Design, Synthesis, and Antifungal Activity of 3-Substituted-2(5H)-Oxaboroles

Next generation antimicrobial therapeutics are desperately needed as new pathogens with multiple resistance mechanisms continually emerge. Two oxaboroles, tavaborole and crisaborole, were recently approved as topical treatments for onychomycosis and atopic dermatitis, respectively, warranting further studies into this privileged structural class. Herein, we report the antimicrobial properties of 3-substituted-2(5H)-oxaboroles, an unstudied family of medicinally relevant oxaboroles. Our results revealed minimum inhibitory concentrations as low as 6.25 and 5.20 μg/mL against fungal (e.g., Penicillium chrysogenum) and yeast (Saccharomyces cerevisiae) pathogens, respectively. These oxaboroles were nonhemolytic and nontoxic to rat myoblast cells (H9c2). Structure–activity relationship studies suggest that planarity is important for antimicrobial activity, possibly due to the effects of extended conjugation between the oxaborole and benzene rings.

O ver the past 70 years, only four classes of antifungal drugs have received FDA approval for the treatment of systemic infections: azoles, 1 allylamines, 2 polyenes, 3 and echinocandins. 4,5Today's treatments depend on new generations of these established classes; however, they have been unable to overcome the challenges posed by new, highly drug resistant pathogens. 6In particular, fungal infections represent an area where new treatments are desperately needed.Existing antifungals provide limited efficacy against Candida auris, C. albicans, Cryptococcus neoformans, and Aspergillus f umigatus, all of which are classified as critical pathogens by the World Health Organization. 7Roughly 1.7 million deaths per year worldwide are a result of fungal infections with immunocompromised patients being most at risk. 8In addition, most antifungal drugs have marked toxicity, which is incompatible with the long treatment regimen and increasingly higher doses that are required to combat infections caused by these multidrug resistant pathogens. 9n 2014, tavaborole (Kerydin, 1), a benzoxaborole, was approved by the FDA for the topical treatment of onychomycosis (i.e., toenail fungal infections). 10,11Although it has only been approved to treat topical infections, there are four other boronic acid-containing drugs approved for the treatment of various diseases, including systemic diseases such as atopic dermatitis [crisaborole (Eucrisa), 2], 12 a β-lactamase inhibitor [vaborbactam (Vabomere), 3] 13 and multiple myeloma [bortezomib (Velcade), 4, and ixazomib (Ninlaro), 5], 14,15 and with many more in various stages of clinical trials (Figure 1). 16Tavaborole and crisaborole are the only clinically approved benzoxaboroles, and both represent first-in-class drugs.Tavaborole inhibits tRNA-synthetase by trapping tRNA in the editing site, 17 and crisaborole is a potent phosphodiesterase-4 (PDE4) inhibitor, which increases intracellular cyclic adenosine monophosphate (cAMP) levels, reducing inflammatory mediators. 12Structure−activity relationship (SAR) studies on tavaborole have shown that the boron is essential for activity, and crystal structures of tavaborole bound to tRNA synthetase reveal critical interactions between the boron's unoccupied p-orbital and oxygen atoms (2′ and 3′) of the tRNA's 3-terminal adenosine. 17Mutations that are predicted to destabilize this interaction (D487G or D487N) abolish the activity of tavaborole, confirming that this interaction is essential for the observed antifungal activity. 18lthough benzoxaboroles have been intensely studied for their potential to treat various diseases, 19−21 no studies have focused on a related structural class such as 3-substituted-2(5H)-oxaboroles.Recent methodology works by our group and others have led to a tractable synthetic route to investigate the activity of this subclass of oxaboroles against a variety of fungal and bacterial pathogens. 22,23We hypothesize that the conformational freedom introduced from the bond rotation around the aryl and oxaborole rings could present structural features distinct from the benzoxaboroles that could be advantageous against pathogenic targets.In particular, this rotational freedom will provide another conformation to present the electrophile to a Lewis base.Herein, we report the synthesis and antimicrobial evaluation of oxaborole analogs.Our studies revealed that they have selective antifungal activity, exhibiting minimum inhibitory concen-   trations (MICs) as low as 5.20 μg/mL with no cytotoxicity against rat myoblast cells (H9c2) or hemolytic activity.
MICs were calculated using either dose-dependent solid media assays (spore forming pathogens, Figure 2B) or microbroth dilutions (nonspore forming pathogens, Figure 2C,D) with biological replicates (Table 1).The greatest activity was observed against S. cerevisiae and P. chrysogenum, with MICs as low as 5.20 and 6.25 μg/mL, respectively.Analysis of the antifungal activity showed clear SARs, where the presence of halogen was overall favorable in the meta position (6c, 6f, 6i, and 6l), less favorable in the para position (6b, 6e, and 6h), and were unfavorable in the ortho position (6d, 6g, and 6j).We hypothesize that a potential steric clash is present in the binding pocket of the target protein, or a preferred conformation is not established, because the substituent in the ortho position prevents coplanarity of the aryl and oxaborole rings.Of the halogen substitutions, iodine was the most favorable (6l) followed by bromine (6i), chlorine (6f), and then fluorine (6c), losing nearly 4-fold efficacy between 6l (MIC = 6.25) and 6c (MIC = 25) regardless of whether the halogen was in the meta or para position.However, this trend reversed for the substitution at the ortho position, further indicating the importance of coplanarity.Finally, bulky (6q−6s) and methoxy (6n−6p) substitutions to the aromatic ring were generally unfavorable regardless of position and caused nearly complete loss of growth inhibition.The control compound tavaborole was more potent compared to 3-substituted-2(5H)-oxaboroles under identical conditions, and attempts to determine if the 3-substituted-2(5H)oxaboroles shared the same fungal target as 1 were unsuccessful.
In general, very little growth inhibition was observed against either Gram-positive or Gram-negative bacterial pathogens.None of the 3-substituted oxaborole derivatives exhibited activity against B. cereus, S. aureus MRSA, or P. aeruginosa.There was modest activity against E. coli, with several compounds possessing MICs between 20 and 50 μg/mL (6b, 6e, 6h, 6k, 6s, 6t, and 6x; Figure 2D).The SAR for the antibacterial is dissimilar to that observed for the antifungal properties.Analogs with aromatic substitutions in the para position were more active than those with substitutions in the meta or ortho positions.For example, compounds 6b, 6e, and 6h are all para-substituted aryl halides and were the only aryl halides possessing activity against E. coli (MIC 21−42 μg/mL).The compounds that exhibited MICs ≤ 25 μg/mL against at least one pathogen, excluding S. cerevisiae, were subsequently evaluated for mammalian cytotoxicity and hemolytic activity.To our delight, none of the compounds (6b−6f, 6h, 6i, 6k− 6m, 6q, 6r, 6t, 6u, 6w, and 6x) had cytotoxic effects against rat myoblast cells (H9c2) and none possessed hemolytic activity at concentrations up to 100 μM (Figure 3).
In summary, a series of 3-substituted-2(5H)-oxaboroles were synthesized and evaluated for growth inhibition against a panel of microbial pathogens.Our studies revealed growth inhibition of fungi (T.mentagrophytes, P. chrysogenum, A. flavus) and yeast (S. cerevisiae) with MICs as low as 6.25 and 5.20 μg/mL, respectively.Structure−activity profiling suggests a strong preference for halogens in the meta position in fungal pathogens.Furthermore, we demonstrated that 3-substituted oxaboroles are nonhemolytic and nontoxic to mammalian cells.Current efforts are directed toward understanding the mechanism of action of these novel compounds.

Figure 3 .
Figure 3. (A) Cell viability assay against H9c2 cell line at 100 μM for 24 h (normalized to PBS no treatment), (B) Hemolytic assay at 50 and 100 μM for 1 h (normalized to triton X-100 positive control).Error bars in both graphs represent standard error of mean (SEM) of replicates (n = 3).