Water-Soluble Diazaborines: Selective Gram-Negative Antibiotic Candidates for Synergy Studies and R&D
Water-Soluble Diazaborines:
Selective Gram-Negative Antibiotic Candidates for Synergy Studies and R&D
Three diazaborines from a peer-reviewed J. Med. Chem. 2026 study — sodium salts ready for synergy work, selectivity research, and follow-on medicinal-chemistry optimisation of a scaffold class where boron-containing drugs (such as the FDA-approved vaborbactam) already have approval
NorrChemica supplies, in research quantities, three diazaborines from the recently published study by Ilina et al. in the Journal of Medicinal Chemistry (2026, 69, 3796–3810) — including the water-soluble sodium salts of the lead compound (paper compound 11, supplied as VI706, also designated Diazaborine 62) and of its closest matched-pair structural reference (paper compound 13, supplied as VI880, Diazaborine 64).
VI706 is the water-soluble sodium salt of the published lead diazaborine (paper compound 11) — same reported MIC as the parent free acid, but in the practical form needed for studies in physiological real-world conditions where DMSO becomes a liability. The lead shows MIC 6.25 μM against E. coli (ATCC 25922), A. baumannii, and S. enterica ser. Typhimurium; FICI 0.25 with colistin; and significant rescue of lethally infected Galleria mellonella larvae at 1.13 mg/kg, with 2.81 mg/kg performance not significantly different from ciprofloxacin at 20 mg/kg under the study conditions. Positioned for antibacterial drug-discovery programmes running combination and synergy studies, selectivity research, and follow-on medicinal chemistry on a scaffold class where boron-containing drugs (such as the FDA-approved vaborbactam) already have approval.
What the study established
Why the water-soluble sodium salts
Free-acid diazaborines in the reported antibacterial series are poorly soluble in water. In aqueous biological formats — MIC plates, plasma-stability assays, in vivo dosing — this creates a practical vehicle problem. DMSO stocks can be used for poorly soluble compounds and are widely employed in drug-discovery workflows, but DMSO is not an inert co-solvent: it can affect protein–ligand binding kinetics, membrane properties, bacterial physiology, and antimicrobial readouts, and at higher concentrations it can interfere with plasma-stability workflows. For the aqueous-formulation-critical assays the published study therefore avoided DMSO altogether and worked with the water-soluble sodium salts of the active compounds.
The published profile of paper compound 11 and its sodium salt (VI706)
VI706 is the experimentally validated water-soluble sodium salt of the published lead diazaborine (paper compound 11). The paper reports identical MIC values for the salts and parent compound, making the lead profile directly relevant to VI706 while preserving the practical aqueous-handling advantage of the salt form.
Published profile · paper compound 11 and sodium salt VI706 · Ilina et al. 2026
- Antibacterial spectrum MIC 6.25 μM against E. coli ATCC 25922 and uropathogenic E. coli CFT073; 12.5 μM against the enterohaemorrhagic clonal strain CB9615 and the multi-drug-resistant uropathogenic UMN026; 6.25 μM against A. baumannii and S. enterica ser. Typhimurium; 12.5 μM against K. aerogenes. MIC against K. pneumoniae is substantially higher (75 μM), reflecting the species-dependent variability typical of FabI-targeted agents and possibly reflecting FabI sequence divergence at the substrate-binding pocket (as the paper discusses for S. aureus).
- Target engagement Preliminary Ki 0.32 μM against isolated E. coli FabI, consistent with the proposed covalent B–NAD+ adduct mechanism (crystallographic basis PDB 5CG1 / 5CG2 for related thiocarbamoylated benzodiazaborines).
- Cytotoxicity window IC50 106 ± 13 μM in HepG2 cells and 240 ± 9 μM in Hs27 fibroblasts (72 h) — ~17-fold over the antibacterial MIC in HepG2 and ~38-fold in Hs27. Considerably more favourable than the reference FabI inhibitor triclosan under matched 72 h assay conditions in this study, which gives IC50 18 ± 2 μM in HepG2 and 21 ± 2 μM in Hs27.
- Plasma stability The sodium salt retains its characteristic 11B NMR resonance at ~+6.5 ppm after 72 h in human plasma at 37 °C — directly supporting the aqueous stability of VI706 under physiological-medium conditions.
- Colistin synergy FICI 0.25 in E. coli ATCC 25922, with four-fold reduction of colistin's MIC (from 1 to 0.25 μg/mL) at 1.25 μM compound concentration. Eight other partners screened in the paper (metronidazole, sulfadiazine, meropenem, amoxicillin, ciprofloxacin, metformin, citric acid, choline chloride) did not show synergy.
- In vivo efficacy Galleria mellonella E. coli infection model: significant rescue at 1.13 mg/kg compared with untreated infected control, and at 2.81 mg/kg producing survival not significantly different from ciprofloxacin at 20 mg/kg under the study conditions. No significant toxicity at the highest dose tested (5.63 mg/kg).
- Resistance MPC/MIC = 32, two-fold higher than ciprofloxacin under matched conditions; spontaneous resistance frequency 2 × 10−8 to 9 × 10−9, in the same order of magnitude as ciprofloxacin (≈10−9).
Boron in approved and clinical-stage antibacterials
Boron is established in approved antibacterials. Vaborbactam (in Vabomere, with meropenem) — a cyclic boronic acid β-lactamase inhibitor — was approved by the FDA in 2017 for complicated urinary tract infections. Its mechanism rests on the boron atom's ability to switch between sp2 trigonal and sp3 tetrahedral hybridisation in the β-lactamase active site — the same geometric flexibility that underpins diazaborine engagement of FabI's NAD+ cofactor.
Boron-containing drugs · antibacterial and beyond
| Compound | Status | Mechanism / use |
|---|---|---|
| Vaborbactam (in Vabomere, with meropenem) | FDA approved 2017 | Cyclic boronic acid β-lactamase inhibitor; adult complicated UTI including pyelonephritis |
| Taniborbactam (with cefepime) | FDA CRL Feb 2024 (CMC); no clinical issues identified | Bicyclic boronate; inhibits Ambler A/C/D serine β-lactamases and selected class B metallo-β-lactamases, notably NDM and VIM |
| Xeruborbactam (QPX7728) | Investigational · Phase 1 | Bicyclic boronate β-lactamase inhibitor; broad serine + metallo-β-lactamase coverage; developed in β-lactam combinations |
| Bortezomib · ixazomib | FDA approved (oncology) | Boron-containing proteasome inhibitors; multiple myeloma — not antibacterial, included for class breadth |
| Tavaborole | FDA approved (dermatology) | Boron-containing antifungal; onychomycosis — not antibacterial |
| Crisaborole | FDA approved (dermatology) | Boron-containing topical PDE4 inhibitor; atopic dermatitis — not antibacterial |
The antibacterial subset of this list — vaborbactam approved, taniborbactam and xeruborbactam in clinical development — establishes the regulatory and mechanistic precedent for cyclic boronate antibacterials. Diazaborines sit within the broader boron-drug lineage as a distinct chemotype: different target (FabI rather than β-lactamases), different binding partner (the NAD+ ribose 2′-OH rather than an enzyme serine residue), and different therapeutic angle (combination with last-resort polymyxins against multi-drug-resistant Gram-negative pathogens, rather than restoration of β-lactam activity against β-lactamase producers). The historical assumption that boron-containing motifs are intrinsically unsuitable for drugs — which originally halted the Sandoz diazaborine programme in the 1980s — is no longer defensible given the regulatory record above.
What researchers can do next with this scaffold
The biggest hammer is not always the right tool for the job. A broad-spectrum antibiotic is not always the right way to treat an infection — not when it also disrupts the gut bacteria the patient depends on. Developing highly selective antibacterials is one of the long-standing challenges in the field and one of its most sought-after, tantalising holy grails. The diazaborine series shows three practical routes to a more selective antibacterial: change the diazaborine structure, change how the compound enters bacterial cells, or combine it with another antibiotic such as colistin. That gives researchers concrete starting points for follow-on work — improving potency, narrowing the bacterial spectrum, testing new combinations, or building analogues around the FabI-binding diazaborine core.
A built-in pathogen-selectivity window
The lead hits E. coli, S. Typhimurium, and A. baumannii — three pathogens that depend on FabI as their only enoyl-ACP reductase. P. aeruginosa, E. faecalis, and many gut microbiome organisms carry redundant isozymes (FabK, FabL, FabV) and are buffered against FabI inhibition. The Gram-negative-pathogen-selective profile is already in place; the open questions are how far structural design and further synergy studies can push it toward true narrow-spectrum therapy, and how much of the human microbiome can be left intact while the target organism is cleared.
Change the diazaborine structure
Masking the B–OH group (compound 49) abolishes Gram-negative activity and switches the scaffold to Gram-positive selectivity (S. aureus MIC 25 μM).
Combine it with another antibiotic
FICI 0.25 with colistin demonstrates that the spectrum can be extended externally through partner choice — not just through structural modification of the warhead.
Four research directions follow from these properties. Each is supported by data already in the published paper but is open follow-on work.
1. Selectivity tuning through minimal structural modification
The paper reports that a single structural change — masking the free B–OH group, as in the tetracyclic derivative compound 49 — abolishes activity against Gram-negative species (no inhibition of E. coli at 50 μM) and switches the compound to Gram-positive selectivity (S. aureus MIC 25 μM). The authors hypothesise that the masked B–OH no longer interacts with peptidoglycan diol residues in the Gram-positive cell envelope, allowing the compound to bypass that interaction and access its intracellular target. This is a research observation, not a fully established model — but it is exactly the kind of structural lever that a focused medicinal-chemistry campaign can exploit. The same scaffold may therefore serve as the starting point for a parallel optimisation track aimed at Gram-positive pathogens, alongside the Gram-negative-selective lead.
2. Expanding the combination space beyond colistin
The paper screened compound 11 against eight partners at single concentrations (metronidazole, sulfadiazine, meropenem, amoxicillin, ciprofloxacin, metformin, citric acid, choline chloride) and identified synergy only with colistin. The combination-discovery space remains largely unprobed. Proper checkerboard assays against outer-membrane permeabilizers (pentamidine and analogues), efflux-pump inhibitors, LPS-pathway modulators, and recently published Gram-negative adjuvants — clofoctol, minocycline, bavachin, the signal-peptidase inhibitor MD3, cannabidiol — have not been performed against this scaffold. Given the size of the polymyxin-adjuvant literature and the clinical urgency of post-mcr-1 colistin resistance, this is direct, low-risk follow-on work. VI706 is a tractable starting point for any group running such checkerboards.
3. Conjugate and delivery strategies
The paper showed that conjugating compound 11 to a phosphonium moiety through a cleavable carbamate linker (compound 68) opened activity in S. aureus (MIC 25 μM) that the parent free acid does not access. Conjugation to L- or D-tryptophan with phosphonium attached (compounds 77 and 78) reached MIC 6.25 μM in S. aureus with ~80% growth inhibition in E. faecalis at 50 μM. These results suggest that the diazaborine warhead retains activity when delivered into different bacterial compartments through different membrane-traversal mechanisms. The conjugate and prodrug space is wide open; linker chemistry can be designed around specific delivery hypotheses.
4. Cross-species FabI engineering and pathogen-selective design
The paper attributes the absence of activity against S. aureus and other Gram-positive species in part to FabI sequence divergence at the substrate-binding pocket between E. coli and S. aureus. The diazaborine scaffold could be tailored against specific FabI orthologs by structure-guided modification — for instance, an M. tuberculosis InhA-selective derivative for tuberculosis research, or a derivative selective for FabI variants in Burkholderia, Francisella, or other priority pathogens. The crystallographic foundation (PDB 5CG1 and 5CG2 for E. coli FabI bound to thiocarbamoylated benzodiazaborines) makes structure-guided optimisation of this kind tractable.
Available from NorrChemica
Three compounds from the published study are supplied in research quantities, in stock, dispatched from Helsinki, Finland, with batch-specific Certificate of Analysis and Safety Data Sheet. Research use only — not for human or veterinary use.
Available Now in the R&D Series
VI706 · Diazaborine 62 · The lead →
Sodium salt of compound 11. ≥95% purity. Full bioactivity profile — FabI inhibitor (preliminary Ki 0.32 μM), colistin synergy (FICI 0.25), and in vivo Galleria mellonella rescue efficacy.
VI880 · Diazaborine 64 · Matched-pair reference →
Sodium salt of compound 13. ≥98% purity. Matched-pair SAR reference isolating the contribution of the 3-amino group; preserves colistin synergy and FabI binding.
VI116 · Diazaborine 1 · Free-acid parent →
Phenyl-series free-acid parent scaffold. ≥95% purity. Structural reference and synthesis-route benchmark for the diazaborine series.
Request a quotation or place an order
Research quantities, dispatched from Helsinki, Finland with batch-specific CoA and SDS. Worldwide DDP available. For larger quantities or other diazaborine derivatives from the series, contact us directly.
Request a Quote → Browse R&D Series© 2026 SynFinn Discovery Oy / NorrChemica · All rights reserved · Content may not be reproduced without written permission