Finding New Scaffolds Among Millions of Molecules: Fujifilm and Academic Researchers Validate Bacterial Signaling Enzyme Inhibitors

One approach to finding new drugs against antibiotic-resistant bacteria is to interfere with the signaling systems they use to sense their environment and adjust their pathogenic behavior. Working with several Japanese research institutions, Fujifilm used computational screening to identify new bacterial enzyme inhibitors and confirmed their activity through chemical synthesis and experiments, turning molecules that had existed only in a virtual database into physical compounds available for further research.

The research focused on histidine kinases. These enzymes participate in bacterial signaling and can influence pathogenicity, antibiotic resistance, and biofilm formation. The team targeted the relatively conserved H-box region, aiming to block the phosphorylation reaction that initiates signaling. Because this bacterial signaling system does not exist in humans, it is an attractive target for new drugs, but that does not mean its inhibitors are inherently safe.

The starting point was waldiomycin, a known natural inhibitor. According to Okayama University, this molecule is toxic to animal cells and has a complex structure, making both synthesis and subsequent modification difficult. The study therefore sought different molecular scaffolds, hoping to retain activity against the target while obtaining chemical structures that are easier to synthesize and modify.

Fujifilm calls the technology AI-AAM, combining artificial intelligence with “amino acid mapping.” According to Osaka University, the researchers first calculated descriptors from part of waldiomycin’s structure to characterize which amino acids the molecule might interact with, then searched a virtual database of approximately 12 million compounds for molecules with similar characteristics. This approach allows compounds with different shapes to be selected without first obtaining a detailed three-dimensional structure of the target protein.

The screening ultimately yielded 714 candidates. Osaka University further stated that the team selected new scaffolds and synthesized 17 compounds and derivatives, testing their effects on histidine kinases from pathogenic bacteria. Two showed stronger inhibitory activity. This provides evidence of progress from computational prediction to experimental validation; the 714 candidates cannot all be regarded as molecules with confirmed activity.

Experimental findings also support the proposed site of action. After the researchers altered amino acid residues in the H-box region, the derivatives’ inhibitory activity decreased markedly. Together with results from mass spectrometry and other analyses, this supports their action in the same region as waldiomycin. These findings give the screening results a mechanistic basis, but inhibition of enzyme activity does not directly establish that an infection can be controlled.

The findings were published in The Journal of Antibiotics on September 30, 2026. Collaborating institutions included Osaka University, Okayama University, Kindai University, the Institute of Microbial Chemistry, Kyushu University, and Fujifilm. The progress so far consists of obtaining new scaffolds and demonstrating their activity experimentally; clinical efficacy has not yet been established. The next steps still require optimizing the molecules and determining whether they can act effectively in an infection setting, and whether their safety and exposure in the body are sufficient to support treatment.