key: cord-1001957-tl0azwa7
authors: Raut, Prakash; Weller, Sasha R.; Obeng, Bright; Soos, Brandy L.; West, Bailey E.; Potts, Christian M.; Sangroula, Suraj; Kinney, Marissa S.; Burnell, John E.; King, Benjamin L.; Gosse, Julie A.; Hess, Samuel T.
title: Cetylpyridinium chloride (CPC) reduces zebrafish mortality from influenza infection: Super-resolution microscopy reveals CPC interference with multiple protein interactions with phosphatidylinositol 4,5-bisphosphate in immune function
date: 2022-02-09
journal: Toxicol Appl Pharmacol
DOI: 10.1016/j.taap.2022.115913
sha: 87cf653aa13d6ee277f62e217b0034a0e5b5723e
doc_id: 1001957
cord_uid: tl0azwa7

The COVID-19 pandemic raises significance for a potential influenza therapeutic compound, cetylpyridinium chloride (CPC), which has been extensively used in personal care products as a positively-charged quaternary ammonium antibacterial agent. CPC is currently in clinical trials to assess its effects on severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) morbidity. Two published studies have provided mouse and human data indicating that CPC may alleviate influenza infection, and here we show that CPC (0.1 μM, 1 h) reduces zebrafish mortality and viral load following influenza infection. However, CPC mechanisms of action upon viral-host cell interaction are currently unknown. We have utilized super-resolution fluorescence photoactivation localization microscopy to probe the mode of CPC action. Reduction in density of influenza viral protein hemagglutinin (HA) clusters is known to reduce influenza infectivity: here, we show that CPC (at non-cytotoxic doses, 5–10 μM) reduces HA density and number of HA molecules per cluster within the plasma membrane of NIH-3 T3 mouse fibroblasts. HA is known to colocalize with the negatively-charged mammalian lipid phosphatidylinositol 4,5-bisphosphate (PIP(2)); here, we show that nanoscale co-localization of HA with the PIP(2)-binding Pleckstrin homology (PH) reporter in the plasma membrane is diminished by CPC. CPC also dramatically displaces the PIP(2)-binding protein myristoylated alanine-rich C-kinase substrate (MARCKS) from the plasma membrane of rat RBL-2H3 mast cells; this disruption of PIP(2) is correlated with inhibition of mast cell degranulation. Together, these findings offer a PIP(2)-focused mechanism underlying CPC disruption of influenza and suggest potential pharmacological use of this drug as an influenza therapeutic to reduce global deaths from viral disease.

24,000 µM hinders microbial membrane lipid integrity by electrostatic attraction and charge neutralization between oppositely charged components of the microbe and CPC resulting in aggregate formation (Yegin et al., 2019) . Therefore, metabolic and biosynthetic processes within bacteria may be perturbed at concentrations both lower and higher than the CMC. Thus, there are both alternative mechanisms of antibacterial action (vs. simple detergent solubilization above the CMC) and evidence of CPC binding to specific bacterial lipids. While cholesterol protects cell membranes from solubilization by quats, leading to greater resistance to quats in eukaryotic cells compared to bacterial cells (Watanabe and Regen, 1994; Marcotte et al., 2005) , there is otherwise a dearth of published literature regarding information on CPC interactions with eukaryotic cell lipids.

In addition to its antibacterial action, CPC also has a demonstrated potential as an antiviral agent (Popkin et al., 2017; Seo et al., 2019; Baker et al., 2020; Koch-Heier et al., 2021) .

To date, most studies exploring antiviral activity of CPC either have not included mechanistic information or have focused on direct interactions between CPC and the viral particle. To our knowledge, information on CPC effects on virus interaction with eukaryotic host cells is lacking.

CPC was shown to be an effective inhibitor of Hepatitis B (HBV). The direct mechanism behind CPC's HBV inhibition ability involves its prevention of capsid assembly via interaction with the viral nucleocapsid protein, which ultimately leads to reduced HBV biogenesis (Seo et al., 2019) .

At 1500-2100 µM, CPC destroys SARS-CoV-2 viruses within tens of seconds (Koch-Heier et al., 2021) . CPC is believed to inhibit enveloped viruses in general by disrupting viral membranes (Baker et al., 2020) . In particular, CPC possesses virucidal activity against several strains of influenza virus, even at doses below its CMC (~15-60 μM, with exposure times of a few minutes), apparently by disrupting the morphology of the viral envelope (Popkin et al., 2017) .

The authors suggested that CPC may target the mammalian host cell-derived lipids within the viral envelope, such as phosphatidylethanolamine.

Furthermore, CPC administered orally to mice ameliorated influenza infection and reduced morbidity in mice (Popkin et al., 2017) . Annually, in the U.S. alone, there are typically 12,000 -61,000 mortalities due to influenza (CDC, 2020) . Thus, there is a current global need for novel forms of drug treatment (Antara, 2018) in order to combat annual seasonal cases of influenza, including those refractory to the annual vaccine. Intriguingly, CPC was tested in a human clinical trial to assess its effectiveness against upper respiratory infections . Very recently, CPC has been the subject of high-profile news reports for its use in multiple clinical trials against SARS-CoV-2. For example, one human clinical trial assessed the use of CPC to reduce oral load of SARS-CoV-2 in patients preparing to undergo dental procedures; the researchers found that CPC mouthwash decreased levels of SARS-CoV-2 in patients' saliva following a 5-minute rinse, as compared to a 5-minute water rinse (Seneviratne et al., 2021) . Thus far, studies have indicated both that CPC may directly destroy viruses and that it may be an effective pharmacological antiviral within animals and humans.

However, the question remains--does CPC act as an antiviral by disrupting eukaryotic host cell/virus interactions? At low micromolar doses far below the CMC, CPC inhibited fusion between the SARS-CoV-2 viral membrane and HEK-293T human kidney cells (Muñoz-There are no published studies regarding CPC effects on immune cell function (Pubmed); in the current study, we explore the effect of CPC on the function of mast cells. Mast cell signaling and function hinge upon the key role of phosphatidylinositol 4,5 bisphosphate (PIP 2 ), a plasma membrane signaling lipid. Mast cells, found throughout human tissues, defend the body against bacterial (Johnzon et al., 2016), viral (Dawicki and Marshall, 2007) , and parasitic (Metcalfe et al., 1997) infections as well as playing key roles in allergy, asthma (Galli et al., 2005) , and neurological function (Theoharides et al., 2016) . Mast cells undergo an allergen/antigen (Ag)-stimulated process called degranulation, in which they release bioactive substances including histamine and serotonin. Mast cell degranulation begins when a multivalent Ag binds to and mediates cross-linking of IgE-bound FcεRI receptors leading to a phosphorylation cascade in which spleen tyrosine kinase (Syk) and phospholipase C gamma (PLCγ) are activated respectively (Kinet, 1999) . The activated PLCγ cleaves phosphatidylinositol 4,5 bisphosphate (PIP 2 ) to generate diacylglycerol (DAG) and inositol 1,4,5-trisphosphate (IP 3 ). The IP 3 binds to its receptor on the endoplasmic reticulum (ER) membrane, triggering an efflux of Ca 2+ from the ER internal stores into the cytosol (Berridge, 1993) . The subsequent activated influx of Ca 2+ into the cytosol is termed store-operated calcium entry (SOCE) (Putney, 1986) . SOCE is a core mediator of the mast cell degranulation pathway and necessary for downstream events such as protein kinase C (PKC) activation (Ozawa et al., 1993) , that lead to degranulation. PIP 2 -dependent SOCE is also key for T Cell function (Punt et al., 2019; Trebak and Kinet, 2019) . In this study, we examine the effects of CPC on degranulation, using the RBL-2H3 mast cell type, which react to environmental stimuli in a fashion analogous to the biochemical responses of primary bone marrow-derived mast cells (Zaitsu et al., 2007; Thrasher et al., 2013; Alsaleh et al., 2016) . In this study, we examine CPC Super-resolution imaging was necessary for the work in this study because PIP 2 and HA clusters are typically too small to image using diffraction-limited microscopy (Curthoys et al., 2019) . Fluorescence photoactivation localization microscopy (FPALM; (Hess et al., 2006) ) was used in order to visualize these molecular interactions on the nanoscale. FPALM can detect molecular distributions in cells and other biological systems at a resolution of ~10-20 nm (Hess et al., 2006; Hess et al., 2009) . FPALM uses repeated cycles of activation, localization, and photobleaching in conjunction with high sensitivity fluorescence imaging, to identify and localize numerous molecules within a sample (Hess et al., 2006) , and is compatible with live-cell (Hess et al., 2007) , 3-dimensional (Juette et al., 2008) and multicolor (Gunewardene et al., 2011) .

In this study, we hypothesize that positively-charged CPC biochemically targets the negatively-charged eukaryotic lipid PIP 2 and, therefore, interferes with PIP 2 -protein interactions and function at ~1,000-fold lower concentrations than those found in personal care products. In order to address these questions, this study will assess CPC effects on the nanoscale distribution, colocalization, clustering, and density of three PIP 2 -binding proteins (MARCKS, HA, and PH) in the PM of eukaryotic cells. Consequent to CPC disruption of these PIP 2 -protein interactions, functional effects on mast cells and to influenza infection in zebrafish are also examined. In this study, we are promulgating a mechanism by which CPC affects the mammalian cell itself, to reduce influenza morbidity. We are also extending the earlier finding of CPC reduction of influenza morbidity in mice (Popkin et al., 2017) to the zebrafish model system, which has been used previously to study influenza infection (Gabor et al., 2014) . We hypothesize that CPC is not simply a virucidal agent; rather, at low-and sub -micromolar doses, CPC is affecting the host cell-virus interaction by interfering with the phosphoinositide PIP 2 and with PIP 2 -HA J o u r n a l P r e -p r o o f interactions. These findings reveal CPC as a potential therapeutic to treat PIP 2 -related diseases like influenza and also may modulate other PIP 2 -dependent eukaryotic cell functions.

Cetylpyridinium chloride (CPC; 99% purity, VWR; CAS no. 123-03-5) was prepared at 150 µM in a pre-warmed Tyrodes buffer (recipe given in (Hutchinson et al., 2011) ) and vortexed. Next, the preparation was sonicated (Branson 1200 ultrasonic cleaner; Branson Ultrasonics, Danbury, CT, USA) at 37 °C for 20 minutes, protected from light, then vortexed again. The solution was then poured into a sterile Erlenmeyer flask for continual stirring until usage, protected from light. A control Tyrodes buffer solution was prepared in tandem.

Following dilution in pre-warmed Tyrodes buffer, exact concentrations were determined using UV-Vis spectrophotometry (Weatherly et al., 2013) and the Beer-Lambert equation (A 260 =ε 260 ℓc), using an ε 260 of 4389 M -1 cm -1 (Bernauer et al., 2015) . Bovine serum albumin (BSA) was added to create a final solution of CPC in BT (Tyrodes buffer containing BSA (Hutchinson et al., 2011) ). For all cell culture experiments, CPC was administered via BT. For zebrafish experiments, CPC was administered via egg water (60 μg/ml Instant Ocean sea salts in autoclaved Milli-Q water, pH 7.0; Aquarium Systems, Mentor, OH) (with no BSA), and sonication and UV-Vis were performed similarly as above. Fresh CPC stocks were prepared on each experimental day.

J o u r n a l P r e -p r o o f RBL-2H3 mast cells were cultured as described (Hutchinson et al., 2011) .

NIH-3T3 mouse fibroblast cells (ATCC, CRL-1658) were cultured as described (Curthoys et al., 2019) .

MDCK London cells (passage 3) were grown at 37 °C with 5% CO2 in T-175 flasks in minimal essential medium (MEM), containing final percentages/concentrations of the following: 5% newborn calf serum (NCS), 2% heat-inactivated fetal bovine serum (FBS), 0.23% sodium bicarbonate solution, 2% MEM amino acids (from 50X stock), 1% MEM vitamins (from 100X stock), 4 mM L-glutamine, and 1% Antibiotic-Antimycotic (from 100X stock). All media and reagents for MDCK culture were from ThermoFisher Scientific unless otherwise noted.

Cells were maintained through washing twice with 1x Dulbecco's phosphate-buffered saline (dPBS, pH 7.4) to remove traces of NCS and FBS, trypsinizing with 0.25% trypsin-EDTA with phenol red, and passaging in a 1:6 dilution every 2-3 days. Virus-infected cells were grown in MEM-BSA/TPCK: this media is similar to the MEM media described above but supplemented with Tosyl phenylalanyl chloromethyl ketone (TPCK) trypsin (Worthington Chemical Corporation) and with Bovine Albumin Fraction V (7.5% solution) instead of NCS and FBS.

RBL-2H3 mast cells were transiently transfected with monomeric red fluorescent protein (mRFP)-MARCKS-ED plasmid (a gift from Dr. Barbara Baird and Dr. David Holowka, Cornell University; "MARCKS", (Gadi et al., 2011) ) via electroporation using Amaxa Nucleofactor kit T (Lonza) as described in (Weatherly et al., 2018) . After electroporation, cells were plated at 1.7 J o u r n a l P r e -p r o o f x 10 5 cells per well (in an eight-well ibidi plate) in a phenol red-free media and incubated overnight at 5% CO 2 / 37 ℃. The next day, the spent media was removed, and cells were incubated with 200 µl of 0, 5, or 10 µM CPC in BT for 30 minutes. Following this incubation, cells were washed in BT, 200 µl BT was added to each well, and images were taken immediately using confocal microscopy. See "Confocal Microscopy" below for imaging details.

For RBL-2H3 cells transiently transfected with monomeric red fluorescent protein (mRFP)-MARCKS-ED plasmid, an Olympus FV-1000 confocal microscope with an Olympus IX-81 inverted microscope, with a 1-mW HeNe-Green laser (543 nm excitation and 560-660 nm emission filter) were used to collect images. An oil immersion 100 x objective with NA 1.4 and image acquisition speed of 2 μs/pixel were used to collect the images. Using the ibidi plate heating system, imaging was conducted at 37 ℃.

Fiji ImageJ software (NIH) was used to analyze confocal microscopy images of RBL-2H3 cells transfected with monomeric red fluorescent protein (mRFP)-MARCKS-ED. Images were converted to 8-bit images, channels split, and further analysis done using the fluorescent channels. Pseudo flatfield correction was used to subtract background, appropriate threshold was applied, and binary masks generated for both the whole cell and the cytoplasm of a given transfected cell. The masks of the whole cell and of the cytoplasm were applied to a given transfected cell in the fluorescent channel to generate regions of interest (ROI) for the whole cell and the cytoplasm, respectively. Next, the area, integrated density, and mean fluorescence per pixel of the whole cell and the cytoplasm were measured from the ROI. The mean fluorescence intensity per pixel of the PM values were generated by 1.) subtracting the integrated density of the cytoplasm from the integrated density of the whole cell, 2.) subtracting the area of the cytoplasm from the area of the whole cell, and 3.) dividing the result of 1.) by the result of 2.). The PM/cytoplasm mean fluorescence per pixel were calculated by dividing the mean fluorescence of the PM by the mean fluorescence of cytoplasm, and results analyzed in GraphPad Prism.

Degranulation response was measured in RBL-2H3 mast cells as in (Weatherly et al., 2013 ). This assay indicates the level of degranulation occurring by measuring the release of enzyme β-hexosaminidase into the surrounding environment. The β-hexosaminidase substrate utilized for detection was 4-methyllumbelliferyl-N-acetyl-β-D-glucosaminide (4-MU), a synthetic substrate that releases a fluorescent product upon enzymatic cleavage. We validated that there is no CPC effect on β-hexosaminidase's ability to bind to its substrate, 4-MU ( Figure   S1 ). Treatment groups included IgE-sensitized, antigen (DNP-BSA; Ag)-stimulated treatment cells, to be exposed to CPC dilutions; spontaneous cells, an indicator of the basal amount of degranulation that normally occurs; high-control cells, cells exposed to Triton-X 100 detergent to lyse all cells and maximize potential β-hexosaminidase release; and no-cell background wells, wells that received no cells that can be used to subtract out background fluorescence. The data CPC Cytotoxicity and Survivability on RBL-2H3 cells Trypan blue exclusion and lactate dehydrogenase (LDH) cytotoxicity assays were used to assess cytotoxicity and survivability on RBL-2H3 cells exposed to CPC. The trypan blue exclusion assay was performed exactly as in Hutchinson 2011 (Hutchinson et al., 2011 , except with overnight plating of 1.2 x 10 6 cells per well of six-well Grenier CELLSTAR plates and with CPC exposure instead of arsenic.

Cytotoxicity was also assessed based upon LDH release from cells (into the supernatant), using an absorbance-based cytotoxicity detection kit per instructions (Sigma), as described in Hutchinson 2011 (Hutchinson et al., 2011 . We confirmed that CPC does not interfere with the LDH assay ( Figure S2 ).

For sample preparation, cells were seeded in growth media (DMEM with Glucose without phenol red, Lonza) with 10% Calf Bovine Serum (ATCC) and antibiotics (Penicillin Streptomycin, 100 ug/ml) at 35,000 cells/ml overnight in 35 mm petri dishes (MatTek). After 20-24 hours, cells were transfected using Lipofectamine 3000 (Invitrogen) with 2 µg of total DNA (1µg of HA-Dendra2 (Gudheti et al., 2013) and 1µg of PAmKate-PH (PLC-δ) (Curthoys et al., J o u r n a l P r e -p r o o f 2019)) and treated with Control (0 µm CPC in BT vehicle), 5µM CPC in BT, or 10 µM CPC in BT for 1 hour in the incubator under same conditions. After one hour, cells were washed twice with PBS (Sigma-Aldrich), fixed at room temperature in 4% paraformaldehyde (PFA) (Alfa Aesar) for 10-15 minutes, and then washed again with PBS two times.

Two color FPALM was performed by previously published methods (Hess et al., 2006; Gunewardene et al., 2011; Curthoys et al., 2019; Sangroula et al., 2020) . Briefly, lasers with wavelengths λ = 405 nm (CrystaLaser, 5mW) for the activation of Dendra2 molecule (Gurskaya et al., 2006) and λ = 558 nm for the readout (CrystaLaser, 100 mW) were focused in the backaperture plane of an oil immersion objective (Olympus 60X 1.45NA) using a second lens (f = 350 mm, Thorlabs) at one focal length away from the back aperture. To match the polarization of the readout laser, the activation laser was passed through a half wave plate (Newport, 10RP42-1) and a linear polarizer (Newport, 5511). For better activation and readout, both lasers were then converted to elliptical (approximately circular) polarization by passing them through a quarter wave plate (Newport, 10RP54-1B). Laser power was recorded by the power meter (Thor labs) to be ~13 mW for the readout laser and about ~75 µW for the activation laser, respectively, after passing straight through the objective lens. The position of the beam in the objective back aperture was then translated to allow total internal reflection fluorescence (TIRF) imaging.

Fluorescence emission was collected by the objective and filtered through a quad-band dichroic (Di01 R405/488/561/635-25x36), and by 405 nm and 561 nm notch filters (Semrock: NF03-561E-25). After the dichroic and notch filter, fluorescence emitted through the tube lens was magnified ~2X using successive achromatic lenses with focal lengths +20 mm and +40 mm.

Fluorescence then reached a second dichroic (Semrock, FF580-FDi02-t3) within the multi-color detection module, which reflected λ < 595 nm and transmitted λ > 595 nm, thus producing two J o u r n a l P r e -p r o o f wavelength ranges which were simultaneously imaged onto adjacent regions of the camera sensor. Fluorescence from the transmitted channel passed through an ET630/92 filter (Semrock, FF01-630/69-25) and fluorescence from the reflected channel passed through a 585/40 filter (Semrock, FF01-585/40-25) before reaching the camera (iXon+ DU897DCS-BV, Andor Scientific, Dublin, Ireland). Typically, ten thousand frames were recorded at ~32 Hz and EM gain of 200.

Point spread functions recorded in the raw images were background subtracted (Sternberg, 1983) and localized by fitting into the 2D Gaussian function (Hess et al., 2006) .

After localization, images were further analyzed for drift subtraction and bleed through correction . Localizations were assigned to either of two channels according to their alpha values (Gunewardene et al., 2011; and further processed through custom built Matlab code for the quantitative analysis of clusters.

We used single linkage cluster analysis (SLCA) (Greenfield et al., 2009; Gudheti et al., 2013; Curthoys et al., 2019; Sangroula et al., 2020; Sneath, 1957) for cluster identification. This technique detects molecules within a maximum distance d max of each other and assigns them to the same cluster. Clusters are analyzed further if they contain at least a minimum total number of molecules. For this analysis, d max = 35 nm. To be considered for analysis, a threshold for the minimum number of molecules per cluster was required to be N min ≥ 50.

The mean density of HA molecules and the mean density of PH molecules was obtained computationally. The process involves binning the molecules into a grid of squares 35 nm X 35 nm and counting the localizations of each species within each bin to create a density map. The 

Microinjections of influenza were performed as in (Gabor et al., 2014) , with the following modifications. Agarose gel (2%) was utilized in 100-mm petri dishes coated with 3% methylcellulose. After anesthetizing the embryos with sodium bicarbonate-buffered MS-222 (200 mg/L), a volume of 1.0 nl (~ 8.7 ×10 2 EID50) APR8 IAV or HBSS (Hank's Balanced Salt Solution, Gibco) was microinjected. The control zebrafish were sibling zebrafish injected simultaneously with treated zebrafish. Needles containing IAV were changed every hour to J o u r n a l P r e -p r o o f guarantee the virus remained viable. Following injection, 45 zebrafish were sorted into plates of 50 mL egg water, which was changed daily. Pulled microcapillary pipettes (1.0 mm outside diameter, 0.6 mm inside diameter, Sutter Instruments, Novato, CA) were used to inject either the virus or HBSS (Goody et al., 2014; Sullivan et al., 2017) .

CPC was administered to influenza infected and control, non-infected zebrafish at a concentration of 0.1 µM in 50 mL of egg water. CPC doses significantly higher than 0.1 µM caused embryo mortality (data not shown). The selection of the infected zebrafish to receive CPC treatment was random. The fish were treated with CPC for 1 hour at 33 °C under lightsensitive conditions (placed in a dark incubator). The fish were then rinsed twice in egg water to remove all traces of CPC and placed into fresh egg water for survivability experiments. Standard CPC treatment occurred at 6 hpi. Specialized treatments were administered 24, 48, and 72 hpi.

Each timed CPC experiment was compared to CPC-treated zebrafish and to untreated zebrafish, for both IAV-and HBSS-injected zebrafish ( Figure 10 ).

The burden of influenza virus in IAV-injected zebrafish was quantified using a TCID 50 assay. Wild-type zebrafish embryos (150 per experimental group) were injected and maintained as described in "Microinjection of V." Water was changed daily, and 25 fish from each group were collected, starting from 0 to 96 hpi. Embryos were placed into a kill tricaine solution (200 µg/mL) for 5 minutes to ensure death before being placed in a sterile microcentrifuge tube. The J o u r n a l P r e -p r o o f residual tricaine solution was removed and replaced with 500 L of RNAlater (ThermoFisher Scientific) and flash frozen with liquid nitrogen. Samples were stored at −80 °C.

MDCK-London cells were plated in 96-well plates (USA Scientific) the night before the TCID 50 assay, with eight wells for samples and with one well at the end of each row dedicated to the control, at a cell density designed to achieve 90% confluency on the day of the experiment.

Each sample well was plated in triplicate to allow for three tests of the serial dilutions to confirm the viral concentration. In addition, cells were not allowed to grow past passage 10. Different passages and aliquots were utilized according to the cells' difference in behavior, such that for experiment 1 there were 17,000 cells per well, experiment 2 had 13,000 cells per well, and experiment 3 was 14,500 cells per well. On the day of the experiment, cells were washed twice with 1x dPBS, pH 7.4, while the virus-containing samples were prepared as follows.

The RNAlater samples were thawed at room temperature, then RNAlater was removed manually with a Pasteur pipette and replaced with 1 mL of MEM-BSA/TPCK. A metal bead was added, and the samples were homogenized with a bullet blender at setting 3 for 5 minutes at 4 °C. The samples were briefly centrifuged at 8000 X g for 1 minute. The samples were placed on ice and moved to the BSL-2 hood. Serial dilutions of the samples were prepared, 1:10 until 10 -7 . The second dPBS wash was removed from the MDCK-London cells, and 50 µL of each sample dilution was added to corresponding plates in triplicate. Controls also received 50 µL of MEM-BSA/TPCK. Plates were centrifuged at room temperature for 5 minutes, 2000 X g before being placed in the 37 °C 5% CO 2 incubator for 90 minutes. The samples were tapped every 15 minutes to ensure the cells did not dry out. Sample media was removed from the MDCK-London cells and 105 µL of MEM-BSA/TPCK was added to each well. Plates were grown at 37 °C with 5% CO2 for 3 days before being counted. Cells were visually assessed for cell damage J o u r n a l P r e -p r o o f such as death, morphological changes, and clearance; images were recorded and results were confirmed by a second researcher. The TCID 50 was calculated by the Spearman-Kärber method (Hierholzer and Killington, 1996) .

Statistical significance was determined by one-way ANOVA with a Tukey's post-test for both the Trypan blue and LDH assays. Significance testing of FPALM images involved one-way ANOVA followed by Dunnett's post-test performed using GraphPad Prism using data imported from the analyzed Matlab files. Standard curve comparison, a triplicate statistical analysis (Mantel-Cox, Log-rank, and Gehan-Breslow-Wilcoxon) in GraphPad Prism was used to determine significance in zebrafish survival studies. Two-way ANOVA with post Spearman's correlation and normality tests of TCID 50 assays in GraphPad Prism was performed to measure significance.

To test our hypothesis that CPC biochemically targets the negatively-charged eukaryotic lipid PIP 2 and, therefore, interferes with PIP 2 -protein interactions, we assessed CPC effects on the cellular localization of the PIP 2 -binding protein MARCKS. RBL-2H3 mast cells were transiently transfected with monomeric red fluorescent protein (mRFP)-MARCKS-ED (Gadi et al., 2011) and incubated overnight. On the next day, the cells were treated with Control buffer (0 µM CPC), 5 µM CPC, or 10 µM CPC for 30 minutes at 37 o C, and live-cell images were J o u r n a l P r e -p r o o f collected with confocal microscopy (Figure 2A ). The ratio of mean fluorescence intensity per pixel of (mRFP)-MARCKS-ED at the PM to the mean fluorescence intensity per pixel of (mRFP)-MARCKS-ED in the cytoplasm were calculated using Fiji ImageJ, and the ratio was calculated: "PM/cytosol mean fluorescence per pixel." This ratio dose-responsively decreases with increasing CPC treatment ( Figure 2B ): 4.9 ± 0.2 for Control (0 µM CPC) samples, 2.4 ± 0.1 for 5 µM CPC samples, and 1.8 ± 0.1 for cells exposed to 10 µM CPC. Comparison of this ratio in CPC-treated and Control cells shows that CPC statistically significantly displaces the PIP 2 -binding protein MARCKS from the PM, into the cytosol (Figures 2A and 2B ).

To test the same hypothesis, in an alternative system (different cell line, PIP 2 -binding protein, and microscopy method), we assessed CPC effects on the overall mean density at the plasma membrane of the PIP 2 -binding protein PAmKate-PH. NIH-3T3 cells were transfected with PAmKate-PH and another putative PIP 2 -binding protein HA-Dendra2 (Curthoys et al., 2019). The next day, super-resolution imaging of the PM of the transfected cells was carried out with TIRF illumination to focus on the plasma membrane, excluding the cytoplasm. Cells were treated with Control buffer (0 µM CPC), 5 µM CPC, or 10 µM CPC for 1 hr at 37 o C, cells were fixed, and images were collected with FPALM. The mean density of PAmKate-PH ("PH") molecules was quantified by taking the ratio of the total number of PH molecules (PH localizations) over the cell area imaged in the same acquisition (following drift correction and bleedthrough correction), and plotted as a function of CPC treatment ( Figure 2C ). The mean density of PH molecules (# PH/m 2 ) was observed to be 700 ± 100 /µm 2 for the Control (0 µM CPC) samples, 510 ± 50/µm 2 for 5 µM CPC samples, and 470 ± 40/µm 2 for cells exposed to 10 µM CPC. Comparison of the density of PH (# PH/m 2 ) in CPC-treated and Control cells shows J o u r n a l P r e -p r o o f that CPC significantly reduces the density of the PIP 2 -binding protein PAmKate-PH present adjacent to the PM ( Figure 2C ).

Degranulation in mast cells is the culmination of a PIP 2 -dependent signaling process (Santos et al., 2013) . Thus, to test our hypothesis that CPC biochemically targets PIP 2 and thus interferes with PIP 2 -dependent cellular functioning, degranulation was measured following CPC exposure using the fluorescence assay described in (Weatherly et al., 2013) , adapted for the use of CPC. In this assay, cells were sensitized with anti-dinitrophenyl (DNP) mouse IgE, to occupy surface IgE receptors, which were then crosslinked with multivalent DNP-BSA antigen (Ag) to initiate cellular signaling. The dose of Ag utilized, 0.0004 μg/mL, was chosen because it elicited a moderate absolute degranulation response of 33% ± 1% (SEM) of the maximal possible granule release when in the absence of CPC. By choosing a moderate Ag dose, potential stimulatory or inhibitory CPC effects can be observed. Figure 3A displays the effects of CPC on degranulation of IgE-sensitized RBL-2H3 mast cells incubated for 1 h in BT containing 0.0004 μg/mL Ag. CPC significantly inhibits degranulation in a dose-responsive manner, beginning with ~25% inhibition at 1 µM ( Figure   3A ). Statistically significant inhibition of degranulation began at 1 μM CPC after 1 hour exposure, resulting in a degranulation response that was 0.7-fold ± 0.1 (SEM) of the 0 μM CPC Control level ( Figure 3A ). At 5 μM CPC, degranulation was reduced to 0.47-fold ± 0.07 (SEM) of the 0 μM CPC Control response. At 10 μM CPC, the response was reduced to 0.30-fold ± 0.05 (SEM) of the 0 μM CPC Control, equivalent to a 70% inhibition of this cellular function. We

Journal Pre-proof confirmed that this inhibition was a true cellular effect, not due to interference of CPC with the reaction of -hexosaminidase with its fluorogenic substrate, which is the signal used to quantify degranulation; also, CPC does not directly affect fluorescence of the background signal ( Figure   S1 ). These data show that CPC inhibits mast cell degranulation in a dose-responsive fashion.

Two cytotoxicity assays were conducted to assess the survivability of RBL-2H3 cells when exposed to a range of CPC doses for 1 hour. To assess the survivability of RBL-2H3 cells upon exposure to various CPC doses, we performed a lactate dehydrogenase (LDH) release assay. LDH is an intracellular enzyme released upon plasma membrane damage or death. LDH release was not significantly altered, compared to 0 μM CPC Control, for exposures up to 15 M CPC for 1 hour, indicating a lack of cytotoxicity or plasma membrane damage ( Figure 3B ); however, 30 μM CPC for 1 hour caused LDH release. Activity of purified LDH was unaffected by CPC, indicating that CPC does not interfere with the detection mechanism of the LDH assay itself ( Figure S2 ).

A trypan blue-exclusion assay was used to assay metabolic viability of RBL-2H3 mast cells upon 1 hour CPC exposure. Figure 3C shows a statistically significant cytotoxic effect only in cells exposed to 30 μM CPC, while concentrations ≤ 15 μM were not cytotoxic as quantified by the assay. Furthermore, in NIH-3T3 cells, there was a lack of CPC cytotoxicity at the doses tested (5 and 10 μM) in a trypan blue assay ( Figure S3 ).

Taken together, the results from these two assays strongly suggest that CPC is not cytotoxic to RBL-2H3 and NIH-3T3 cells when exposed for 1 hour at concentrations ≤ 15 μM.

These data suggest that cell death is not the mechanism for the effects of CPC documented in this manuscript (such as degranulation and re-distribution of PIP 2 -binding proteins).

Considering the relationship between PIP 2 and the influenza viral protein hemagglutinin (HA) (Curthoys et al., 2019) , the requirement of HA clustering for influenza infectivity (Ellens et al., 1990; Takeda et al., 2003) , the reported antiviral properties of CPC (Popkin et al., 2017) , and the disruption of PIP 2 -binding protein MARCKS by CPC, we employed super resolution microscopy to test whether CPC affects nanoscale HA clustering. We used FPALM (Hess et al., 2006) with TIRF excitation to image the PM of NIH-3T3 cells expressing both HA-Dendra2 and PAmKate-PH Domain, which binds to and labels PIP 2 (Gambhir et al., 2004; Curthoys et al., 2019) . Clusters were identified using SLCA (see methods). Quantification of HA clusters showed a wide range of density with the mean value of 12700 ± 4300 HA/µm 2 for control ( Figure 4A ). This wide variation of density in HA cluster was absent in cells treated with CPC: the mean HA density in CPC-treated cells decreased significantly by ~7 9% when compared against the control, with the mean HA density of 2700 ± 400 HA/µm 2 for 5 µM and 2700 ± 400 HA/µm 2 for 10 µM CPC-treated cells ( Figure 4A) . Similarly, the number of HA molecules in a cluster was reduced significantly by CPC. On average, 340 ± 110 HA molecules were found in an HA cluster which decreased significantly by ~7 4% and ~7 6% to 88 ± 9 HA molecules and 81 ± 13 HA molecules for 5 µM and 10 µM CPC-treated cells, respectively ( Figure 4B ). The mean area of an HA cluster was observed to be 0.041 ± 0.004 µm 2 for Control, 0.044 ± 0.006 µm 2 for 5 µM, and 0.039 ± 0.003 µm 2 for 10 µM CPC ( Figure 4C ). The mean perimeter of an HA cluster was observed to be 0.90 ± 0.08 µm for control 1.0 ± 0.1 µm for 5 µM, and 0.09 ± 0.06 J o u r n a l P r e -p r o o f µm for 10 µM CPC ( Figure 4D ). Both the mean area and mean perimeter of an HA cluster were statistically unaffected by CPC treatment. The mean density of HA molecules (within each cell) was quantified by dividing the total number of HA molecules (HA localizations) in a given cell over the imaged area of each cell. The mean density was 195 ± 8 HA /µm 2 in the Control, 190 ± 12 HA/µm 2 for 5 µM, and 180 ± 12 HA/µm 2 for 10 µM CPC treatment: no statistically significant CPC effect was observed.

We also studied the effect of CPC in PIP 2 clusters along with the HA, as PIP 2 has also been previously reported to cluster at the PM (Curthoys et al., 2019) (van den Bogaart et al., 2011; Wang and Richards, 2012) . In order to visualize the nanoscale distribution of PIP 2 , we used FPALM to image a PAmKate-tagged version of the PIP 2 -binding Pleckstrin Homology (PH-) domain from PLC-δ (Gambhir et al., 2004; Curthoys et al., 2019) , called PAmKate-PH and abbreviated as "PH domain" (Figure 5 ). Cluster analysis was performed using SLCA (see Methods). The mean density of a PH domain cluster was 2700 ± 500 /µm 2 for Control, 2400 ± 300 /µm 2 for 5 µM, and 2100 ± 100 /µm 2 for 10 µM CPC treated cells ( Figure 5A ); no statistically significant CPC effect was detected on mean density. However, the mean number of PAmKate-PH molecules forming a cluster decreased significantly (p < 0.01). On average, 270 ± 30 molecules formed a PH domain cluster (Control) while only 170 ± 13 and 190 ± 21 PH domain molecules on average were found in a cluster for 5 µM and 10 µM CPC-treated cells, respectively ( Figure 5B ). The mean area of a PH domain cluster (Control) was found to be 0.105 ± 0.009 µm 2 , which appeared to decrease to 0.08 ± 0.01 µm 2 and 0.085 ± 0.006 µm 2 for 5 µM CPC and 10 µM CPC-treated cells, respectively ( Figure 5C ); however, no statistical significance J o u r n a l P r e -p r o o f was observed by one-way ANOVA followed by Dunnett's multiple comparison test against Control, and the Kruskal-Wallis test was barely insignificant (p = 0.0503). The mean perimeter of PH domain clusters was statistically unchanged by CPC: 1.7 ± 0.1 µm for Control, 1.4 ± 0.1 µm for 5 µM, and 1.53 ± 0.08 µm for 10 µM CPC ( Figure 5D ).

CPC dramatically reduced colocalization of HA-Dendra2 and PAmKate-PH in NIH-3T3 cells ( Figure 6 ). In order to quantify the effect of CPC on HA and PH domain co-clustering at the nanoscale, the mean number of HA molecules per grid pixel within PH clusters (N HA-PH ; see Methods) and the mean number of PH molecules within HA clusters (N PH-HA ) were determined from two-color FPALM results. The analysis shows a significant decrease in N PH-HA in 10 µM CPC-treated cells, decreasing by approximately 71% as compared to control, which was statistically significant ( Figure 7A) . A similar result was obtained for the number of HA localizations within PH domain clusters, where the Mean Pixel Sum N HA-PH dropped approximately by 61% compared to Control, again statistically significant ( Figure 7B ).

Viral load and mortality assessment were performed ± CPC exposure at 0.1 M for 1 hour in Danio rerio (zebrafish, ZF), an in vivo influenza virus infection (IAV) model. In order to determine appropriate dose and timing of CPC exposure, first, acute (1 hour) exposures of 2-4 dpf ZF embryos to CPC alone (no IAV injection) were performed (Figure 8 ). All doses 1 M and above were 100% lethal to 2-4 dpf ZF embryos within 3 days post CPC exposure, while 0.5 M and 0.25 M CPC were lethal to a fraction of animals. In contrast, 99.5% of 200 exposed ZF embryos survived for at least 3 days following a 1 hour, 0.1 M CPC exposure. Next, longer J o u r n a l P r e -p r o o f exposures to this chosen 0.1 M CPC dose were performed, but lethality (15% of embryos) began within 3 hours of continuous CPC exposure, and mortality increased with length of CPC exposure (43% lethal at 6 hours of 0.1M CPC, 96% lethal at 12 hours, 100% lethal at 24 hours) (these data are not plotted). Thus, exposure to CPC at 0.1 M for 1 hour exposure was chosen for anti-influenza pharmacological studies.

Specifically, 1 hour of 0.1 M CPC treatment in mild to moderate IAV infections was shown to significantly increase the chance of survival if administered within the first 48 hpi ( Figure 9 ). For zebrafish exposed for 1 hour to 0.1 M CPC at 6 hpi, survival was analyzed at 5 dpi: average survival of CPC-treated zebrafish was 83.2%, versus 60.1% for CPC-untreated, IAV-infected zebrafish ( Figure 9A ). For zebrafish exposed for 1 hour to 0.1 M CPC at 24 hpi, survival was analyzed at 5 dpi: average survival of CPC-treated zebrafish was 87.1%, versus 60.1% for CPC-untreated, IAV-infected zebrafish ( Figure 9B ). For zebrafish exposed for 1 hour to 0.1 M CPC at 48 hpi, survival was analyzed at 5 dpi: the average survival of CPC-treated zebrafish was 87.7%, versus 60.1% for CPC-untreated, IAV-infected zebrafish ( Figure 9C ). Importantly, neither CPC nor mock HBSS injection causes mortality within 5 dpi ( Figure S4 ).

If the one-hour CPC treatment is withheld until 72 hpi, CPC is unable to rescue the embryos from influenza mortality ( Figure S5 ). This CPC reduction in influenza mortality correlates to CPC reduction in viral burden despite a similar starting viral load (Figure 10 ). Figure 10 shows that there is no difference in viral burden (TCID 50 /mL) at 0 hpi for IAV-infected zebrafish with or without the CPC treatment (1 hour, 0.1 M). However, starting at 24 hpi, and especially significant at 48 hpi, there is a decrease in viral burden by as much as 1 log-fold due to CPC treatment ( Figure 10) . Although it appears that viral burden drops, even in non-CPC treated fish, at the later time points (Figure 10 ), J o u r n a l P r e -p r o o f the reason is due to a combination of zebrafish mortality (fish that die due to influenza at earlier time points are not assessed at 96 hours), viral clearance, and removal of fish from the experiment.

Cetylpyridinium chloride possesses antiviral activity, including against influenza, in human , mouse (Popkin et al., 2017) , and, now, zebrafish with a short (1 hour), low-dose (0.1 M) exposure ( Figure 9 ). However, its anti-influenza mode of action, particularly at low-micromolar doses relevant to safe potential pharmacological usage, is unknown. At the CPC doses (≤ 10 M) tested in zebrafish and cells in this study, CPC is far below its CMC (Mandal and Nair, 2002; Varade et al., 2005; Abezgauz et al., 2010; Shi et al., 2011) and, thus, is not acting as a detergent (confirmed by lack of cytotoxicity, Figure 3B -C and lack of zebrafish mortality, Figure S4 ). Also, these low-micromolar doses are well below levels of CPC known to act as direct virucidal agents (Popkin et al., 2017; Baker et al., 2020; Koch-Heier et al., 2021) . Thus, because CPC is not acting via a detergent lysis or direct virucidal mechanism to alleviate zebrafish viral load and mortality following influenza infection, another anti-infective mechanism must be involved. In this study, the role for low-micromolar CPC effects on mammalian host cells and on interactions between the host cell and viral components was explored. We hypothesized that CPC affects the mammalian cell itself, to reduce influenza morbidity. We found that CPC disrupts interactions between PIP 2 and multiple PIP 2 -binding proteins, including MARCKS, PH-domain from PLC, and influenza HA, suggesting a PIP 2focused mechanism underlying CPC disruption of influenza infectivity.

One of the central molecular players in mast cell signaling is PIP 2 . Thus, we hypothesized that, via its positive charge at the end of a lipidic tail, CPC biochemically targets the negatively-charged eukaryotic lipid PIP 2 and, therefore, interferes with PIP 2 -protein interactions that are essential to mast cell function. These proteins could be MARCKS, PLC , or other PIP 2 -binding proteins that play crucial roles in mast cell function. Thus, we performed experiments to test this CPC mode-of-action hypothesis and found that CPC treatment caused MARCKS (in mast cells) and PH molecules (in NIH-3T3 cells), two well-known PIP 2 binding proteins (McLaughlin et al., 2002) , to fall off the PM (Figure 2 ).

Concurrently, we discovered that CPC inhibits degranulation of antigen-stimulated mast cells in a dose-responsive manner, beginning at 1 µM ( Figure 3A) . These data represent one of the first studies of CPC effects on eukaryotic cells, despite its widespread use in personal care products. These findings indicate that further research into the effects of CPC on immune cell function are necessary and relevant to consumer health. The doses tested are non-cytotoxic, as indicated by the LDH and trypan blue cytotoxicity assays ( Figure 3B-C) ; thus, the CPC effect on mast cells is not due to cell death. Thus, CPC is likely interfering with one or more components of the antigen-stimulated mast cell signaling cascade, resulting in its suppressive effect on degranulation. Further research will be necessary to pinpoint the mechanism of CPC disruption of mast cell degranulation, which may be an immunotoxic side effect of CPC. Research is also needed to determine the long-term effects of CPC, including potential toxicity, in a variety of cell types and in mammalian model systems.

Our results suggest that other PIP 2 -dependent physiological processes, including influenza infection, may be altered by CPC. Because of a recently-identified relationship between HA and PIP 2 in which PIP 2 modulates HA clustering, HA modulates PIP 2 clustering, J o u r n a l P r e -p r o o f and both molecules are co-localized at the nanoscale in the PM (Curthoys et al., 2019) , we hypothesized that CPC may also affect the plasma membrane distribution of HA, via PIP 2 interference. Based on our previously published results (Curthoys et al., 2019) , we hypothesized that HA interacts with PIP 2 through charge-charge interactions between the negatively-charged PIP 2 head and the HA cytoplasmic tail (CT), which has two basic residues and an estimated net charge of +6 per trimer at physiological pH, and also has several conserved, acylated cysteines within the HA CT which could interact with lipids within the inner leaflet of the PM (Veit et al., 2013) . This information led us to hypothesize that CPC blocking of this HA-PIP 2 interaction would affect HA clustering, which is crucial for infectivity (Ellens et al., 1990) and the viral lifecycle (Takeda et al., 2003) . We used super-resolution microscopy (FPALM) to image HA-Dendra2 and the PIP 2 -binding PH domain tagged with PAmKate (Flesch et al., 2005; Szentpetery et al., 2009; Curthoys et al., 2019) (Figure 6 ). Localized molecular coordinates of HA-Dendra2 and PAmKate-PH were then quantified using SLCA (Curthoys et al., 2019; Sangroula et al., 2020; Sneath, 1957) to determine the properties of clusters formed by the HA and PH domain.

The density of HA in clusters in control cells varied widely ( Figure 4A ), consistent with published results showing that HA clusters exist on many length scales (Hess et al., 2005) .

Significant disruption of HA clusters by CPC was observed, particularly with respect to the density and the number of HA molecules forming each HA cluster in CPC-treated cells ( Figure   4A -B). Since HA is a transmembrane protein, CPC is not expected to cause HA to dissociate from the PM as we observe with PH and MARCKS proteins (Figure 2 ), but rather is disrupting the clusters through redistribution of the HA which is present in the PM. We speculate that this disruption could be due to CPC interactions with PIP 2 or through interference with the J o u r n a l P r e -p r o o f interaction between the HA CTD and the PIP 2 headgroup. These data are important because dense HA clusters are crucial for efficient viral entry through membrane fusion (Takeda et al., 2003) , and CPC significantly reduces HA cluster density, thus illuminating a potential mechanism of antiviral properties of CPC at low concentrations.

This mechanism is also consistent with our findings that CPC treatment increases survival ( Figure 9 ) and reduces viral load ( Figure 10 ) in influenza-infected zebrafish. There is a minimal difference in effect of CPC treatments at 5 and 10 M on HA clustering (Figure 4) , suggesting that 5 M CPC may be a sufficient cellular dose for maximizing anti-influenza CPC activity. At 72 hpi, the infection is too severe for CPC treatment to make a difference ( Figure   S5 ). CPC does seem to rescue older fish better than younger fish (data not shown), indicating possible CPC disruption of developmental processes. The zebrafish embryos were injected with 1.0 nl (~ 8.7 ×10 2 EID50) of APR8 IAV and selected at random for CPC treatment using 0.1 µM for 1 hour (standard CPC treatment at 6 hpi and specialized treatments at 24, 48, and 72 hpi). In contrast, in the Popkin study, wildtype mice were infected intranasally (LD50, 8.0 x 10 3 pfu/mouse) with the mouse-adapted influenza strain A/PuertoRico/8/1934 H1N1 (PR8, ATCC) while CPC was given orally (5 μL, in formulation) 15 minutes prior and then twice a day for 5 consecutive days (Popkin et al., 2017) . While the influenza and CPC exposures were different for this zebrafish study and the published mouse study, both found that CPC improves survivability following influenza infection. PIP 2 is known to cluster at the PM (van den Bogaart et al., 2011; Wang and Richards, 2012; Curthoys et al., 2019) , and the interactions of PIP 2 with proteins control a large number of cellular functions (McLaughlin et al., 2002; McLaughlin and Murray, 2005; Di Paolo and De Camilli, 2006; Balla, 2013; O'Donnell, 2018) . Quantification of PAmKate-PH (PIP 2 ) clusters J o u r n a l P r e -p r o o f visualized by FPALM showed significant disruption by CPC ( Figure 5 ), thus further confirming the effect of CPC on PIP 2 -binding proteins. Results also uncovered a major disruption of HA-Dendra2 co-localization with PAmKate-PH (Figures 6 and 7) , further confirming that CPC can disrupt interactions between PIP 2 and proteins at the PM. These results are important because PIP 2 -binding proteins have been previously shown to have polybasic domain and clusters to negatively-charged PIP 2 and PIP 3 through charge-charge interaction (Heo et al., 2006) , and CPC disrupts the clustering of several of these proteins. The CPC-induced reduction of the mean number of PH domain molecules forming a cluster also explains the observation that the only PH cluster property that is affected is the mean number of PH molecules per cluster, which would be due to CPC ejecting or screening some of the PH-domain-containing proteins close to the plasma membrane.

In addition to interactions with HA, PIP 2 plays numerous roles in modulating cytoskeletal function and membrane biochemistry that may be involved in influenza pathogenesis (Raucher et al., 2000; Di Paolo and De Camilli, 2006; Guerriero et al., 2006; Heo et al., 2006; Altan-Bonnet and Balla, 2012) . PIP 2 regulates cytoskeleton-PM adhesion (Raucher et al., 2000) , promotes the production of actin filaments within the cell (Janmey et al., 2018) , and is capable of removing capping proteins like CapZ from actin filaments, to amplify growth (Schafer et al., 1996; Toker, 1998) . Furthermore, PIP 2 is able to interact with several actin-binding proteins such as geloslin (Janmey and Stossel, 1987) , cofilin (Yonezawa et al., 1990) and Ezrin/Radixin/Moesin (ERM) proteins (Hirao et al., 1996) , with ERM proteins working to crosslink the PM and actin filaments (Arpin et al., 2011) . Several actin-binding proteins found in purified influenza virus (Shaw et al., 2008) have known PIP 2 -interactions (Catimel et al., 2008; Catimel et al., 2009) . The relationship between PIP 2 and actin filaments further implicates PIP 2 as having a role in J o u r n a l P r e -p r o o f additional cellular processes like chemotaxis and cell migration (Wu et al., 2014) . PIP 2 also plays a role in allowing the adapter protein AP-2 to associate with clathrin-coated pits (Gaidarov and Keen, 1999) . Any of these interactions could potentially be affected by CPC, both in the context of viral infection and normal cell function, and should be examined in future research.

A plethora of studies in different viruses show that phosphoinositides, and in many cases PIP 2 , interact with viral proteins through multiple basic residues, and that these interactions are crucial for the life cycles of multiple viruses (Ono et al., 2004; Chukkapalli et al., 2008; Chukkapalli and Ono, 2011; Altan-Bonnet and Balla, 2012; Johnson et al., 2016; Yandrapalli et al., 2016; GC et al., 2017; Budicini et al., 2018) . More recently, phosphoinositide kinase inhibitors, which affect cellular PIP 2 levels, have been shown to inhibit Zaire Ebola virus and SARS-CoV-2 (Kang et al., 2020) . Phosphoinositide kinase inhibitors also inhibit mast cell degranulation (Santos et al., 2013) . Targeting interactions between viral proteins and host cell phosphoinositides could be a novel therapeutic approach; CPC is already being used in a clinical trial against the SARS-CoV-2. Our study presents the effect of CPC on PIP 2 -binding proteins and illuminates a mechanism for its antiviral properties at relatively non-cytotoxic concentrations, paving the path for the future study of the effect of CPC in modulating PIP 2 and PIP 2 -binding proteins and other membrane proteins that interact with the PIP 2 .

In future work, we aim to further deeply explore the CPC, PIP 2, and HA interactions in multiple cell types including polarized epithelial cells. Future virology research will be necessary to pinpoint the exact mechanism of action of CPC anti-influenza action within the viral lifecycle.

In conclusion, here we report that low-dose CPC acts as an effective anti-influenza agent in zebrafish. The underlying mechanism of action is likely related to the ability of CPC to disrupt J o u r n a l P r e -p r o o f Journal Pre-proof HA cluster density, the number of HA molecules per cluster, and the nanoscale co-localization of HA with the PIP 2 reporter PAmKate-PH, which we infer is due to CPC interference with PIP 2 .

In fact, CPC disrupts multiple (MARCKS, PH, and HA) PIP 2 -dependent protein interactions and subcellular localizations. Super-resolution microscopy makes possible these mechanistic investigations.

CPC also interferes with PIP 2 -dependent cellular functions (mast cell degranulation) and organismal interaction with influenza (as evidenced by CPC reduction of zebrafish mortality and viral burden). In summary, CPC targets the key eukaryotic signaling lipid PIP 2 and reduces influenza morbidity. While other factors could also play a role in vivo, our data suggest that CPC modulation of phosphoinositide-dependent host cell-viral protein interactions is likely a part of the mechanism of viral inhibition. Zebrafish embryos injected with HBSS and exposed to various doses of CPC were monitored for cumulative survival over the course of 5 days (no observable death occurred after 3 days).

Survival curve values presented are for a total of 204 zebrafish, assayed via four independent experiments. Statistically significant differences are represented by ****p < 0.0001, determined by the Mantel-Cox test, the Logrank test, and the Gehan-Breslow-Wilcoxon test of survival curves in GraphPad Prism. 

Effect of Hofmeister anions on micellization and micellar growth of the surfactant cetylpyridinium chloride

Silver Nanoparticle-Directed Mast Cell Degranulation Is Mediated through Calcium and PI3K Signaling Independent of the High Affinity IgE Receptor

Phosphatidylinositol 4-kinases: hostages harnessed to build panviral replication platforms

Cetylpyridinium chloride blocks herpes simplex virus replication in gingival fibroblasts

Molecular evolution of hemagglutinin gene of Influenza A virus

Emerging role for ERM proteins in cell adhesion and migration

Interaction of Cetylpyridinium Chloride with Giant Lipid Vesicles

Repurposing Quaternary Ammonium Compounds as Potential Treatments for COVID-19

Phosphoinositides: tiny lipids with giant impact on cell regulation

Opinion on cetylpyridinium chloridesubmission II (P97)

Inositol trisphosphate and calcium signalling

A comparison between chlorhexidine and some quaternary ammonium compounds with regard to retention, salivary concentration and plaqueinhibiting effect in the human mouth after mouth rinses

Investigating Phospholipid Binding Residues in the C Terminus of Ebola Virus Matrix Protein, VP40

The PI(3,5)P2 and PI(4,5)P2 Interactomes

PI(3,4,5)P3 Interactome

Disease Burden of Influenza

Interaction between the human immunodeficiency virus type 1 Gag matrix domain and phosphatidylinositol-(4,5)-bisphosphate is essential for efficient gag membrane binding

Molecular determinants that regulate plasma membrane association of HIV-1 Gag

Sequestration of phosphoinositides by mutated MARCKS effector domain inhibits stimulated Ca(2+) mobilization and degranulation in mast cells

Phosphoinositide-AP-2 interactions required for targeting plasma membrane clathrin-coated pits

Electrostatic Sequestration of PIP2 on Phospholipid Membranes by Basic/Aromatic Regions of Proteins

Membrane association and localization dynamics of the Ebola virus matrix protein VP40

Cationic antiseptics: diversity of action under a common epithet

Studying the immune response to human viral infections using zebrafish

Cations induce shape remodeling of negatively charged phospholipid membranes

In vitro assessment of the virucidal activity of four mouthwashes containing Cetylpyridinium Chloride, ethanol

Self-organization of the Escherichia coli chemotaxis network imaged with superresolution light microscopy

Actin mediates the nanoscale membrane organization of the clustered membrane protein influenza hemagglutinin

Phosphatidylinositol 5-kinase stimulates apical biosynthetic delivery via an Arp2/3-dependent mechanism

Superresolution imaging of multiple fluorescent proteins with highly overlapping emission spectra in living cells

Does PtdIns(4,5)P2 concentrate so it can multi-task?

Structural Characterization of the Interaction between a

Lipopolysaccharide and lipoteichoic acid

PI(3,4,5)P3 and PI(4,5)P2 lipids target proteins with polybasic clusters to the plasma membrane

Ultra-high resolution imaging by fluorescence photoactivation localization microscopy

Dynamic clustered distribution of hemagglutinin resolved at 40 nm in living cell membranes discriminates between raft theories

Ultrahigh resolution imaging of biomolecules by fluorescence photoactivation localization microscopy

Quantitative electron microscopy and fluorescence spectroscopy of the membrane distribution of influenza hemagglutinin

Virus isolation and quantitation

Protein/Plasma Membrane Association: Possible Involvement of Phosphatidylinositol Turnover and Rho-dependent Signaling Pathway

Inorganic arsenite inhibits IgE receptor-mediated degranulation of mast cells

Regulation of actin assembly by PI(4,5)P2 and other inositol phospholipids: An update on possible mechanisms

Modulation of gelsolin function by phosphatidylinositol 4,5-bisphosphate

The Ebola Virus matrix protein, VP40, requires phosphatidylinositol 4,5-bisphosphate (PI(4,5)P2) for extensive oligomerization at the plasma membrane and viral egress

The Role of Mast Cells in Bacterial Infection

Three-dimensional sub-100 nm resolution fluorescence microscopy of thick samples

Inhibition of PIKfyve kinase prevents infection by Zaire ebolavirus and SARS-CoV-2

Bleed-through correction for rendering and correlation analysis in multi-colour localization microscopy

THE HIGH-AFFINITY IGE RECEPTOR (FcεRI): From Physiology to Pathology

Inactivation of SARS-CoV-2 through Treatment with the Mouth Rinsing Solutions ViruProX((R)) and BacterX((R)) Pro

Cyclic voltammetric technique for the determination of the critical micelle concentration of surfactants, self-diffusion coefficient of micelles, and partition coefficient of an electrochemical probe

Cetylpyridinium Chloride: Mechanism of Action, Antimicrobial Efficacy in Biofilms, and Potential Risks of Resistance

Influence of the lipid composition on the membrane affinity, and the membrane-perturbing ability of cetylpyridinium chloride

Plasma membrane phosphoinositide organization by protein electrostatics

PIP 2 and Proteins: Interactions, Organization, and Information Flow

Mast Cells

Quantification of phosphoinositides reveals strong enrichment of PIP2 in HIV-1 compared to producer cell membranes

Critical Micellle Concentrations of Aqueous Surfacant Systems

Randomized, double-blind, placebo-controlled clinical trial to assess the safety and effectiveness of a novel dual-action oral topical formulation against upper respiratory infections

Cetylpyridinium chloride-containing mouthwashes reduce in vitro SARS-CoV-2 infectivity

Tamas Balla: Leading the way with phosphoinositides

Phosphatidylinositol (4,5) bisphosphate regulates HIV-1 Gag targeting to the plasma membrane

Different isozymes of protein kinase C mediate feedback inhibition of phospholipase C and J o u r n a l P r e -p r o o f Journal Pre-proof stimulatory signals for exocytosis in rat RBL-2H3 cells

Cetylpyridinium Chloride (CPC) Exhibits Potent, Rapid Activity Against Influenza Viruses in vitro and in vivo

Role of charged lipids in membrane structures -Insight given by simulations

A Model For Receptor-Regulated Calcium Entry

Phosphatidylinositol 4,5-Bisphosphate Functions as a Second Messenger that Regulates Cytoskeleton-Plasma Membrane Adhesion

Efficacy of two alcohol-free cetylpyridinium chloride mouthwashes -a randomized double-blind crossover study

Triclosan disrupts immune cell function by depressing Ca(2+) influx following acidification of the cytoplasm

Inhibitors of PI(4,5)P2 synthesis reveal dynamic regulation of IgE receptor signaling by phosphoinositides in RBL mast cells

Dynamics of Capping Protein and Actin Assembly In Vitro: Uncapping Barbed Ends by Polyphosphoinositides

Efficacy of commercial mouth-rinses on SARS-CoV-2 viral load in saliva: randomized control trial in Singapore

Cetylpyridinium chloride interaction with the hepatitis B virus core protein inhibits capsid assembly

Charge Shielding of PIP2 by Cations Regulates Enzyme Activity of Phospholipase C

Cellular proteins in influenza virus particles

Determination of the critical premicelle concentration, first critical micelle concentration and second critical micelle concentration of surfactants by resonance Rayleigh scattering method without any probe

Alterations to Influenza Virus Hemagglutinin Cytoplasmic Tail Modulate Virus Infectivity

The Application of Computers to Taxonomy

Biomedical Image Processing

PIP2 is a necessary cofactor for ion channel function: how and why?

Using Zebrafish Models of Human Influenza A Virus Infections to Screen Antiviral Drugs and Characterize Host Immune Cell Responses

Live cell imaging with protein domains capable of recognizing phosphatidylinositol 4,5-bisphosphate; a comparative study

Influenza virus hemagglutinin concentrates in lipid raft microdomains for efficient viral fusion

Mast cells, brain inflammation and autism

In vitro modelling of rat mucosal mast cell function in Trichinella spiralis infection

The synthesis and cellular roles of phosphatidylinositol 4,5-bisphosphate

Calcium signalling in T cells

Agents That Increase the Permeability of the Outer Membrane

Membrane protein sequestering by ionic protein-lipid interactions

PIP2 signaling in lipid domains: a critical re-evaluation

Effect of salt on the micelles of cetyl pyridinium chloride

Palmitoylation of influenza virus proteins

The effector domain of myristoylated alanine-rich C kinase substrate binds strongly to phosphatidylinositol 4,5-bisphosphate

Segregation of PIP2 and PIP3 into distinct nanoscale regions within the plasma membrane

Quaternary Ammonium-Based Surfactants That Can Recognize Cholesterol-Rich Membranes and Proton-Ionizable Analogs That Cannot

A microplate assay to assess chemical effects on RBL-2H3 mast cell degranulation: effects of triclosan without use of an organic solvent

Antimicrobial agent triclosan disrupts mitochondrial structure, revealed by super-resolution microscopy, and inhibits mast cell signaling via calcium modulation

Multivalent Cation-Bridged PI(4,5)P2 Clusters Form at Very Low Concentrations

Antibacterial and Antiplaque Effects of a Novel, Alcohol-Free Oral Rinse with Cetylpyridinium Chloride

The role of phosphoinositide-regulated actin reorganization in chemotaxis and cell migration

Self assembly of HIV-1 Gag protein on lipid membranes generates PI(4,5)P2/Cholesterol nanoclusters

Cetylpyridinium chloride produces increased zeta-potential on Salmonella Typhimurium cells, a mechanism of the pathogen's inactivation

Inhibition of the Interactions of Cofilin, Destrin, and Deoxyribonuclease I with Actin by Phosphoinositides

Estradiol activates mast cells via a non-genomic estrogen receptor-alpha and calcium influx

We thank Dr. Robert Wheeler, Siham Hattab, and Bailey Blair for use of and help with the ibidi heating system and confocal microscope; Dr. Joshua Kelley for guidance regarding image analysis; Dr. Juyoung Shim for helpful discussions, and Liz Saavedra Perez for the confirmation of TCID 50 assay results, and Patricia Byard for administrative assistance.