The Global Beta-Lactamase Database (GBLD) is a next-generation, integrative computational resource designed to transcend the limitations of sequence-centric repositories. While existing databases such as BLDB have significantly advanced the cataloguing of β-lactamase sequences, functional annotations, and select crystal structures, they offer no standardized pipeline for homology modelling, structural validation, surface electrostatic analysis, molecular docking, or molecular dynamics (MD) simulation across the breadth of known variants. GBLD addresses this critical gap by providing a uniform, reproducible, end-to-end computational workflow from raw sequence retrieval to atomistic MD simulation enabling researchers to interrogate substrate recognition, inhibitor binding, and structural dynamics for any β-lactamase variant of interest.
Beta-lactamases are the most widespread mechanism of antibiotic resistance among Gram-negative pathogens, comprising over 7,739 unique sequences spread across Ambler classes A, B (B1, B2, B3), C, and D. The existing BLDB currently indexes 1,805 three-dimensional structures across all classes (Class A: 681; B1: 470; B3: 104; C: 249; D: 285), yet the vast majority of the 7,739+ known sequences lack any experimentally resolved or computationally validated structure. This structural gap means that for hundreds of clinically significant variants particularly newly emerging ones researchers have no standardized way to understand active site geometry, substrate binding pose, or inhibitor interaction profile.
Existing databases are also functionally constrained. BLDB provides sequence BLAST, and links to PDB structures where available. The integrative β-lactamase database published by ASM (2019) improved on earlier tools by integrating sequence, structure, MIC, and kinetic data, but still does not offer a simulation or docking pipeline. The absence of a unified platform integrating sequence, structural, and experimental data. GBLD directly resolves this unmet need.
GBLD is built on four core objectives that collectively exceed the scope of all prior β-lactamase databases:
- Universal structural coverage — Generate high-quality homology models for all β-lactamase variants lacking experimental structures, using SWISS-MODEL with standardized template selection criteria
- Structural validation and quality control — Apply GROMACS energy minimization and PROCHECK Ramachandran analysis (>90% residues in favoured region as acceptance threshold) to guarantee model reliability
- Substrate and inhibitor docking — Execute both blind and targeted molecular docking using AutoDock Vina for all variants with clinically relevant substrates (e.g., cefotaxime, meropenem, avibactam)
- Molecular dynamics simulation — Perform atomistic MD simulations on docked complexes to capture dynamic interaction profiles, conformational flexibility, and binding stability over simulation time