Occidiofungin, an actin binding antifungal with in vivo efficacy in a vulvovaginal candidiasis infection

Current antifungal treatment options are plagued with rapidly increasing occurrence of resistance, high degree of toxicity and a limited spectrum of activity. The need to develop a novel antifungal with a unique target, wider spectrum of activity, and reduced toxicity to the host, is urgent. We have identified and characterized one such compound named occidiofungin that is produced by the soil bacterium Burkholderia contaminans MS14. This study identifies the primary cellular target of the antifungal, which was determined to be actin. Actin binding metabolites are generally characterized by their ability to inhibit polymerization or depolymerization of actin filaments, which presumably accounts for their severe toxicity. Occidiofungin, instead, has a subtler effect on actin dynamics that triggers apoptotic cell death. We were able to demonstrate the effectiveness of the antifungal in treating a vulvovaginal yeast infection in a murine model. This discovery puts occidiofungin in a unique class of actin-binding antifungal compounds with minimal reported toxicity to the host. The results of this study are important for the development of a novel class of antifungals that could fill the existing gap in treatment options for fungal infections. Author summary Widespread resistance to antifungal compounds currently in use has been alarming. Identification and development of a new class of antifungals with a novel cellular target is desperately needed. This study describes the assays carried out to determine the molecular target and evaluate efficacy of one such novel antifungal compound called occidiofungin. Occidiofungin modified with a functional alkyne group enabled affinity purification assays and localization studies in yeast. These studies led to the identification of the actin binding property of occidiofungin. Actin-binding by secondary metabolites often exhibit severe host toxicity, but this does not appear to be the case for occidiofungin. We have previously been able to administer occidiofungin to mice at concentrations in the range of 5 mg/kg without any serious complications. We were able to demonstrate the effectiveness of the antifungal in treating a vaginal fungal infection in a murine model. The results outlined in this manuscript establish that occidiofungin is an efficacious compound with a novel molecular target, putting it in a completely new class of antifungals.


Introduction
Alkyne derivatization of occidiofungin. 194 In order to localize occidiofungin in yeast and to identify its cellular binding partners, 195 methods to fluorescently label or add a functional purification tag to occidiofungin were needed. 196 To this end, occidiofungin was chemically modified with a terminal alkyne through acylation of 197 the free amino group of the diaminobutyric acid residue at position 5 (S1 Figure) Table). To determine whether alkyne-OF still had the same apoptosis inducing bioactivity as 204 the native occidiofungin, S. cerevisiae was treated with alkyne-OF and apoptotic assays such as 205 TUNEL, reactive oxygen species (ROS) detection, and phosphatidylserine externalization assays 206 were performed. Double stranded DNA breaks, the generation of ROS, and the externalization of 207 phosphatidylserine were observed in the alkyne-OF treated cells, supporting the same mechanism of action (S4 Figure A-C). Although this alkyne modification moderately reduced the inhibitory 209 activity of the compound, the functionalized derivative has the same apoptotic bioactivity and was 210 therefore used to identify the fungal target.

212
Identification of occidiofungin interacting proteins. 213 Alkyne-OF was used in a pull-down assay to identify intracellular proteins that directly or 214 indirectly interact with occidiofungin ( Fig 4A). In brief, alkyne-OF was incubated with S.  Table). The culled protein list was grouped based on gene ontology 227 including cellular localization and/or molecular function. The resulting distribution is presented in  In vitro analysis of the interaction of occidiofungin with purified actin. 245 Typical assays for characterizing actin binding natural products are the in vitro F-actin 246 polymerization and depolymerization experiments. However, the addition of occidiofungin was 247 found to have no effect on the polymerization or depolymerization properties of F-actin (S5 248 Figure). Therefore, additional studies were required to confirm that actin was the biological target 249 for occidiofungin. Biotinylation of alkyne-OF following incubation with F-or G-actin and 250 streptavidin agarose beads was performed to determine whether occidiofungin directly associated 251 with purified actin in vitro. F-or G-actin incubated with the wild type occidiofungin or DMSO 252 were used as controls for potential non-specific interaction of actin with the agarose beads. The 253 eluant from the biotinylated alkyne-OF had a single band at approximately 42 kDa, the expected 254 size for actin. As shown in Fig 5A, the biotinylation of alkyne-OF was required for the co-255 purification of F-or G-actin with the streptavidin beads (Lane 5 and 8) as actin was not present in the control lanes that exposed actin to native OF or the carrier solvent DMSO (lanes 6, 7, 9, and 10). In this in vitro interaction assay, occidiofungin was shown to directly bind to F-or G-actin. 258 To further support this observation, a dissociation constant of 1.0 ± 0.8 μM was determined from 259 three independent ITC experiments using rabbit skeletal muscle G-actin (Fig 6). The ITC data also 260 showed a 1:1 binding ratio for occidiofungin to G-actin. The ITC experiments were not adaptable 261 to observe F-actin binding, so a co-sedimentation assay, which is commonly reported for 262 identifying actin associated proteins, was performed [51, 52]. Phalloidin was used as a positive 263 control in the assay (Fig 7). Phalloidin had an estimated dissociation constant (Kd) of 8 nM with  Figure). Similar to that shown using labeled phalloidin, F-actin also appeared to aggregate 278 following exposure to occidiofungin. Fluorescence visualization of this interaction following 279 treatment with alkyne-OF and native occidiofungin demonstrated a high degree of aggregation of the filaments which was not observed in the controls. Occidiofungin does not prevent 281 polymerization or depolymerization of F-actin, but it does bind to F-actin causing it to aggregate.

282
Given these unique observations for the bioactivity of occidiofungin, additional in vivo studies 283 were conducted to verify that actin is its biological target. where S is the maximal X bound, Kd is the dissociation constant and X is the concentration of free 306 ligand.

308
In vivo analysis of occidiofungin interactions with actin. 309 In vivo visualization of the localization of occidiofungin was done in intact yeast cells.    (Fig 9A and 9B). Actin cables are formed by bundling F-actin. In Figure 8, the 339 punctate structures in these cells are still likely filamentous actin, but occidiofungin appears to 340 disrupt the organization of F-actin to form cables at sub-inhibitory concentrations.

349
Six to eight-week-old BALB/c mice that were intravaginally infected with C. albicans were 350 dosed once per day with occidiofungin for three days. The occidiofungin treated groups were 351 compared to a vehicle control group. Three groups of six mice were treated with 100, 50, and 0 µg 352 of occidiofungin suspended in 0.3% Noble agar. The occidiofungin treated groups reduced fungal 353 load by more than two logs (Fig 10). The reduction in fungal load with both treatment groups was 354 statistically significant from vehicle control (p<0.001). There was no statistically significant 355 difference between the treated groups (p=0.33), suggesting that the lower limit of occidiofungin 356 dosing was not achieved in the experiment. During the course of the study, the mice were examined 357 for outward signs of distress or irritation. No behavioral changes including sluggishness, 358 stretching, or reluctance to consume food was observed. Furthermore, no vaginal bleeding or 359 swelling was observed following treatment.

379
It is likely that occidiofungin undergoes self-assembly, forming micellar structures at the 380 concentrations tested in our assay. Additionally, cellular processes that rely on the maintenance of 381 the actin cytoskeleton such as endocytosis, hyphae formation, and nuclear DNA positioning were 382 shown to be disrupted with the addition of occidiofungin. We also demonstrate that occidiofungin 383 is capable of treating a murine vulvovaginal candidiasis infection without any signs of toxicity.

384
Furthermore, occidiofungin demonstrated efficacy at a concentration that is ten-fold lower than 385 azole-based treatment methods [61].

386
Given that we have observed fungicidal activity as low as 100 nM concentrations, the 387 number of bound occidiofungin to F-actin is likely to be a lot lower than 24:1 for its fungicidal activity. We hypothesize that occidiofungin binding to F-actin interferes with the binding of other 389 actin associated proteins leading to disruption of cellular activities involving actin dynamics.

390
Cables are necessary for a multitude of cellular functions including hyphal formation (disrupted 391 following treatment as seen in Figure 2), endocytosis (reduced following treatment as seen in 392 Figure 3) and proper positioning of the mitotic spindle during cell division (accumulation of 393 multinucleated cells following treatment as seen in S2 Table). Studies aimed at understanding the 394 events following the binding of occidiofungin to actin will need to be conducted to determine the by the low MICs against several different types of fungi. Susceptibility to occidiofungin can be 408 seen in pathogenic strains that are resistant to treatment with azoles and echinocandins (S1 Table). 409 Additionally, occidiofungin has the advantage of inducing cell death in fungi via a mechanism that   in the presence of the dye for 60 minutes at 30˚C, followed by two washes with PBS and then added to a microscope slide for visualization. Images were obtained using an Olympus FV1000 526 confocal microscope with a 40x/0.9 dry objective. 527 Derivatization of occidiofungin. 528 Occidiofungin was purified from a liquid culture of Burkholderia contaminans MS14 as