Investigation of Salicylanilides as Botulinum Toxin Antagonists

Abstract

Botulinum neurotoxin serotype A (BoNT/A) is recognized by the CDC as the most potent toxin and as a Tier 1 biowarfare agent. The severity and longevity of botulism stemming from BoNT/A is of significant therapeutic concern and early administration of antitoxin-antibody therapy is the only approved pharmaceutical treatment for botulism. Small molecule therapeutic strategies have targeted both the heavy chain (HC), and the light chain (LC) catalytic active site and α-/β-exosites. The LC translocation mechanism has also been studied, but an effective, non-toxic inhibitor remains under-explored. In this work, we screened a library of salicylanilides as potential translocation inhibitors. Potential leads following a primary screen were further scrutinized identifying sal30 which has a cellular IC₅₀ value of 141 nM. Inquiry of salicylanilide sal30’s mechanism of action was explored through DQ-BSA fluorescence, confocal microscopy, and V-ATPase inhibition assays. The summation of these findings imply that endo-lysosomal proton translocation through the protonophore mechanism of sal30 causes endosome pH to increase, that in turn prevents LC translocation into cytosol which requires acidic pH. Thus, inhibition of BoNT/A activity by salicylanilides likely occurs through disruption of pH-dependent endosomal LC translocation. We further probed BoNT inhibition by sal30 using additivity analysis studies with bafilomycin A1, a known BoNT/A LC translocation inhibitor, which indicated the absence of synergy between the two ionophores.

Graphical Abstract

Botulinum neurotoxins (BoNTs) are the most potent toxins known, and of the seven immunologically distinct serotypes (A-G), four (A, B, E, and F) are known to cause botulism in humans. BoNT architecture features heavy chain (HC) and light chain (LC) domains linked by a disulfide bond. The HC is responsible for neuronal cell surface receptor binding and translocation of the LC into the cell cytosol; the catalytic LC, which cleaves N-ethylmaleimide sensitive factor attachment protein receptor (SNARE) proteins inside the neuronal cell cytosol.^1-3 BoNT/A is the most toxic among these serotypes, with an estimated intravenous median lethal dose (LD₅₀) of 1-2 ng/kg in humans.¹ BoNT/A intoxication disables acetylcholine (Ach) exocytosis at neuromuscular junctions, as a result of proteolytic cleavage of synaptosomal-associate protein SNAP-25, a soluble SNARE protein.^3-5 Prevention of neurotransmission causes flaccid paralysis and eventually death by asphyxiation.^4,6 Treatment of botulism and BoNT exposure using the national stockpile of a finite amount of equine-derived antibody is expensive and risky, and involves extensive hospitalization due to toxin longevity of the LC within the neuronal compartment. Recently developed equine-derived Botulism Antitoxin Heptavalent (BAT) has been shown to reduce symptom onset (≤2 days) and resulted in shorter hospital stays.⁷ Alternatively,the use of small molecule inhibitors with the ability to prevent BoNT/A internalization can address this shortcoming.

Several types of small molecule inhibitors of BoNT have been studied, including cellular receptor binding- plant and animal based lectins⁸, and heavy chain (HC) binding- synthetic G_T1b based glycoconjugate antagonists^9,10; reversible^11-13, slow-binding¹⁴ and irreversible^15,16 (covalent) inhibitors of the LC active site; LC exosite (α and β) inhibitors^17,18; and pH-dependent translocation antagonists^9,19. In particular, bafilomycin A1 (BafA1) is known to inhibit BoNT (A-G) and tetanus toxin translocation by preventing endosomal acidification.¹⁹ Acidification of endosomes is attributed to proton transfer from the cytoplasm to the lumen through proton pumps and vacuolar H+-ATPase (V-ATPase).^20,21 BafA1 inhibits V-ATPase function, preventing LC translocation thereby delaying botulism-induced muscle paralysis.^9,19

Monensin and nigericin are also protonophores, and act as H⁺ shunts known to delay onset times of BoNT/A- and BoNT/B- treated muscles at low concentrations.¹⁹ Nigericin, a microbial toxin and a protonophore, permeabilizes K⁺ and H⁺ through the membrane, controlling the intracellular pH.^19,22 Monensin has been shown to form zwitter-ionic complexes with monovalent and divalent cations mechanizing as a transporter across the lipid membrane.^19,23-25 Niclosamide, a salicylanilide and an approved anti-helminthic with anti-neoplastic and antiviral effects, exerts its effect by neutralizing endo-lysosomal pH.²⁵ Unlike BafA1, it does not affect a protein target, and is considered a proton carrier or a protonophore. In addition to its use as anti-parasitic, Niclosamide has been investigated as a cancer therapeutic.²⁶ Salicylanilides such as rafoxamide, oxyclozanide, brotianide, and closantel are also protonophores used to treat bacterial virulence, which can be repurposed to combat other infections.^27-30 From the prominent use of pH-modulating translocation antagonists in treating bacterial virulence and toxicity³¹, it is evident that this is a viable approach to delay onset of BoNT/A-induced paralysis. To this end, we examined a previously prepared library of salicylanilides, sal1-sal60 (Figure 1, and compound list supp.), as potential BoNT/A antagonists. We note that the content of the salicylanilide library was geared to both modulate pKa and hydrophobic properties of the salicylanilide scaffold. Using such logic, we hypothesized that library member cell permeability could be improved. ³⁴ Successful mining of this library includes investigations for treatment of Clotridioides difficile infection (CDI) in a CDI mouse model, as well as reduction of SARS-CoV-2 replication in a rodent model.^32,33

Figure 1.

General structure of salicylanilide and library of halogenated salicylanilides used for screening

Similar to previously studied salicylanilides, the mechanism is hypothesized to proceed via neutralization of endosomal pH to prevent release of toxin into the cytoplasm.^25,28,35 With the history of this salicylanilide library noted, we initiated a screen for BoNT/A inhibition in enzyme and cell assays examining SNAP-25 protection and studied their mechanism of action using fluorescence spectroscopy. To complement these studies we also investigated potential synergistic effects of a selected salicylanilide when utilized with the V-ATPase inhibitor BafA1.

RESULTS

Salicylanilide screen via cell-based activity assay.

To begin this effort, salicylanilides were examined for their ability to protect SNAP-25 cleavage in hiPSC-derived GABA neurons. Salicylanilides were pooled in groups A-J (Compound list supp.) and numbered from sal1-sal60 for initial screening in our cell-based activity assay. Compound pools were screened at a concentration of 100 μM. However, at this concentration most of the pools proved to be cytotoxic, except in the case of pools G and I. Consequently, groups were reevaluated at lower concentrations of 20 μM, and 4 μM in our cellular assay (Figure S1).

BoNT/A intoxication protection by selected Salicylanilides.

Under this new protocol, pools C, D, E, F, H, and J were scrutinized for deconvolution and thus a secondary screen was initiated wherein a total thirty salicylanilides from these selected pools were singularly tested at 5 μM (Figure S2). The analysis narrowed the library members to salicylanilides sal17 (niclosamide), sal30, sal31, sal36, sal50, and sal51 (Figure 2A). The summation of this assay (Figure 2A) led us to the most potent compound sal30, with an IC₅₀ value of 141 nM (86-228 nM 95% CI). To further probe sal30’s inhibition properties a time dependent BoNT/A inhibition assay (Figure 2B) was undertaken using bafilomycin A1 (bafA1) as positive control. Here, cells were initially exposed to BoNT/A and incubated for varying amount of time before treatment with sal30 or BafA1. This assay demonstrated that the efficacy of sal30 as well as BafA1 decreases with increasing incubation times before treatment. Overall, this data suggests that salicylanilide sal30 may be acting via similar inhibition mechanism as bafilomycin A1, preventing endosomal acidification.

Figure 2.

A) Salicylanilides selected for IC50 assessment based on initial and secondary cell activity assay results. Cellular IC50 values of salicylanilides sal17, sal30, sal31, sal36, sal50, and sal51 determined by monitoring cleaved SNAP-25 over a range of concentrations. Cellular IC50 values were established through Western blot analysis and graphs were prepared in PRISM using nonlinear regression and a variable four parameter fit . B)Time dependent inhibition of BoNT/A by sal30. The graph on the right depicts densitometry data.

In-vitro inhibition evaluation of LC by salicylanilides.

Although endosomal acidification would be the most likely target of the salicylanilides, we still took recourse to examine inhibition of the LC’s enzymatic activity as a possible mechanism of BoNT/A inhibition by salicylanilides. To test this, the entire salicylanilide library was screened at a concentration of 5 μM in SNAPtide assay- a robust fluorescence resonance energy transfer (FRET) based BoNT/A-LC catalytic activity.³⁶ Interestingly, many salicylanilides presented inhibition of LC activity (Figure S3); however this was determined to be detergent dependent (Figure S4-S6), hence, non-specific aggregation and not active site LC enzyme inhibition.³⁷

Synergy analysis of salicylanilide with BafA1.

It has been reported how the salicylanilide Niclosamide, a proton ionophore exhibits a synergistic inhibitory effect when used with BafA1 against rhinovirus infection.²⁵ Considering these reports, it was logical to explore potential synergistic, additive, or antagonistic behavior that the combination of sal30 and BafA1 may exhibit on BoNT/A-mediated SNAP-25 cleavage in the hiPSC-derived GABA neuron cell model. Thus, cells were infected with BoNT/A before addition of the inhibitors at indicated concentrations (Figure S7). After a 24-hour incubation period to allow for SNAP-25 cleavage, cells were lysed and analyzed by Western blot analysis. Nonlinear fit of quantified SNAP-25 protection (Figure 3A) revealed no significant difference in SNAP-25 cleavage between treatment with individual drugs or a combination of drugs; moreover, at best the effects were additive.

Figure 3.

A) Protection of SNAP-25 cleavage in BoNT/A treated hiPSC-derived GABA neurons by sal30, BafA1, and the two in combination. Data was fit in Graphpad Prism using Nonlinear fit-inhibitor vs response (three parameter least squares ordinary fit). B) Loewe additivity analysis using Synergyfinder web application.38

Synergy assessment of bafA1 and sal30.

To quantify additivity of the two drugs, Loewe additivity analysis was performed on Synergyfinder web application.³⁸ The Loewe model assumes Loewe Additivity Consistency Conditions (LACC)- that the effect of two compounds is equivalent at different concentrations.³⁹ Loewe consensus scoring was used to assess synergy, where a score of −20 indicates highly antagonistic behavior, +20 indicates high synergy, and score close to 0 indicates additive effect, i.e. the drug combination provides the sum of the effects of individual components. Based on the synergy score of −1.55 from Loewe model, the combination of sal30 and bafA1 exhibits additive effect as no synergy is observed.

Mechanism of Inhibition.

To evaluate whether sal30 has an effect analogous to BafA1, its impact on lysosomal pH in Neuro-2A cells was studied using an assay based on DQ-BSA fluorescence. Under normal acidic conditions of endosomes, DQ-BSA substrate undergoes proteasomal degradation releasing the fluorophore 4,4-difluoro-4-bora-3a,4a-diaza-s-indacene (BODIPY).⁴⁰ Under neutral pH, degradation of DQ-BSA is reduced, resulting in decreased fluorescence. BafA1 (V-ATPase inhibitor) and monensin (protonophore), both known to disrupt the lysosomal pH gradient, were used as positive controls. After 6 hours of incubation in the presence of 500 nM inhibitor, reduced BODIPY fluorescence in samples treated with sal30 (73.9±1.6%), as well as bafA1 (62.8±2.7%) and monensin (81.6±3.9%) was observed, when compared to DQ-BSA fluorescence of DMSO control (Figure 4A). Furthermore, concentration dependence of fluorescence was observed (250 nM and 125 nM), with BafA1 having the most effect on lysosomal pH, which correlates with bafA1 being the most potent in preventing BoNT/A induced SNARE cleavage in neuronal cells (Figure 3A).

Figure 4.

A) Neuro-2A cells seeded in 96 well plates and grown overnight were treated with DMSO or inhibitor at indicated concentrations for 6 hours, followed by 1 hour treatment with DQ-BSA (10 μg/mL). After washing, fluorescence at 515 nM (ex = 495 nm) was recorded. Values and error bars indicate mean ± SD, n=3. B and C) Lysosomes in HeLa cells were prelabelled by internalization of DQ-BSA (Green) for 1 hour. Subsequently, cells were treated with DMSO or inhibitor for 6 hours and nuclei were stained with Hoechst (Blue) and observed through sequential scanning with a LSM 880 Confocal microscope using an oil immersion objective. Pixel intensity was normalized to cell count (150 cells). Values and error bars represent mean ± SD D) V-ATPase assay measuring ATPase activity through lysosomal Dx-OG fluorescence, after treatment with inhibitors, n=2.

Effect of sal30 on endo-lysosomal pH.

To further probe the effect of sal30 on endo-lysosomal pH, confocal microscopy was used to visualize sal30 treated HeLa cells by means of pH-dependent DQ-BSA green (em = 520 nm) and Hoechst (em = 450 nm) stains. As shown in figure 4B, DMSO-treated control cells showed higher lysosomal degradation of DQ-BSA, indicated by green fluorescence, whereas positive control cells treated with BafA1 and cells treated with sal30 both showed lower fluorescence of digested DQ-BSA. Quantification of fluorescence pixel intensity per cell body (average of 150 cells) for each compound treatment is depicted in Figure 4C, quantitatively corroborating our DQ-BSA assay results demonstrating that pH of the endosome is increased by sal30 similar to bafA1, and as a result decrease in DQ-BSA fluorescence is observed.

After establishing the effect of salicylanilides sal30 on lysosomal pH, a V-ATPase activity kinetic assay was used to assess whether it modulates V-ATPase function analogous to BafA1.^21,41 The crux of this assay is that it measures V-ATPase induced acidification of lysosomal fractions by measuring the pH dependent fluorescence of Dextran Oregon Green dye (Dx-OG-488, ex = 496 nm, em = 524 nm). HEK 293T cells, preincubated with Dx-OG dye, were lysed and the organelle fraction was separated. Test compounds sal30, monensin, and BafA1 were then added to the lysosomal fractions and baseline fluorescence was measured for 5 minutes. Next, V-ATPase substrates adenosine triphosphate (ATP) and magnesium chloride (MgCl₂) were added to induce lysosomal acidification, and Dx-OG fluorescence was measured continuously for 25 minutes. Finally, nigericin, which neutralizes the pH of acidified vesicles, was added to restore fluorescence. As seen in Figure 4D, the relative fluorescence of Dx-OG was consistent in BafA1 treated samples throughout and after V-ATPase substrate addition, reflecting the inhibition of V-ATPase activity by BafA1. However, sal30, and monensin treated samples showed decreased Dx-OG fluorescence upon addition of V-ATPase substrates, analogous to that seen with the DMSO negative control, indicating the lack of inhibition by these compounds. Upon neutralization of lysosomal pH with nigericin, Dx-OG fluorescence was restored to pre-activation baseline in each case.

DISCUSSION

The finite amounts, high cost, and risks associated with horse serum antitoxin-antibody treatment of botulism are of major hindrance to treating BoNT intoxication.^42-44 Potent small molecule inhibitors that are cell permeable have a good chance of entering motor neurons and thus counter the effect of toxin by preventing LC translocation. In addition to significantly lowering cost, small molecule treatment would shorten recovery time, as they can access the intracellular compartment and prevent toxin entry into neurons for a longer time period than antibodies, which cannot neutralize or clear the toxin after cell association or entry. Most of the efforts in the search for a small molecule BoNT inhibitor have been focused on active site inhibitors of the BoNT/A LC, but off-site targeting and difficulties translating in vitro to in vivo efficacy are presenting challenges still to be overcome.^13,16,17,36 Ionophores represent an alternative inhibition strategy by blunting translocation of LC into cytosol via disruption of the endosomal pH gradient required for LC release, resulting in delay of toxin on-set. BafA1 is a potent BoNT inhibitor, however its clinical use has stalled as its V-ATPase inhibitory action can alter ATP hydrolysis, and thus drug toxicity.^19,45 In fact, while bafA1 has been shown to have anticancer and antiviral properties, its toxicity profile has prevented its use for clinical applications. However, proton ionophores can have a similar endo-lysosomal pH gradient depletion effect, without actively affecting ATP hydrolysis. Proton ionophores inhibit cell entry of the BoNT enzymatic domain and thus will have no effect on the intracellular duration of LC that has already entered the neurons, thus requiring administration before (as a preventive in a high-risk situation) or shortly after exposure.⁴⁶ Since BoNT can remain in circulation for several days and during this time can still be distributed to the peripheral nervous system, an action period of several days would be desirable.³ Moreover, most toxin is taken up by the peripheral motor neurons within the first 48 hours after exposure, and thus a duration of inhibitory action for 48 hours would be expected to prevent the most severe botulism.

Salicylihalamides and salicylanilides have been studied as anthelmintic drugs mechanizing as ionophores to suppress bacterial virulence and toxicity.^25,32,33,47 Building upon these findings, a library of salicylanilides was screened in a cell-based assay to observe protection against BoNT/A-mediated SNAP-25 cleavage, an indicator of inhibitory potency. As we posited from our previous work^32,33, substituents on the anilide ring have an inductive effect on salicylanilide’s acid/base properties as well as hydrophobicity. Salicylanilides assessed in Fig. 2A, specifically sal17, sal30, and sal50 possess electron withdrawing functionality that contribute to acidity of the anilide embedded within the salicylanilide scaffold. However, because these compounds exhibit similar pKa (~5.8 phenolic proton) the improved IC₅₀ value of sal30 (141 nM); we hypithesize to be attributed to improved cell permeability, which ultimately translates to efficient shuttling of protons across the cellular membrane. It is noteworthy that we were able to repurpose sal30 for BoNT/A inhibition, as it is the same compound that showed 3.4 log reduction of viral infectivity against SARS-CoV-2 replication in rodent model in our previous study³³.

With the knowledge salicylanilides can act as ionophores our lead sal30 was further appraised and compared to BafA1 in a cellular assay probing treatment delay time on BoNT/A inhibition. Data from this study revealed that both compounds were able to similarly prevent BoNT intoxication with delayed application, thereby suggesting that sal30 may affect BoNT/A LC translocation. Based on the ability of salicylanilide protonophores such as niclosamide to synergistically inhibit rhinovirus infection²⁵, assessment of potential synergy between BafA1 and sal30 was undertaken. The synergy experiments did not conclusively delineate a combinatorial effect of the compounds. Although retesting synergy at a wide range of concentrations may yield more compelling results, the preliminary synergy experiments presented here negated the possibility of synergistic BoNT inhibition.

In vitro SNAPtide assay examination was imperative to eliminate LC inhibition mechanism before undertaking assays to investigate endosomal pH effect. The initial screen showed potential LC metalloprotease inhibition. However, such inhibition could also be the result of colloidal aggregation, a known malefactor which can cause non-specific inhibition as the compound aggregates could get absorbed by the protein resulting in enzyme denaturation.^37,48 SNAPtide assay (Figure S4) indicated that masking the phenol group of sal60 does not inhibit LC, as compared to sal23 with an unmasked phenol. Moreover, using non-ionic detergent such as Triton X-100, known to disrupt aggregates^37,49, further confirmed the non-specific inhibition as IC₅₀ values increased with increasing detergent concentrations (Figure S5).

Effect of salicylanilides on endosomal pH was explored after elimination of LC inhibition. The translocation of BoNT/A LC is a pH-dependent event that takes advantage of the endo-lysosomal pH gradient.^3,5 Because salicylanilides are known protonophores^25,27-29, we hypothesized that the distribution of this pH-dependent step was the primary mechanism of action. Thus, we compared sal30 to two molecules with known mechanisms of pH modulation: BafA1 causes an increase in lysosomal pH by inhibiting V-ATPase, whereas monensin is a protonophore that transports protons across endosomal membranes. Armed with this knowledge, we examined how these molecules impact the degradation of DQ-BSA BODIPY. All three compounds inhibited the degradation of DQ-BSA in Neuro-2A cells. We then focused our efforts upon the endo-lysosomal compartment, by directly observing and quantifying the degradation of DQ-BSA localized to endo-lysosomes using confocal microscopy. Initially, confocal microscopy was attempted using N2A cells, but due to smaller size of the cells and staining issues, puncta could not be resolved and visualized under the conditions used. Nevertheless, since the mechanistic investigation of sal30 was the purpose of this experiment, HeLa cells were examined for ease of staining and visualization of the endosomal puncta, which in turn showed the pH effect exerted by sal30 on endosomes. The DQ-BSA fluorescence assay along with our confocal microscopy labors established that sal30 causes change in lysosomal pH, similar to BafA1 and monensin.

Because protonophore activity and inhibition of ATPase both could explain the observed effect of sal30 on the lysosomal compartment, we conducted studies to distinguish these two possible mechanisms of action. Sal30, BafA1, and monensin were tested in a V-ATPase inhibition assay, which clearly showed that unlike BafA1, neither sal30, nor monensin inhibited V-ATPase activity. Therefore, we posit that sal30, which is known to act as a protonophore in Neuro-2A, HeLa and HEK293T cell models, is able to disrupt BoNT/A LC translocation by the same process. BafA1 mechanism of action upon BoNT/A is through V-ATPase inhibition, which ultimately affects ATP hydrolysis. Protonophores such as sal30 can have similar antagonistic effects on BoNT/A, however such consequences do not alter V-ATPases.

CONCLUSION

Small molecule inhibitors of BoNT/A have the potential to address longevity of BoNT induced paralysis. In this work, we introduced salicylanilides as potential BoNT/A inhibitors, and our screening efforts resulted in a potent LC translocation inhibitor sal30 with an IC₅₀ value of 141 nM. Compound sal30 is able to disrupt LC translocation from endosomes by preventing endosome acidification. Future studies will seek to optimize potency of the salicylanilide scaffold blocking of BoNT’s translocation domain; specifically efforts will be directed at the additional tuning of sal30’s acid/base and hydrophobicity properties and determine their utility as a new class of anti-botulism therapeutics that have a longer duration of action and are more easily produced.

METHODS

All salicylanilide compounds (sal1-sal60, see compound list supp.) were synthesized and purified in our laboratory previously with >95% purity by HPLC.^32,33

Enzyme Assays.

BoNT/A LC (20 μL of 25 nM, final assay concentration = 10 nM) was added in 40 mM HEPES, 0.01% Triton X-100, pH 7.4 in a black half-volume 96 well plate (Corning, cat. # 3964). Five μL of a DMSO inhibitor stock (Final assay concentration = 5 μM) was added to assay buffer, resulting in a final concentration of 1% DMSO. A negative control sample containing enzyme with no inhibitor in 1% DMSO was included. The assay mixtures were allowed to incubate at r.t. for 30 min, before addition of 25 μL of SNAPtide substrate flP6 (List Labs) to initiate the enzyme reaction, giving a final substrate concentration of 4 μM. The assay plate was immediately inserted into a Spectramax i3x plate reader (Molecular Devices), and the fluorescence was monitored in kinetic mode for 20 min at r.t., ex = 490 nm, em = 523 nm, cutoff = 495 nm. The initial rate of the enzyme activity (RFU/s) was obtained by calculating the slope of the fluorescence versus time. Initial rates of reactions containing inhibitors were then normalized to the initial rate of the no inhibitor negative control reaction to give relative enzyme activity expressed as a percentage of full activity. For IC50 determination, varying inhibitor concentrations were used and enzyme-inhibitor mixtures were incubated for 30 mins prior to the addition of SNAPtide substrate and fluorescence was measured vide supra. Data were analyzed in Excel, then Graphpad Prism 8. Dose-response curves were fitted with four parameter least-squares fit. To measure detergent dependence, IC50 was measured with varying buffer composition by altering Triton X-100 percentage).

Cell activity screening assay.

The hiPSC derived GABA Neurons and culture medium were purchased from Cellular Dynamics International (Fujifilm, CDI, R1013), and cultured in PLO-matrigel (Sigma, P4957, and BD Bioscience, 354277) coated coated 96-well TPP plates (Midsci, TP92696) for 7 days prior to the assay. For the inhibition assay, 200 LD₅₀ Units of BoNT/A (specific activity 1.7 x 10⁸ U/mg) was added to the cells in 50 μL stimulation medium (modified neurobasal containing 2.2 mM CaCl₂ and 56 mM KCl (Invitrogen, custom media) and supplemented with B27 and glutamax, both from Thermo Fisher Invitrogen, 17504044 and 35050061), and the cells were incubated at 37 °C in a humidified 5 % CO₂ atmosphere for 8 min. The toxin was removed, cells washed 2 times in 300 μL of culture medium, and the inhibitors were added immediately at 100 uM in 1 % DMSO (Sigma, D2650) (except I, which was added at 50 uM in 2.5 % DMSO) all in duplicate. Positive control (+C) was toxin without inhibitor in culture media, and negative control (−C) was culture media, both with 1 % DMSO added. Cells were incubated for 7.5 h post toxin addition at 37 °C, 5 % CO₂ to allow for SNAP-25 cleavage, and the inhibitor mixtures were aspirated and cells lysed in 50 μL of 1 x LDS lysis buffer (Thermo Fisher Invitrogen, NP0007). The samples were analyzed by Western blot using a monoclonal anti-SNAP-25 antibody (Synaptic Systems, Germany) as described previously^50,51, and bands were visualized using Phosphaglo chemiluminescent reagent (KPL) on a Azure C600 imaging system equipped with a CCD camera. Screening was performed at 4 μM and 20 μM in a similar manner.

Cell IC₅₀ determination.

Neuronal hiPSC derived cells (Fujifilm CDI, R1013) were prepared and exposed to BoNT/A as described in the cell activity screening assay. The toxin was removed, cells washed 2 times in 300 μl of culture medium (Fujifilm CDI, R1013), and the inhibitors were added immediately at the indicated concentrations in culture medium with 1 % DMSO (Sigma, D2650), all in triplicate (this was finished at ~ 18-20 min post start of toxin addition). Positive control (+C) was toxin without inhibitor in culture media, and negative control (−C) was culture media, both with 1 % DMSO added. Cells were incubated for 7.5 h post toxin addition at 37°C, 5 % CO2 to allow for SNAP-25 cleavage, and the inhibitor mixtures were aspirated and cells lysed in 50 μL of 1 x LDS lysis buffer (Invitrogen). The samples were analyzed by Western blot using a monoclonal anti-SNAP-25 antibody (Synaptic Systems, Germany, 111011) and a secondary anti-mouse antibody (KPL, 5220-0468) as described previously^50,51, and bands were visualized using Phosphaglo chemiluminescent reagent (KPL, 5430-0054) on an Azure C600 imaging system equipped with a CCD camera. Before harvest, cells were observed by light microscopy. No signs of cytotoxicity were observed in any wells.

Time-dependent Inhibition Assay.

Similar cell culture and intoxication protocols were used as the cell activity screening assay. The toxin was removed, cells washed two times in 300 μL of culture medium, and the inhibitors were added at 5μM (sal30) or 2 μM (bafilomycin A1) in 1 % DMSO at the indicated times, all in triplicate. Positive control (+C) was toxin without inhibitor in culture media, and negative control (−C) was culture media, both with 1 % DMSO added. Cells were incubated for 24 h post first toxin addition at 37 °C, 5 % CO₂ to allow for SNAP-25 cleavage, and the inhibitor mixtures were aspirated and cells lysed in 50 μL of 1 x LDS lysis buffer (Invitrogen). The samples were analyzed by Western blot using a monoclonal anti-SNAP-25 antibody (Synaptic Systems, Germany) as described previously^50,51, and bands were visualized using Phosphaglo chemiluminescent reagent (KPL) on an Azure C600 imaging system equipped with a CCD camera. No signs of cytotoxicity were observed in any wells. The percent protection was determined by densitometry of the Western blot data. The percentage of uncleaved versus cleaved SNAP-25 was measured for each sample, in triplicate, and the positive control sample (toxin only with no inhibitor added) was set to 0% protection and detection of only uncleaved SNAP-25 was considered 100% protection.

Synergy assessment.

Similar cell culture and intoxication protocols were used as the cell activity screening assay. The toxin was removed, cells washed 2 times in 300 μL of culture medium, and the inhibitors were added at at the indicated concentrations with a final concentration of 1 % DMSO in culture media, all in triplicate. At the time of inhibitor addition, 30 min had elapsed post toxin addition due to time required for toxin exposure and washing of the cells. Positive control (+C) was toxin without inhibitor in culture media, and negative control (−C) was culture media, both with 1 % DMSO added. Cells were incubated for 24 h post toxin addition at 37 °C, 5 % CO₂ to allow for SNAP-25 cleavage, and the inhibitor mixtures were aspirated and cells lysed in 50 μL of 1 x LDS lysis buffer (Invitrogen). The samples were analyzed by Western blot using a monoclonal anti-SNAP-25 antibody (Synaptic Systems, Germany) as described previously^50,51, and bands were visualized using Phosphaglo chemiluminescent reagent (KPL) on an Azure C600 imaging system equipped with a CCD camera. Data was analyzed in GraphPad Prism6. Synergy analysis was performed using SynergyFinder web application tool by inputting the SNAP-25 protection response data of drug combinations in CSV format.³⁸

DQ-BSA Fluorescence Assay.

Neuro-2A (cholinergic murine neuroblastoma) purchased from ATCC (CCL-131) were grown in Eagle’s minimum essential medium(EMEM-gibco-11090) with Earl’s salts containing 2 mM glutamine, 1.5 g/L sodium bicarbonate, 1 mM sodium pyruvate, 0.1 mM non-essential amino acids, and 10% fetal bovine serum. Cells were seeded overnight in clear bottom 96-well plates (Corning, 3904) at 2E04 cells per well in 100 μL media. The following day, 50 μL compound stocks of sal30, monensin, bafA1, and DMSO control in EMEM (500 nM, 250 nM, 125 nM) were added to designated wells in triplicate, and incubated at 37 °C for 6 hours. After the treatment, cells were incubated with DQTM Green BSA (Invitrogen, D12050) at 10 μg/mL for 1 hour at 37 °C. Cells were washed with phenol red free media three times and 50 μL of fresh phenol red free media was added. The fluorescence at 515 nm was measured with 495 nm excitation. Assay was performed in triplicate.

DQ-BSA fluorescence visualization using confocal microscopy.

HeLa cells purchased from ATCC (CCL-2) were cultured in Eagle’s minimum essential medium (EMEM, 30-2003) containing 10% FBS. Cells from a confluent 75 cm² culture flask, were seeded in 35 mm microscopy dishes at 10⁵ cells in 2 mL media overnight. The following day, media was removed and replaced with culture media containing DQTM Green BSA (Invitrogen, D12050) at 10 μg/mL and cells were incubated for 1 hour at 37 °C. Media containing dye was removed and cells were washed with PBS followed by treatment with compound stocks of sal30 (500 nM), bafA1(100 nM), and DMSO(1%) control in culture media, and incubation for 6 hours. During the last 30 minutes, Hoechst 33342 was added to the final concentration of 5 μg/mL. After incubation period, media was aspirated and cells were washed with PBS before imaging, and cells were imaged using Zeiss LSM 880 Airyscan confocal laser scanning microscope with a 63x oil immersion lens using DAPI and Alexa fluor 488 filters to visualize nucleus (450 nm) and lysosome (520 nm) respectively. Image analysis was done using Fiji opensource tool.⁵² DQ-BSA BODIPY fluorophore pixel intensity of each cell was calculated after auto-threshold, and data was normalized and plotted in GraphPad Prism.

V-ATPase activity assay.

V-ATPase activity assay was performed as indicated by Aldrich et al. with minor modifications.⁴¹ HEK293T/17 purchased from ATCC (CRL-11268) were cultured in Dubelco’s minimum essential medium (Gibco, 10566) with 10% fetal bovine serum and 1% penicillin-streptomycin. From a confluent 75 cm² dish, cells were incubated with Dx-OG488 (35 μg/mL) overnight. The following day, cells were washed with PBS and incubated with FBS free media for 2 hours. During the last 15 minutes, carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP) was added to a final concentration of 1 μM. Cells were pelleted and resuspended in 2 x 750 μL fractionation buffer (composed of 50 mM KCl, 90 mM K-gluconate, 1 mM EGTA, 50 mM sucrose, 5 mM glucose, 20 mM HEPES, mini protease inhibitor cocktail-pH 7.4) containing 1 μM FCCP. Cells were broken by spraying through a 25 G needle, and spun down at 10,000 rpm for 20 seconds at 4 °C. Supernatant containing organelles was collected and centrifuged at 13,200 rpm for 20 minutes. Pellet containing the organelles was resuspended in pre-warmed fractionation buffer containing 1% bovine serum albumin. Organelle suspension was split into 15 aliquots and added to 96 well plate (Corning, 3904). Compound pretreatment (with 100 nM BafA1/500 nM monensin/500 nM sal30) was applied to designated wells (in triplicate), and incubated at 37 °C for 1 hour. Baseline fluorescence was measured at emission intensity-524 nm upon excitation at 496 nm at 15 second intervals for 5 minutes. V-ATPase was activated by addition of 5 mM ATP and 5 mM MgCl₂, and 524 nm fluorescence was measured for 25 minutes at 15 second intervals. Lysosomal proton gradient was dissipated by addition of 1 μg/mL nigericin at the end of the measurement.

Abbreviations

ATP

adenosine triphosphare

BafA1

bafilomycin A1

BoNT/A

botulinum neurotoxin serotype A

BSA

bovine serum albumin

CDC

Centers for Disease Control and Prevention

DCHA

2,4-dichlorocinnamic hydroxamic acid

Dx-OG

dextran oregon green

FRET

fluorescence resonance enrgy transfer

HC

heavy chain

LC

light chain

SNAP-25

synaptosomal-associated protein of 25 kDa

SNARE

soluble N-ethylmaleimide-sensitive factor attachment receptor