Plant Alkaloids Inhibit Membrane Fusion Mediated by Calcium and Fragments of MERS-CoV and SARS-CoV/SARS-CoV-2 Fusion Peptides

To rationalize the antiviral actions of plant alkaloids, the ability of 20 compounds to inhibit calcium-mediated fusion of lipid vesicles composed of phosphatidylglycerol and cholesterol was investigated using the calcein release assay and dynamic light scattering. Piperine, tabersonine, hordenine, lupinine, quinine, and 3-isobutyl-1-methylxanthine demonstrated the most potent effects (inhibition index greater than 50%). The introduction of phosphatidylcholine into the phosphatidylglycerol/cholesterol mixture led to significant changes in quinine, hordenine, and 3-isobutyl-1-methylxanthine efficiency. Comparison of the fusion inhibitory ability of the tested alkaloids, and the results of the measurements of alkaloid-induced alterations in the physical properties of model membranes indicated a potent relationship between a decrease in the cooperativity of the phase transition of lipids and the ability of alkaloids to prevent calcium-mediated vesicle fusion. In order to use this knowledge to combat the novel coronavirus pandemic, the ability of the most effective compounds to suppress membrane fusion induced by fragments of MERS-CoV and SARS-CoV/SARS-CoV-2 fusion peptides was studied using the calcein release assay and confocal fluorescence microscopy. Piperine was shown to inhibit vesicle fusion mediated by both coronavirus peptides. Moreover, piperine was shown to significantly reduce the titer of SARS-CoV2 progeny in vitro in Vero cells when used in non-toxic concentrations.


Calcein Leakage Assay
The choice of membrane-forming lipids is related to the enrichment of the viral lipid envelope with negatively charged and neutral glycerophospholipids, CHOL, and SM [45][46][47]. Small unilamellar vesicles were prepared from the mixtures of DOPG/CHOL (80/20 mol%), DOPC/DOPG/CHOL (40/40/20 mol%), and POPC/SM/CHOL (60/20/20 mol%) by extrusion. The resulting liposome suspension contained 3 mM lipid. Lipid stock in chloroform was dried under a gentle stream of nitrogen. Dry lipid film was hydrated by a buffer (35 mM calcein, 10 mM HEPES, pH 7.4). The suspension was subjected to five freeze-thaw cycles and then passed through a 100 nm nuclepore polycarbonate membrane for 13 times. The calcein that was not entrapped in the vesicles was removed by gel filtration with a Sephadex G-50 column to replace the buffer outside the liposomes with calcein-free solution (0.15 M NaCl, 10 mM HEPES, pH 7.4). Calcein inside the vesicles fluorescence very poorly, because of strong self-quenching at millimolar concentrations, while the fluorescence of disengaged calcein in the surrounding media is due to liposome fusion [41].
The structure/function study performed by [49] showed that the fragment of MERS-CoV spike protein S2 subunit (RSARSAIEDLLFDKV), which is highly conserved throughout the coronavirus family, induces the dipalmitoylphosphatidylcholine/SM/CHOL liposome syncytium formation, and the IEDLLF core sequence has a key role in membrane perturbation. A shorter α-helical fragment of MERS-CoV fusion peptide (RSAIEDLLFD-KVT) and the homologous highly conserved fragment of SARS-CoV/SARS-CoV-2 fusion peptides (RSFIEDLLFNKVT) [52] were chosen to mimic viral associated fusion. The chosen peptides were further referred to FP-MERS-CoV and FP-SARS-CoV-2, respectively. FP-SARS-CoV-2 and FP-MERS-CoV were added to POPC/SM/CHOL vesicles (60/20/20 mol%) to induce membrane fusion. The results of adding the different concentrations of peptides into aqueous solutions, bathing calcein-loaded POPC/SM/CHOL vesicles, are shown in Figure S2c To test the fusion inhibitory action of alkaloids, a fusion inducer (calcium or modeling fusion peptides at the appropriate concentrations) was added to liposomes, pre-incubated with 400 µM of alkaloids (except for 40 µM of tabersonine). Since tabersonine at a concentration of 400 µM completely inhibited the calcium-mediated fusion of both DOPG/CHOL and DOPC/DOPG/CHOL vesicles, in order to understand whether its inhibitory effect depended on the lipid composition of liposomes, we reduced its concentration by 10-fold. Both concentrations of tabersonine (40 and 400 µM) were used to test its inhibitory action on the FP-SARS-CoV-2 and FP-MERS-CoV mediated fusion.
We found that the osmotic pressure gradient resulting from the addition of a fusion inducer or an inhibitor did not affect the leakage of calcein from liposomes (data not shown).
The degree of calcein release was determined using a Fluorat-02-Panorama spectrofluorometer (Lumex, Saint-Petersburg, Russia). The excitation wavelength was 490 nm and the emission wavelength was 520 nm. Triton X-100 was added to a final concentration of 1% to each sample in order to completely disrupt liposomes, and the intensity after releasing the total amount of calcein from liposomes was measured. To describe the fusion of the liposomal membranes, the relative fluorescence of leaked calcein (RF, %) was calculated using the following formula: where I and I 0 were the calcein fluorescence intensities in the sample, in the presence and absence of the fusion inducer/inhibitor, respectively, and I max was the maximum fluorescence of the sample after lysis of liposomes by Triton X-100. Factor of 0.9 was introduced to calculate the dilution of the sample by Triton X-100. The inhibition index (II) was calculated to characterize the relative efficiency of tested alkaloids to inhibit the fusion of liposomes compared to the action of the fusion inducer alone on the same liposome preparation: where RF U and RF S were the maximum relative fluorescence of the leaked calcein at the vesicle fusion induced by the calcium or modeling fusion peptides in the absence (RF U , U = Ca 2+ /FP-MERS-CoV/FP-SARS-CoV-2) and presence of alkaloids (RF S ), respectively. The estimation of RF S was made, taking into account the intrinsic effect of alkaloids (the calcein leakage due to the potent lipid disordering action of alkaloid alone, RF A ). Figure S3 and   Figure S3c). Time dependences of calcein leakage were fitted with two-exponential functions with characteristic times, t 1 and t 2 , related to the fast and slow components of the release process.

Confocal Fluorescence Microscopy
The visualization of changes in the morphological features and behavior of vesicles in the suspension under the action of fusion inducers was performed by marking the liposome membranes by the fluorescent labeled lipid, Rh-DPPE. Giant unilamellar vesicles were prepared from the mixture of POPC/SM/CHOL (60/20/20 mol%) and 1 mol% of fluorescent lipid probe Rh-DPPE by the electroformation method (standard protocol, 3 V, 10 Hz, 1 h, 25 • C) using Nanion Vesicle Prep Pro (Munich, Germany). The resulting liposome suspension contained 1 mM lipid in 0.5 M sorbitol solution. The measurements of the liposome fusion were added; 10% of PEG-8000, 50 µM of FP-SARS-CoV-2, and 200 µM of MERS-CoV were used to induce membrane fusion and were incubated for 1-2 min at room temperature (25 ± 1 • C). Vesicles were imaged through an oil immersion objective (65 ×/1.4HCX PL) using an Olympus (Hamburg, Germany). A helium-neon laser with a wavelength of 561 nm was used to excite Rh-DPPE. The temperature during observation was controlled by air heating/cooling in a thermally insulated camera.

Dynamic Light Scattering
Small unilamellar vesicles composed of DOPG/CHOL (80/20 mol%) and POPC/SM/ CHOL (60/20/20 mol%) were prepared by the extrusion and were treated with 400 µM of cotinine, melatonin, 3-isobutyl-1-methylxanthine, lupinine, piperine, hordenine, and 40 or 400 µM of tabersonine for 30 min before the addition of 20 mM CaCl 2 , 50 µM FP-SARS-CoV-2, and 200 µM MERS-CoV solution. Three different control samples were used: unmodified liposomes and vesicles incubated with fusion inducer, and the inhibitor alone. We found that the addition of alkaloid alone into the liposome suspension did not result in a noticeable alteration in vesicle size (data not shown). The hydrodynamic diameter (d, nm) and zeta-potential (ζ, mV) were determined on a Malvern Zetasizer Nano ZS 90 (Malvern Instruments Ltd., Malvern, United Kingdom) by gradual titration of liposome suspension in PBS at 25 • C.

Membrane Boundary Potential Measurements
Virtually solvent-free planar lipid bilayers were prepared using a monolayer-opposition technique [53] on a 50-µm diameter aperture in a 10-µm thick Teflon film separating the two compartments of the Teflon chamber. The steady-state conductance of K + -nonactin was modulated via the two-sided addition of the colchicine, cotinine, 1,7-dimethylxanthine, capsaicin, synephrine, 3-isobutyl-1-methylxanthine, lupinine, piperine, tabersonine, and hordenine from different mM stock solutions in ethanol or methanol to the membranebathing solution (0.1 M KCl, pH 7.4) to obtain a final concentration ranging from 5 µM to 1 mM. The aperture was pretreated with hexadecane. Lipid bilayers were made from DOPG/CHOL (80/20 mol%). The conductance of the lipid bilayers was determined by measuring membrane conductance (G) at a constant transmembrane voltage (V = 50 mV). In the subsequent calculations, the membrane conductance was assumed to be related to the membrane boundary potential (ϕ b ), the potential drop between the aqueous solution and the membrane hydrophobic core, by the Boltzmann distribution [54]: where ξ is the ion mobility, ze is the ion charge, k is the Boltzmann constant, and T is the absolute temperature. Ag/AgCl electrodes with 1.5% agarose/2 M KCl bridges were used to apply V and measure G.
The current was measured using an Axopatch 200B amplifier (Molecular Devices, LLC, Orleans Drive, Sunnyvale, CA, USA) in the voltage clamp mode. Data were digitized using a Digidata 1440A and analyzed using pClamp 10.0 (Molecular Devices, LLC, Orleans Drive, Sunnyvale, CA, USA) and Origin 8.0 (OriginLab Corporation, Northampton, MA, USA). Data were acquired at a sampling frequency of 5 kHz using low-pass filtering at 200 Hz.
All experiments were performed at room temperature (25 • C).

Differential Scanning Microcalorimetry
Differential scanning microcalorimetry experiments were performed by a µDSC 7EVO (Setaram, Caluire-et-Cuire, France). Giant unilamellar vesicles were prepared from a mixture of DPPG/CHOL (90/10 mol%) by the electroformation method (standard protocol, 3 V, 10 Hz, 1 h, 55 • C). The liposome suspension contained 3 mM lipid and was buffered by 5 mM HEPES at pH 7.4. The tested alkaloids (colchicine, cotinine, 1,7-dimethylxanthine, capsaicin, synephrine, 3-isobutyl-1-methylxanthine, lupinine, piperine, and hordenine) were added to aliquots to obtain a final concentration up to 400 µM of alkaloids (except for 40 µM of tabersonine). The liposomal suspension was heated at a constant rate of 0.2 C·min −1 . The reversibility of the thermal transitions was assessed by reheating the sample immediately after the cooling step from the previous scan. The temperature dependence of the excess heat capacity was analyzed using Calisto Processing (Setaram, Caluire-et-Cuire, France). The peaks on the thermograms were characterized by the maximum temperatures of the main phase transition (T m ) of DPPG/CHOL mixture and the half-width of the main peak (T 1/2 ) indicating the size of the lipid cooperative unit.

In Vitro Antiviral Activity Assessment
To analyze the antiviral activity of piperine, monkey kidney epithelial cells, Vero (ATCC CCL81) were grown in culture vials (Nunc, Roskilde, Denmark) in DMEM culture medium supplemented with 10% FBS. The cell concentration was adjusted to 5 × 10 5 cells/mL, and cells were seeded into 96-well culture plates (0.1 mL per well). The plates were incubated for 24 h at 37 • C in an atmosphere of 5% CO 2 . Piperine was dissolved in DMSO, and then two-fold serial dilutions of 200 to 1.56 µg/mL in serum-free DMEM/F12 nutrient medium were prepared. Afterwards, 100 µL/well of each dilution of piperine was added to Vero cells and incubated for 1 h at 37 • C in a 5% CO 2 atmosphere. Serum-free DMEM/F12 medium was added to control wells instead of piperine. Aliquots of SARS-CoV-2 (10 3 TCID50/mL) (isolate 17612) were mixed 1:1 with dilutions of piperine and incubated for 1 h at 37 • C. After 1 h incubation, piperine was removed from the plates, 100 µL/well of the SARS-CoV2/piperine mixture was added to the corresponding wells and incubated for 1 h at 37 • C in a 5% CO 2 atmosphere, the virus-containing medium was added to the wells of the viral control, and serum-free DMEM/F12 was added to the cell control. The virus was then removed, the cells were washed three times with DPBS, 100 µL/well of dilutions of piperine (200-1.56 µg/mL) were added to the wells, serum-free DMEM/F12 medium was added to the control wells, and the plates were incubated for 24 h at 37 • C in a 5% CO 2 atmosphere. After 24 h, supernatants from the experimental and control wells with the virus were collected and titrated by TCID 50 assay in Vero cells after incubation for 72 h. The visual assessment of the cytopathogenic effect (CPE) of the virus was performed using an Olympus CKX41 light inverted microscope (Olympus, Tokyo, Japan).
To prove the statistical significance of difference between RF-, t 1 -, t 2 -, d-, ζ-, T m -, and T 1/2 -values at the addition of alkaloids and the control values of related parameters (in the absence of alkaloids) the nonparametric signed-rank U-test (Mann-Whitney-Wilcoxon's U-test) was employed (*-p ≤ 0.01, **-p ≤ 0.05).
Pearson's correlation coefficient was applied to estimate the potent relationship between the inhibitory action of alkaloids on the vesicle fusion and the alkaloid-induced alterations in physical properties of lipid bilayers (∆T m , ∆T 1/2 , ∆ϕ b ). The coefficients were calculated using Microsoft Excel (Microsoft Corp., Redmond, WA, USA). . Some alkaloids were able to inhibit calcium-mediated fusion, and this ability strictly depended on the alkaloid type ( Figure S4b,c). Table 1 summarizes the mean values of maximum marker leakage caused by calcium-induced fusion of DOPG/CHOL vesicles pretreated by different alkaloids (RF S ). The mean maximum leakage at the addition of fusion inducer alone is about 90%. A significant and reliable decrease in RF S (indicating the inhibition of calcium-mediated fusion) is caused by pretreatment of DOPG/CHOL vesicles with dihydrocapsaicin (RF S is about 70%), capsaicin, synephrine (RF S is about 50%), 3-isobutyl-1-methylxanthine, quinine, piperine (RF S is about 30%), tabersonine, hordenine (RF S is about 20%), and lupinine (RF S is about 10%).

DOPG/CHOL DOPC/DOPG/CHOL
Alkaloid RF S , % t 1 , min t 2 , min RF S , % t 1 , min t 2 , min  Table 1 also demonstrates the kinetic parameters of liposome fusion in the presence of alkaloids. It should be noted that the time dependences of the marker release due to vesicle fusion are well-described by two-exponential dependences with characteristic times related to fast and slow components (t 1 and t 2 , respectively). Analysis of Table 1 indicates that t 1 and t 2 values vary in the ranges of about 1-10 and 10-120 min, respectively. The fast component might characterize the calcium sorption on the negatively charged DOPGenriched membranes and the subsequent aggregation of the lipid vesicles, while the slow one might be related to the topological rearrangements of membrane lipids and further dye leakage at the liposome fusion. Colchicine, capsaicin, and tabersonine significantly accelerated the marker release kinetics at calcium-mediated fusion of DOPG/CHOL vesicles (about 5-10-fold decrease in both characteristic times is observed) ( Table 1). Melatonin and quinine drastically slowed down the fusion process increasing t 2 from about 70 to more than 100 min (Table 1).
(II is about 25%) ≤ capsaicin ≈ synephrine (II is about 45%). The latter (the most effective) group includes 3-isobutyl-1-methylxanthine (II is about 65%), piperine ≈ quinine (II is about 70%), tabersonine (II is about 75%), hordenine (II is about 80%), and lupinine (II is about 85%). Figure 2a shows the results of size measurements after calcium addition to unmodified DOPG/CHOL liposomes and vesicles pretreated by alkaloids of various anti-fusogenic activity, cotinine, melatonin, 3-isobutyl-1-methylxanthine, lupinine, piperine, tabersonine, and hordenine. The diameter of DOPG/CHOL liposomes in the absence of any modifiers is equal to 80 ± 20 nm. CaCl2 induces liposome fusion, and the vesicle size increased up to 145 ± 40 nm (Figure 2a). Cotinine and melatonin practically did not affect the size-increasing effect of calcium, as the liposome diameter was about 140 nm. The other tested compounds (lupinine, 3-isobutyl-1-methylxanthine, tabersonine, piperine, and hordenine) significantly inhibited calcium-mediated liposome fusion: the diameter of the vesicles was equal to 77 ± 13 nm ( Figure 2a). The data obtained are in good agreement with the II values assessed by the calcein release assay (Figure 1a).    Figure 2a shows the results of size measurements after calcium addition to unmodified DOPG/CHOL liposomes and vesicles pretreated by alkaloids of various anti-fusogenic activity, cotinine, melatonin, 3-isobutyl-1-methylxanthine, lupinine, piperine, tabersonine, and hordenine. The diameter of DOPG/CHOL liposomes in the absence of any modifiers is equal to 80 ± 20 nm. CaCl 2 induces liposome fusion, and the vesicle size increased up to 145 ± 40 nm (Figure 2a). Cotinine and melatonin practically did not affect the sizeincreasing effect of calcium, as the liposome diameter was about 140 nm. The other tested compounds (lupinine, 3-isobutyl-1-methylxanthine, tabersonine, piperine, and hordenine) significantly inhibited calcium-mediated liposome fusion: the diameter of the vesicles was equal to 77 ± 13 nm ( Figure 2a). The data obtained are in good agreement with the II values assessed by the calcein release assay (Figure 1a). Figure 2b presents the results of estimation of the ζ-potential of DOPG/CHOL liposomes in the presence of 20 mM CaCl 2 without and with different alkaloids. The ζ-potential of unmodified DOPG/CHOL liposomes was equal to about −70 mV. The addition of calcium led to an increase in ζ-potential up to about −50 mV. The ζ-potential value of DOPG/CHOL vesicles pretreated with alkaloids was 10-20 mV less than that of untreated ones (Figure 2b). The observed incomplete compensation of negative surface charge of DOPG-enriched liposomes was expected at the introduction of positively charged calcium ions and alkaloid molecules (lupinine, tabersonine, and hordenine). The disagreement of alkaloid abilities to inhibit calcium-mediated vesicle fusion and to increase ζ-potential (Figure 2a,b) clearly demonstrates that the anti-fusogenic activity of alkaloids is not due to partial compensation of the negative membrane surface charge by alkaloids or a competition between alkaloids and calcium ions for the interaction with negatively charged lipid groups.   To study the dependence of the inhibition effect of alkaloids on the lipid composition of fusing vesicles, DOPC was introduced into membrane composition. It should be noted that the cell membranes are enriched of phosphatidylcholine species compared to virions [47,55,56]. The decrease in the proportion of negatively charged DOPG in membrane composition from 80 to 40 mol% was accompanied by an increase in CaCl 2 concentration from 20 to 40 mM to reach the RF S value of about 80%. Figure S4d-f (Supplementary Materials) demonstrates the calcein leakage caused by the calcium-mediated fusion of DOPC/DOPG/CHOL liposomes. RF S resulted from calcium-mediated fusion of DOPC/DOPG/CHOL vesicles pretreated by different alkaloids reliably decreases in the series: caffeine ≈ pentoxifylline ≈ cotinine ≈ quinine (RF S is about 65%) ≥ atropine ≈ dihydrocapsaicin ≈ capsaicin (RF S is about 55%) > lupinine (RF S is about 40%) > piperine (RF S is about 25%) > tabersonine (RF S is equal to 1%) ( Table 1).
The kinetic parameters of the time dependence of the calcein leakage caused by calciummediated fusion of DOPC/DOPG/CHOL liposomes are shown in Table 1. Colchicine and capsaicin accelerate the release kinetics due to calcium-mediated fusion of DOPC/DOPG/CHOL vesicles as well as of DOPG/CHOL liposomes. The huge inhibition of calcium-mediated fusion of DOPC/DOPG/CHOL liposomes in the presence of tabersonine did not allow for determining the characteristic times of dye release kinetics. Similar to DOPG/Chol vesicles, the pretreatment of DOPC/DOPG/CHOL liposomes with colchicine and capsaicin led to about five-fold decrease in t 2 -value. Unlike the case of DOPG/CHOL vesicles, the kinetics of calcein leakage due to fusion of DOPC/DOPG/CHOL liposomes was also accelerated by piperine (by six times), while theophylline and atropine slightly slowed down the fusion process, increasing t 2 from 50 to about 100 and 120 min, respectively (Table 1). Figure 1b shows the II values, characterizing the relative ability of different alkaloids to suppress the fusion of DOPC/DOPG/CHOL vesicles. Moreover, 3,9-dimethylxanthine, 7-(β-hydroxyethyl)theophylline, colchicine, 1,7-dimethylxanthine, theophylline, melatonin, berberine, synephrine, 3-isobutyl-1-methylxanthine, and hordenine did not demonstrate a significant ability to inhibit the fusion of DOPC/DOPG/CHOL liposomes. Caffeine, pentoxifylline, cotinine, atropine, dihydrocapsaicin, capsaicin, and quinine were characterized by low or medium efficiency to suppress the fusion of vesicles made of ternary lipid mixture, while the II values in the presence of lupinine, piperine, and tabersonine were about 50, 70, and 100% (Figure 1b). Comparison of Figure 1a,b clearly demonstrates that the anti-fusogenic activity of alkaloids strictly depends on the lipid composition of the vesicles.

The Influence of Alkaloids on the Physical Properties of the Model Lipid Membrane
It has been repeatedly shown that the elastic characteristics of the lipid bilayer play a huge role in the membrane fusion. The ordering of the lipid head groups and acyl chains, transbilayer lateral pressure profile, phase state of lipids, spontaneous curvature of the membrane, area per lipid molecule, and other characteristics can influence fusion [57]. Many of these parameters influence each other, but all of them are directly dependent on the lipid composition of the membrane. The key role of lipids is also supported by the fact that viral infections can alter cell lipid synthesis, regulating it according to their needs [58]. The ability of alkaloids and their derivatives to change the physical properties of a bilayer upon intercalation was demonstrated in a number of studies [59][60][61][62]. In particular, alkaloids affect the lipid packing [63,64], change lipid phase transition temperatures [65,66], etc. Recently, we performed a detailed study of alkaloid effects on the physical properties of the POPC membrane [67]. To understand the relationship between the alkaloid effects on the fusion and modulation of lipid matrix under their action, the influence of tested alkaloids on the properties of PG and CHOL-enriched bilayers were performed. Figure 3 presents the typical heating thermograms of DPPG/CHOL-liposomes in the absence (control) and in the presence of tested alkaloids. The value of the phase transition temperature (T m ) of DPPG/CHOL mixture, and the half-width of the main peak (T 1/2 ) are summarized in Table 2. T m is the point at which thermally induced lipid melting occurs [68], while T 1/2 describes the sharpness of the phase transition or the inverse cooperativity of this process [69]. The temperature of the main transition (T m ) of untreated DPPG/CHOL vesicles is equal to 40.4 • C, with a half-width of the peak (T 1/2 ) of about 0.8 • C (Figure 3, control).

The Influence of Alkaloids on the Physical Properties of the Model Lipid Membrane
It has been repeatedly shown that the elastic characteristics of the lipid bilayer play a huge role in the membrane fusion. The ordering of the lipid head groups and acyl chains, transbilayer lateral pressure profile, phase state of lipids, spontaneous curvature of the membrane, area per lipid molecule, and other characteristics can influence fusion [57]. Many of these parameters influence each other, but all of them are directly dependent on the lipid composition of the membrane. The key role of lipids is also supported by the fact that viral infections can alter cell lipid synthesis, regulating it according to their needs [58]. The ability of alkaloids and their derivatives to change the physical properties of a bilayer upon intercalation was demonstrated in a number of studies [59][60][61][62]. In particular, alkaloids affect the lipid packing [63,64], change lipid phase transition temperatures [65,66], etc. Recently, we performed a detailed study of alkaloid effects on the physical properties of the POPC membrane [67]. To understand the relationship between the alkaloid effects on the fusion and modulation of lipid matrix under their action, the influence of tested alkaloids on the properties of PG and CHOL-enriched bilayers were performed. Figure 3 presents the typical heating thermograms of DPPG/CHOL-liposomes in the absence (control) and in the presence of tested alkaloids. The value of the phase transition temperature (Tm) of DPPG/CHOL mixture, and the half-width of the main peak (T1/2) are summarized in Table 2. Tm is the point at which thermally induced lipid melting occurs [68], while T1/2 describes the sharpness of the phase transition or the inverse cooperativity of this process [69]. The temperature of the main transition (Tm) of untreated DPPG/CHOL vesicles is equal to 40.4 °C, with a half-width of the peak (T1/2) of about 0.8 °C (Figure 3, control).   Table 2. The parameters characterized the effects of alkaloids on the physical properties of lipid bilayers: T m -the main transition temperature of DPPG/CHOL (90/10 mol%); T 1/2 -the half-width of the main peak in the presence of 400 µM of alkaloids (40 µM of tabersonine); ∆ϕ b (max)-the maximum changes in the boundary potential of DOPG/CHOL (80/20 mol%) membranes. Colchicine, cotinine, and 1,7-dimethylxanthine did not affect the thermotropic phase behavior of DPPG/CHOL mixture ( Figure 3, Table 2). Synephrine, 3-isobutyl-1-methylxanthine, lupinine, and hordenine decreased the T m by approximately 0.3 to 0.7 • C and increased the T 1/2 by 0.4 to 0.6 • C ( Figure 3, Table 2). The data are in agreement with the study of hordenine influence on the phase behavior of dimyristoylphosphatidylglycerol [61]. The introduction of capsaicin, piperine, and tabersonine was accompanied by a sharp decrease in T m (0.9-2.1 • C) and increase in T 1/2 (0.7-1.2 • C) ( Figure 3, Table 2). The correlation coefficients between LogD o/w (Table S1 Supplementary Materials) and -∆T mand ∆T 1/2 -values ( Table 2) are in the range of 0.74-0.77, demonstrating a good correlation between the lipophilicity of the alkaloid molecules and their ability to alter lipid packing.
The correlation coefficient between the alkaloid-induced changes in T m of DPPG/CHOL ( Table 2) and their II values in DOPG/CHOL vesicles (Figure 1a) was equal to 0.52, the corresponding coefficient characterizing the interdependence of alkaloid-induced changes in T 1/2 ( Table 2) and the II values (Table 1) was equal to 0.63. The correlation coefficient between the alkaloid-induced changes in T m ( Table 2) and their II values in DOPC/DOPG/CHOL vesicles (Table 1) was equal to 0.64, while the interdependence of compound-induced changes in T 1/2 ( Table 2) and alkaloid II values of DOPC/DOPG/CHOL liposomes (Figure 1b) was characterized by a coefficient of 0.68. The observed correlation between the parameters characterizing the lipid melting, especially the cooperativity of lipid phase transition, and the fusion inhibition index indicates a relationship between the ability of alkaloids to disorder membrane lipids and to inhibit the liposome fusion. The incorporation of alkaloids into the polar region of the membrane, leading to an increase in lateral pressure, positive curvature stress, and an area per lipid molecule in this region, contributes to both a reduction in the temperature/cooperativity of the phase transition and an increase in the energy of fusion intermediates characterized by lipid surfaces of a negative curvature. Moreover, a comparison of the data obtained ( Figure  3, Table 2) and our recently published results [65] demonstrate that the tested alkaloids have different effects on the phase behavior of DPPG/CHOL and DPPC. This fact might be related to the observed difference in their ability to inhibit the fusion of DOPG/CHOL and DOPC/DOPG/CHOL vesicles (Figure 1, Table 1). Alkaloid-induced curvature stress should depend on the depth of molecule insertion into lipid bilayer and its orientation in the membrane. The significant inhibitory action of 3-isobutyl-1-methylxanthine on the fusion of DOPG/CHOL liposomes might be related to the induction of the high posi-tive curvature stress due to both the electrostatic interaction between xanthine residue and DOPG on the membrane surface and the tendency of the isobutyl side chain to be embedded into the membrane hydrocarbon core. Xanthine derivatives with lower hydrophobicity, i.e., caffeine, pentoxifylline, 1,7-dimethylxanthine, 3,9-dimethylxanthine, and 7-(β-hydroxyethyl)theophylline (Table S1, Supplementary Materials) did not inhibit calcium-mediated fusion of DOPG/CHOL liposomes. Introduction of DOPC into the membrane lipid composition was accompanied by a two-fold decrease in the DOPG content, which led to significant changes in the membrane orientation and embedment of xanthines. Thus, caffeine and pentoxifylline demonstrated a weak inhibitory effect on a fusion of DOPC/DOPG/CHOL vesicles, while 3-isobutyl-1-methylxanthine did not affect this process. The complete loss of the inhibiting ability of β-phenylethylamine derivatives, synephrine and hordenine, in membranes with lower content of DOPG compared to DOPGenriched bilayers, could be explained in a similar way. An increase in the inhibitory effect of tabersonine on the fusion of DOPC/DOPG/CHOL vesicles compared to DOPG/CHOL liposomes might indicate its deeper immersion into the DOPC/DOPG/CHOL membranes than into DOPG/CHOL bilayers due to a decrease in the electrostatic interactions between positively charged tabersonine and negatively charged DOPG molecules.
Considering that liposome fusion is triggered by calcium ions, we also considered the possibility of alkaloids influencing fusion by changing the boundary potential of the membranes. Table 2 shows the maximum changes in ϕ b at the adsorption of different alkaloids (-∆ϕ b (max)). The dependences of ∆ϕ b on the concentrations of tested alkaloids are presented in Figure S5 Table 2). Colchicine, synephrine, 3-isobutyl-1-methylxanthine, lupinine, and piperine led to reduction in ϕ b by about 20-40 mV ( Figure S5 in the Supplementary Materials, Table 2). Cotinine, 1,7-dimethylxanthine, and hordenine practically did not alter the ϕ b -value ( Figure S5 Supplementary Materials, Table 2). There was no correlation between ∆ϕ b of DOPG/CHOL membranes after the addition of alkaloids and their II values. This fact demonstrates the minor role of changes in the bilayer electrical properties by alkaloids in their ability to inhibit the calcium-mediated membrane fusion.

Liposome Fusion Mediated by Fragments of Coronavirus Fusion Peptides
The COVID-19 pandemic has caused major challenges to healthcare systems across the globe. The lack of specific treatment, virus mutations, the emergence of new strains that have become resistant to vaccines, and many other factors, have led to the search for new drugs to treat COVID-19. To demonstrate the pharmacological applications of alkaloids as coronavirus fusion inhibitors-the abilities of the most effective compounds to suppress the membrane fusion induced by FP-MERS-CoV and FP-SARS-CoV-2 were studied. These peptides are able to induce calcein release due to fusion of POPC/SM/CHOL liposomes ( Figure S2c

Antiviral Evaluation
Piperine is an alkaloid found in black pepper, one of the most widely used spices. The in vitro and in vivo antiviral activity of the alkaloid against the MERS-CoV in Vero cells, and in a mice model, was already demonstrated [70]. Here, we evaluated in vitro antiviral activity of piperine against the SARS-CoV-2 virus.
According to results of the cytotoxicity analysis performed by Hegeto et al. [71], the value of IC 50 of piperine against Vero cells was equal to 183.33 µg/mL. Based on this finding, the maximum piperine concentration of 200 µg/mL was chosen. The incubation with 200 µg/mL of piperine for 72 h led to death of all cells in wells; all lower piperine concentrations used were nontoxic to Vero cells (data not shown).
To determine the infectious activity of viral progeny, ten-fold dilutions of the supernatants collected from the experimental wells and viral controls in serum-free DMEM/F12 medium were prepared. The resulting dilutions were added to a 96-well culture plate with 80-90% Vero cell monolayer and incubated for 72 h. After that, visual assessment of the cytopathogenic effect (CPE) of the virus was performed. The control samples without any additives demonstrated that the virus titer was 10 4 TCID 50 /mL. Samples with a piperine concentration of 1.56 µg/mL resulted in 10 3 TCID 50 /mL; alkaloid concentrations of 3.12-25 µg/mL had 10 2 TCID 50 /mL, and in samples with piperine, concentrations of 50 and 100 µg/mL CPE were not observed at all. Thus, the significant reduction of the titer of SARS-CoV2 progeny in Vero cells at 1.56-100 µg/mL of piperine was clearly demonstrated.

Conclusions
New broad-spectrum antiviral drugs are needed due to the increase in the number of viral infections. Virus fusion inhibitors could be effective antiviral agents because fusion is a necessary step for the lifecycle of the virus. We used plant secondary metabolites, in particular, alkaloids, as a new class of fusion inhibitors. It was shown that the ability of alkaloids to inhibit calcium-mediated fusion of liposomes depends on their lipid composition. A correlation between the anti-fusogenic activity of alkaloids and their disordering effects on membrane lipids was found. Additionally, the ability of piperine to suppress the fusion induced by fragments of the coronavirus fusion peptides (MERS-CoV and SARS-CoV/SARS-CoV-2) was demonstrated. We also showed that piperine dramatically reduced the titer of SARS-CoV2 progeny in vitro in Vero cells when used in non-toxic concentrations. Thus, we hypothesize that the antiviral activity of piperine is related to its lipid-associated action. Moreover, to our knowledge, except for one bioassay with SARS-CoV-2-S pseudotyped particles [72], this is the first study demonstrating the in vitro activity of piperine against SARS-CoV-2 virus. It should be noted that piperine with curcumin is being studied in several clinical trials for the treatment of COVID-19 [73][74][75]. According to [75], administration of oral curcumin with piperine substantially reduced the morbidity and mortality due to COVID-19. We hope that, over time, the number of effective alkaloids in the treatment of COVID-19 will increase. The use of plant metabolites can successfully complete the existing therapeutic strategies, which will make it possible to combat the SARS-CoV-2 virus more effectively.