rather than HS2, expanding known MraY interaction space. The observed membrane perturbation suggests an auxiliary mechanism that may enhance bactericidal action, paralleling other agents with dual effects. Importantly, moderate cytotoxicity and in vivo efficacy validate the potential of these leads. The strategy’s successful extension to tubulin-binding natural products underscores its broad applicability to complex, mid-sized molecules for rapid SAR and functional tuning.

This work introduces a practical, generalizable in situ build-up library strategy that streamlines the optimization of complex natural product-derived inhibitors by fragmenting scaffolds into core and accessory modules and ligating them via hydrazone chemistry directly on assay plates. Applied to MraY inhibitors, the approach rapidly identified potent muraymycin-based analogues with broad-spectrum antibacterial activity, low resistance emergence, and in vivo efficacy. Structural studies uncovered a previously unobserved HS6-centered binding mode, expanding the mechanistic repertoire of MraY inhibition. Converting hydrazones to stable amide/anilide analogues retained potency and improved suitability for in vivo studies. The methodology also translated to tubulin-binding natural products, delivering functional analogues, highlighting its versatility. Future directions include: optimizing hydrophobic substituents to further improve selectivity index and reduce cytotoxicity; enhancing activity against challenging Gram-negative species (e.g., P. aeruginosa); expanding accessory fragment diversity for broader SAR; and applying the platform to other membrane targets or for rapid conjugation with functional moieties.

- Hydrazone stability: Hydrazones are hydrolytically labile, limiting in vivo utility and necessitating conversion to stable linkages for advanced studies. - Reaction variability: Although most in situ ligations reached ≥80% conversion, some hydrazone formations were partial or failed depending on hydrazine structure, potentially confounding initial activity readouts. - Screening throughput/data depth: Initial library enzymatic and MIC screens were semi-quantitative (n=1 for library-level MICs), requiring resynthesis and full validation for hits. - Spectrum gaps: Activity against Pseudomonas aeruginosa was limited or variable among stable analogues (MIC 8–128 µg/mL), indicating room for further optimization. - Off-target/cytotoxicity: While moderate (HepG2 IC50 12–25 µM), broader selectivity profiling (e.g., versus human GPT) and in vivo toxicity studies are needed to ensure safety margins. - Tubulin case caution: For tubulin-binding cores with higher intrinsic activity, low ligation conversion can mask true SAR by core-driven effects; rigorous purity verification (LC–MS) is needed for candidate hits.