Schiff bases are organic compounds defined by the azomethine group, a carbon-nitrogen double bond formed when a primary amine reacts with an aldehyde or ketone. This simple condensation reaction produces a family of molecules with a remarkable range of biological activities, including antiviral, antimicrobial, antifungal, antitumor, and antibacterial properties, which has made them a staple of medicinal chemistry. Aryl Schiff bases, in which the imine bond is attached to an aromatic ring, offer greater conjugation and stability than their aliphatic counterparts. Beyond biology, these versatile molecules serve as corrosion inhibitors, catalysts, metal-cation complexing ligands, optical switching materials, nonlinear optical materials, and sensors, making them one of the most broadly useful compound classes in modern chemistry.

The molecule at the heart of the new study, bis(2-nitrobenzylidene)hexane-1,6-diamine, contains two independent imine bonds, each of which can adopt either an E or Z configuration around the carbon-nitrogen double bond. In principle this yields four geometrical isomers, but because identical groups are attached to both imine bonds, only three distinct diastereomers are possible: ZZ, EZ, and EE. The team optimized all three structures at the B3LYP/6-311G(d,p) level of theory in the gas phase using the Gaussian 16 suite of programs. The energy ranking was unambiguous. The EE isomer, BSB-EE, lies 6.97 kilocalories per mole below the EZ isomer and 13.62 kilocalories per mole below the ZZ isomer, a direct consequence of steric hindrance in the Z-configurations where the bulky nitro-substituted phenyl rings crowd one another.

Symmetry also leaves a striking fingerprint on the dipole moments of the three isomers. BSB-EE, with its mirror-symmetric arrangement, has a calculated dipole moment of exactly zero, making it a genuinely non-polar molecule. By contrast, the EZ and ZZ diastereomers are strongly polar. Selected bond metrics remain remarkably consistent across all three forms, with carbon-nitrogen double bond lengths of 1.269 and 1.267 angstroms in the E and Z configurations respectively, and single carbon-nitrogen bonds of 1.453 and 1.457 angstroms. The dihedral angle around the imine linkage cleanly distinguishes the configurations, measuring close to 179 degrees for the E form and only 7 to 9 degrees for the Z form. These geometric signatures suggest that BSB-EE could serve as a robust scaffold where minimal intermolecular dipole interactions are desired, or where self-assembly is governed by shape and pi-conjugation rather than polarity.

Vibrational analysis provided a bridge between the computed structures and laboratory reality. The molecules, with the formula C20H22N4O4, exhibit 144 vibrational modes, and all calculated wavenumbers were scaled by a factor of 0.967 following the recommendations of the Computational Chemistry Comparison and Benchmark Data Base. The characteristic imine stretching vibration, the classic signature of Schiff base formation, appears in the 1600 to 1640 per centimeter region. For BSB-EE the calculated asymmetric C=N stretch falls at 1648 per centimeter, in close agreement with the experimentally measured value of 1633 per centimeter reported in earlier work. Aliphatic carbon-hydrogen stretches were computed at 2859 through 2969 per centimeter, aromatic stretches at 3072 through 3116 per centimeter, and the N=C-H stretch at 2982 per centimeter, all consistent with experimental benchmarks and confirming that the theoretical treatment faithfully reproduces the molecule’s vibrational behavior.

Optical properties revealed an equally clear stereochemical trend. Using time-dependent density functional theory, the team calculated the ultraviolet-visible absorption spectra of all three isomers. All three absorb in the near-ultraviolet to visible range, but BSB-EE absorbs at the shortest wavelength, 320 nanometers, while BSB-EZ absorbs at 340 nanometers and BSB-ZZ at 350 nanometers. More importantly, the oscillator strength, which measures the probability and intensity of the main electronic transition, decreases steadily from BSB-EE to BSB-EZ to BSB-ZZ. The dominant transitions are pi to pi-star promotions within the conjugated bis-Schiff base system, and the pronounced drop in oscillator strength indicates that the key transition is significantly more allowed in the EE geometry, likely because of superior orbital overlap. For applications requiring maximal visible absorption or fluorescence, BSB-EE emerges as the most promising candidate of the three.

Frontier molecular orbital analysis deepened the picture of electronic reactivity. The energy gap between the highest occupied and lowest unoccupied molecular orbitals governs kinetic stability, with large gaps indicating stability and small gaps indicating reactivity. BSB-EE has the largest gap at 4.561 electron volts, compared with 4.140 electron volts for BSB-EZ and 4.092 electron volts for BSB-ZZ, all falling in a moderate range of kinetic stability. Interestingly, BSB-EE is the strongest electron donor among the three isomers, with the most negative HOMO energy, while the EZ and ZZ forms are marginally better electron acceptors. Global reactivity descriptors including electronegativity, chemical hardness, ionization potential, and electron affinity differ only modestly across the isomers, reinforcing the central lesson that small geometric changes can meaningfully shift electronic distribution and reactivity without wholesale transformation of the molecule’s chemical character.

Perhaps the most surprising results came from the nonlinear optical calculations. Nonlinear optical materials respond to intense electromagnetic fields such as laser irradiation, producing phenomena like second-harmonic generation, third-harmonic generation, and two-photon absorption, with applications in optoelectronics, telecommunications, and photonics. Schiff bases typically combine electron-donating and electron-accepting groups linked by conjugated pi-systems, a push-pull architecture that enhances intramolecular charge transfer and boosts nonlinear response. When compared against urea, the standard reference material, the bis-Schiff base derivatives showed strongly enhanced polarizability. BSB-EZ displayed the strongest second-order response, while BSB-EE showed no measurable second-order activity, a consequence of its centrosymmetry. Yet BSB-EE posted a large positive second hyperpolarizability of 16.48 times ten to the minus forty electrostatic units, indicating powerful third-order nonlinearity, whereas the EZ and ZZ isomers showed negative gamma values of smaller magnitude, implying distinctly different third-order response characteristics relevant to Kerr effects and two-photon processes.

The team then turned to biology, docking the optimized BSB-EE structure into two protein targets from the RCSB Protein Data Bank: 6LU7, the main protease of SARS-CoV-2, a critical enzyme in viral replication, and 1HSG, the crystal structure of HIV protease complexed with an inhibitor. Using AutoDock 4.2, the ligand achieved favorable binding energies at both active sites. In 6LU7, hydrogen bonds form between the imine nitrogen atoms and the residues ASP30 and TYR32 at distances of 2.37 and 2.29 angstroms, while in 1HSG a hydrogen bond forms with THR216 at 2.03 angstroms. The phenyl rings of the ligand settle completely into hydrophobic pockets of the active sites, supplemented by van der Waals, pi-alkyl, and pi-anion electrostatic interactions with surrounding amino acid side chains. Benchmarking against doxorubicin, a known inhibitor with a binding affinity of minus 6.98 kilocalories per mole, revealed superior predicted inhibitory potential for BSB-EE against 6LU7, underscoring its promise as a lead compound.

The authors are careful to note the limitations of their approach. Gas-phase calculations neglect solvation, specific solute-solvent interactions, and crystal packing, and docking scores are only predictions that require experimental validation through enzymatic assays and binding studies, followed by structure-activity optimization and thermodynamic and kinetic analyses. Future work will include solvent-embedded infrared and nonlinear optical analyses and molecular dynamics-assisted docking to solidify structure-property-bioactivity correlations under physiologically relevant conditions. Natural bond orbital analysis already points to intramolecular charge transfer between electron-rich carbon and oxygen atoms and electron-deficient, positively polarized nitrogen sites, consistent with the push-pull delocalization that underpins the molecule’s nonlinear optical promise. Taken together, the study delivers a systematic rationale for treating BSB-EE as a versatile lead, one molecule whose shape alone tunes its stability, its optics, and its potential as a weapon against viral proteases.

Subject of Research: Computational analysis of the physicochemical properties and protein-binding potential of bis-Schiff base isomers derived from hexamethylenediamine and o-nitrobenzaldehyde

Article Title: Physicochemical properties (IR, FMOs, NLO, and molecular docking) of bis-Schiff bases derived from hexamethylenediamine and o-nitrobenzaldehyde

Article References: Rezvan, V. H., & Rajendran, S. (2026). Physicochemical properties (IR, FMOs, NLO, and molecular docking) of bis-Schiff bases derived from hexamethylenediamine and o-nitrobenzaldehyde. Discover Chemistry, 3(1), Article 557. https://doi.org/10.1007/s44371-026-00992-z

Image Credits: AI Generated

DOI: 10.1007/s44371-026-00992-z

Keywords: bis-Schiff base, DFT, B3LYP, frontier molecular orbitals, nonlinear optical properties, molecular docking, SARS-CoV-2 main protease, HIV protease, geometrical isomers, UV-Vis spectroscopy, FTIR, medicinal chemistry

The core insight behind SynCraft is what the authors call the synthesis cliff. In medicinal chemistry, small structural changes can produce disproportionately large swings in synthetic feasibility, much as activity cliffs describe tiny modifications that dramatically alter biological potency. A molecule that is unreachable by any known reaction sequence may sit only one or two atom-level edits away from a close analogue that a retrosynthesis planner can route with ease. The trick, then, is not to abandon the original design and start over, but to find the minimal set of structural modifications that carries the molecule over the cliff while preserving everything that made it attractive in the first place: its binding interactions, its pharmacophore, its structural novelty.

Existing approaches to this problem tend to fall into two camps, and both come with costs. Post hoc filtering simply discards generated molecules that fail a synthesizability check, which wastes much of the creative output of generative models and biases libraries toward mundane chemistry. Projection-based methods go further, forcing an infeasible molecule onto the nearest point in a predefined space of synthetically accessible compounds. The problem is that this projection can wreck the very features that mattered. A key pharmacophore can be disrupted, a crucial hydrogen-bond donor can vanish, and the structural novelty that motivated the design in the first place can be flattened into something indistinguishable from existing patent literature. SynCraft rejects both strategies in favor of something more precise.

The framework reframes synthesizability optimization as a structural editing problem and hands that problem to a large language model, but with a crucial twist. Rather than asking the model to write out complete molecular structures as SMILES strings, a task at which language models are notoriously fragile because a single misplaced character can render an entire string chemically invalid, SynCraft asks the model to predict executable sequences of atom-level edits. Each edit is a discrete, verifiable operation on the existing structure, applied through cheminformatics tooling such as RDKit. This division of labor plays to the strengths of each component: the language model contributes chemical intuition and reasoning about which modifications are likely to help, while deterministic software guarantees that every edit is applied correctly and that the resulting molecule is always structurally valid.

To teach models this editing skill, the researchers assembled a training corpus of 3,332 molecular pairs, each accompanied by reasoning traces that explain why the edits transform an inaccessible molecule into a synthesizable one. The benchmarking results show that SynCraft outperforms state-of-the-art baselines at generating synthesizable analogues with high structural fidelity, meaning the edited molecules remain close to the original designs rather than drifting into unrelated chemical territory. Notably, the framework performed comparably whether it was running on proprietary frontier models or on open-weight backends, an important practical finding because it suggests the approach does not depend on access to the most expensive commercial systems. Laboratories with modest budgets can deploy the same chemistry on locally hosted models.

Perhaps the most striking validation of the approach is retrospective: SynCraft, guided by interaction-aware prompting that describes the protein-ligand contacts a molecule must preserve, successfully replicated the kind of editing judgments that expert medicinal chemists make intuitively. In tests on inhibitors of PLK1, a well-studied cancer target, the system’s proposed edits mirrored expert reasoning about which structural features to protect and which to sacrifice. The framework then went a step further into prospective territory, rescuing high-scoring RIPK1 candidate molecules that had been generated by earlier molecular generation models but discarded because they could not be synthesized. These were, in effect, good ideas trapped behind a synthetic wall, and SynCraft found the edits that let them through without destroying their binding profiles.

To make the workflow practical rather than merely demonstrative, the researchers packaged SynCraft as an agent skill, a modular capability that an autonomous AI agent can invoke. They demonstrated an end-to-end rescue workflow on candidate molecules generated for the SARS-CoV-2 main protease, the target that anchored much of the pandemic-era antiviral effort. The workflow verified that edited molecules retained their binding modes, checking interaction preservation with tools such as the protein-ligand interaction profiler, so that the rescue operation could be audited at the level of individual hydrogen bonds and hydrophobic contacts rather than trusted on faith. This kind of verification matters in a field where black-box outputs have historically been greeted with justified skepticism by experimentalists.

The broader significance of the work lies in how it repositions large language models within chemistry. Much of the early excitement, and disappointment, around LLMs in molecular design came from attempts to have them generate molecules directly, treating chemical structures as just another text format. The results were often syntactically broken or chemically nonsensical, and studies measuring the robustness of chemical language models to different molecular representations revealed how brittle the approach was. SynCraft belongs to a more mature second wave in which the language model serves as a reasoning engine over structured, tool-verifiable operations, a pattern that echoes the chain-of-thought techniques that unlocked reasoning in general-purpose models. The model thinks; deterministic chemistry software acts; validators check the result.

Accessibility may prove as consequential as accuracy. The SynCraft implementation, reproduction scripts for every experiment, and the agentic skill package have been released under an MIT license on GitHub, with the test sets, training corpus, inference outputs and case-study structures deposited openly on Figshare. Any group running generative drug design pipelines can therefore bolt synthesizability rescue onto their existing workflow without licensing barriers. As generative models continue to flood virtual screening campaigns with candidates, the gap between what AI can imagine and what chemists can build has been the field’s most persistent embarrassment. A framework that closes that gap with minimal, principled edits, while preserving the binding interactions that make a molecule worth making, turns discarded designs into a usable reservoir of drug candidates, and suggests that the future of AI-driven discovery lies less in generating more molecules than in knowing exactly which few atoms to change.

Subject of Research: Using large language models to optimize the synthesizability of AI-generated drug candidate molecules through predicted structural edit sequences

Article Title: Guiding large language models to predict edit sequences for molecular synthesizability optimization

Article References: Li, J., & Lai, L. (2026). Guiding large language models to predict edit sequences for molecular synthesizability optimization. Nature Machine Intelligence. https://doi.org/10.1038/s42256-026-01304-x

Image Credits: AI Generated

DOI: 10.1038/s42256-026-01304-x

Keywords: large language models, molecular synthesizability, drug discovery, generative AI, medicinal chemistry, SynCraft, molecular editing, retrosynthesis, SARS-CoV-2 main protease, RIPK1, PLK1, cheminformatics