Green Bismuth Catalyst Forges Antimicrobial Pyridopyrimidines in One Pot

A team of chemists and microbiologists has unveiled a one-pot, three-component strategy for building pyrido[2,3-d]pyrimidine derivatives using bismuth(III) triflate, a low-toxicity Lewis acid catalyst, in aqueous ethanol. The study, published in Results in Chemistry, combines laboratory synthesis with an unusually thorough computational workup, including density functional theory calculations, molecular docking, molecular dynamics simulations, pharmacophore mapping and drug-likeness screening. The result is a blueprint for how green chemistry and in silico drug discovery can be woven together to accelerate the hunt for new antimicrobial agents at a time when resistance to existing antibiotics and antifungals continues to climb.

The synthetic route itself is elegantly simple. Substituted aromatic aldehydes, methyl cyanoacetate and barbituric acid are combined in a 1:1 mixture of ethanol and water at 85 degrees Celsius with just 0.03 mol percent of Bi(OTf)3. The catalyst orchestrates a domino Knoevenagel-Michael sequence: it coordinates to the ester oxygen of methyl cyanoacetate, sharpening the electrophilicity of the carbonyl carbon, while the aldehyde-derived enolate attacks to form a carbon-carbon bond. The resulting alpha,beta-unsaturated intermediate is then struck by the weakly nucleophilic nitrogen of barbituric acid, and intramolecular cyclization closes the fused pyridopyrimidine ring. Eight derivatives, bearing nitro, bromo, hydroxyl, methyl, methoxy or unsubstituted phenyl groups, were isolated in yields ranging from 76 to 91 percent within two to five hours, with minimal side products and no need to purify intermediates.

What makes the protocol genuinely green is the solvent system and the catalyst. Bismuth(III) salts have gained a reputation as environmentally benign alternatives to conventional Lewis acids because they are air- and moisture-stable, functionally tolerant, commercially available and far less corrosive than many metal halides. Compared with earlier methods, the advantages are clear. A DMAP-catalyzed protocol requires ultrasonic irradiation in toxic DMF, ZrO2 nanoparticles demand nanomaterial synthesis and raise agglomeration concerns, and the sulfonated SBA-15 catalyst involves a cumbersome multistep preparation. The bismuth route runs in a drinkable solvent mixture under ordinary reflux, and the catalyst can be recovered, although reusability tests showed a gradual decline: the model reaction yielded 79 percent in the first cycle, 74 percent in the second, 65 percent in the third and only 33 percent in a fourth run that stretched to nine hours. The authors suggest that immobilization strategies could extend catalyst lifetime in future work.

On the computational side, the team optimized all eight molecules at the B3LYP/6-31G(d,p) level of density functional theory, confirming that each geometry corresponds to a true energy minimum. Frontier molecular orbital analysis revealed that the nitro-substituted compounds 4a and 4b possess the smallest HOMO-LUMO gaps, at roughly 0.135 electron volts, making them the most electronically soft and reactive members of the series, primed for donor-acceptor interactions with biological macromolecules. In contrast, the bromo, methyl and unsubstituted derivatives 4c, 4f and 4g showed the largest gaps and the greatest hardness, indicating higher kinetic stability. Mulliken charge analysis and molecular electrostatic potential maps pinpointed the carbonyl oxygens and ring nitrogens as the electron-rich hotspots most likely to engage in hydrogen bonding with protein targets, while the N-H hydrogens carried the complementary positive charge.

The biological evaluation used the broth dilution method to determine minimum inhibitory concentrations against four bacterial strains, the Gram-negative Escherichia coli and Pseudomonas aeruginosa and the Gram-positive Staphylococcus aureus and Streptococcus pyogenes, benchmarked against ampicillin, and three fungal strains, Candida albicans, Aspergillus niger and Aspergillus clavatus, benchmarked against griseofulvin. The structure-activity trends were striking. The unsubstituted compound 4g and the methoxy-bearing 4h were the standout antibacterial agents, with MIC values as low as 65 and 62 micrograms per milliliter respectively against E. coli and S. aureus, figures that actually beat ampicillin on those strains. The methyl derivative 4f was the most potent antifungal, matching griseofulvin’s 100 micrograms per milliliter against A. niger. Electron-donating groups such as methyl and methoxy generally enhanced activity, likely by improving lipophilicity and membrane penetration, while nitro groups tended to raise MIC values.

To explain these observations mechanistically, the researchers docked all eight compounds into two clinically significant enzymes: bacterial beta-lactamase, the molecular engine of antibiotic resistance, and sterol 14-alpha demethylase CYP51, the fungal enzyme that builds ergosterol membranes. Using AutoDock 4.2 with a Lamarckian genetic algorithm, they found that compounds 4b and 4h bound beta-lactamase most favorably, with estimated free energies of binding of minus 9.30 and minus 9.43 kilocalories per mole, stabilized by networks of hydrogen bonds, pi-alkyl contacts and pi-cation interactions with residues such as ARG 661, PRO 298 and TYR 611. Against CYP51, compound 4b again led with minus 7.79 kilocalories per mole, and the methyl-substituted 4f formed six hydrogen bonds within the active site, consistent with its antifungal potency. Torsional free energies were nearly uniform across the series, indicating that differences in affinity stem from noncovalent interaction networks rather than ligand flexibility.

Molecular dynamics simulations then stress-tested the most promising complexes over 100-nanosecond production runs using the Maestro-Desmond package with the OPLS3e force field. The 4f-beta-lactamase complex held a backbone root mean square deviation averaging 1.60 angstroms, comfortably below the 2.0 angstrom threshold that validates docking reliability, and stabilized after an initial 30-nanosecond equilibration. The 4d-CYP51 complex averaged 1.40 angstroms, closely tracking the reference antifungal ligand VT1 at 1.35 angstroms. Root mean square fluctuation analysis showed that flexibility was confined to loop and terminal regions, leaving the binding pockets rigid. Contact histograms revealed that water-bridged hydrogen bonds, involving residues such as Ser64, Gln120, Thr122 and Tyr132, together with hydrophobic contacts, were the main glue holding the complexes together throughout the simulations.

The pharmacoinformatic layer of the study added yet another dimension. POM analysis, which classifies bioactivity based on dipolar interactions between electron-rich and electron-deficient centers, identified three distinct antitumor pharmacophore regions built from NH donor to carbonyl oxygen acceptor pairs, plus a single antibacterial pharmacophore involving an NH-CO motif. OSIRIS toxicity screening found no major mutagenic, tumorigenic, irritant or reproductive risks across the series, with compounds 4d and 4g earning the highest drug scores at 0.70 and 0.87. Molinspiration calculations confirmed that all eight molecules respect Lipinski’s criteria almost entirely: molecular weights below 500 daltons, five or fewer rotatable bonds, moderate polarity and acceptable lipophilicity, with only the heavily nitrated pair showing a single violation.

Taken together, the work demonstrates a rare degree of coherence between experiment and computation. The compounds with the best MIC values, notably 4f, 4g and 4h, also posted strong docking energies and stable dynamic behavior, while the electrostatic potential maps independently flagged the same carbonyl and ring nitrogen atoms that docking identified as interaction points. The authors conclude that pyrido[2,3-d]pyrimidines represent a promising scaffold for further antimicrobial development, and that targeted substitution of the aromatic ring is the key lever for tuning activity. With antimicrobial resistance declared one of the top global public health threats, a synthetic method that is cheap, green and high-yielding, paired with a validated computational pipeline for prioritizing candidates, offers exactly the kind of integrated approach the field has been calling for. The next step will be translating these laboratory and in silico hits into lead compounds capable of surviving the far harsher tests of animal models and, eventually, the clinic.

Subject of Research: Green synthesis and antimicrobial evaluation of pyrido[2,3-d]pyrimidine derivatives catalyzed by bismuth(III) triflate

Article Title: Bi(OTf)₃-catalyzed one-pot synthesis and antimicrobial evaluation of bioactive pyridopyrimidines supported by DFT, molecular dynamics, POM and pharmacophore analyses

Article References: Mehta, A. W., Abdel-Megid, M., Patil, R. C., Ahmed, S., Salem, M. E., Abu-Rayyan, A., Shtaiwi, A., Hajam, Y. A., Bhat, A. R., Eissa, M. E., Agisho, H. A., Mujahid, M. H., Yamari, I., Raza, N., & Zbancioc, A. M. (2026). Bi(OTf)₃-catalyzed one-pot synthesis and antimicrobial evaluation of bioactive pyridopyrimidines supported by DFT, molecular dynamics, POM and pharmacophore analyses. Results in Chemistry, 30, Article 103849. https://doi.org/10.1016/j.rechem.2026.103849

Image Credits: AI Generated

DOI: 10.1016/j.rechem.2026.103849

Keywords: pyridopyrimidines, bismuth(III) triflate, multicomponent reaction, green chemistry, antimicrobial activity, molecular docking, molecular dynamics, DFT, beta-lactamase, CYP51, drug discovery, pharmacophore