Prenol, but Not Vitamin C, of Fruit Binds to SARS-CoV-2 Spike S1 to Inhibit Viral Entry: Implications for COVID-19

Despite great success of vaccination against COVID-19, vaccines alone cannot provide 100% protection against COVID-19. Although repurposing of remdesivir and dexamethasone for COVID-19 has been approved by the Food and Drug Administration, these drugs exhibit different side effects (1–3). Therefore, a specific and effective, but nontoxic, molecule is needed to control this viral pandemic.

It has been known for the centuries that consuming fruits may be beneficial for fighting infection in the body. Fruits contain vitamin C or ascorbic acid and during the first and second waves of COVID-19 cases, vitamin C almost disappeared from local stores in many parts of the world. However, clinical trials do not provide a clear answer to whether vitamin C is beneficial for COVID-19 patients. Although the i.v. use of vitamin C is feasible in COVID-19 patients (4), in clinical trials, vitamin C does not significantly reduce the length of hospital stay, readmission rate, admission to intensive care, need for advanced oxygen support, and mortality (5, 6). In an open-label, randomized, and controlled trial, the authors (7) do not find significantly better outcomes in the COVID-19 patient group who were treated with high-dose i.v. vitamin C in addition to the main treatment regimen at discharge. According to Hui et al. (8), genetically predicted circulating levels of vitamin C do not associate with susceptibility to severe COVID-19, hospitalization as a result of COVID-19, any COVID-19 infection, or pneumonia. Therefore, we analyzed vitamin C and a few other major components of fruit for any possible anti–COVID-19 activities.

Because angiotensin-converting enzyme 2 (ACE2), being a beneficial molecule, converts a vasoconstrictor (angiotensin II) to a vasodilator (angiotensin 1–7) (9–12) and the spike protein on the surface of SARS-CoV-2 binds to ACE2 (13–16) for entering into human cells, we screened different components of fruit and observed that prenol, but not other major constituents of fruit (vitamin C, rutin, and cyanidin), was capable of suppressing the binding of SARS-CoV-2 spike S1 with ACE2. Accordingly, prenol, but not vitamin C, inhibited the entry of pseudotyped SARS-CoV-2 into human ACE2 (hACE2)-expressing HEK293 cells and suppressed SARS-CoV-2 spike S1–induced inflammation in human A549 lungs cells, suggesting that prenol, not vitamin C, of fruit may be beneficial for COVID-19. Consistently, oral treatment with prenol led to a decrease in proinflammatory molecules in the lungs, reduction of fever, attenuation of arrhythmias, and enhancement of locomotor activities in an animal model of COVID-19, highlighting the possible therapeutic implication of prenol in COVID-19. However, vitamin C, not prenol, inhibited the entry of pseudotyped vesicular stomatitis virus (VSV) into human cells, suggesting that vitamin C, not prenol, may be beneficial for VSV infection.

Prenol, vitamin C, cyanidin, rutin, and methyl butane were purchased from Sigma (St. Louis, MO). 3-Hydroxy-(2, 2)-dimethyl butyrate (HDMB) was purchased from Santa Cruz Biotechnology (Dallas, TX). Methyl cellulose was received from Spectrum Chemical (Gardena, CA). SARS-CoV-2 spike-pseudotyped lentiviral particles, VSV-G-pseudotyped particles, and hACE2-expressing HEK293 cells were purchased from GeneCopoeia (Rockville, MD). rCOVID-19 Spike protein S1 (14–685) was purchased from Abeomics (San Diego, CA). Recombinant hACE2 protein (18–739) was purchased from MyBiosource (San Diego, CA). rSARS-CoV-2 Spike Receptor-Binding Domain (RBD; N501Y) and SARS-CoV-2 Spike RBD (E484K) were purchased from Biovision (Milpitas, CA). rSARS-CoV-2 Spike Protein RBD (Delta Variant) was bought from Abbexa (Cambridge, UK). rSARS-CoV-2 Omicron variant (B.1.1.529) spike S1 was purchased from Genscript (Piscataway, NJ). Human A549 lung carcinoma cell line and F-12K medium were obtained from ATCC (Manassas, VA). HBSS, 0.05% trypsin, and antibiotic-antimycotic mixture were bought from Mediatech (Washington, DC). FBS was obtained from Atlas Biologicals (Fort Collins, CO). ACE2:SARS-CoV-2 Spike Inhibitor Screening Assay Kit was purchased from BPS Bioscience (San Diego, CA). Mouse TNF-α ELISA kit was bought from ThermoFisher (Waltham, MA).

We used the ACE2:SARS-CoV-2 Spike inhibitor screening assay kit (BPS Bioscience, San Diego, CA) to examine the effectiveness of different fruit components (prenol, vitamin C, cyanidin, and rutin) and structural analogues of prenol (methyl butane and HDMB) in inhibiting the binding between SARS-CoV-2 spike S1 and ACE2. It was performed following manufacturer’s instructions as described by us recently (17, 18). In brief, this is an Amplified Luminescent Proximity Homogenous Assay Screen (AlphaScreen)-based assay in which a 96-well nickel-treated plate provided by the manufacturer was coated with ACE2 solution. After washing with immuno buffer and treatment with blocking buffer, different concentrations of prenol, vitamin C, cyanidin, rutin, methyl butane, and HDMB were added to each well followed by addition of SARS-CoV-2 Spike (RBD)-Fc. After incubation with blocking buffer, plates were treated with anti-mouse Fc-HRP followed by addition of an HRP substrate. A Perkin Elmer multimode microplate reader, Victor X5, was used to monitor the resultant chemiluminescence.

To study the interaction among SARS-CoV-2 spike S1, hACE2, and prenol, we performed in silico structural analysis as described earlier (17, 19, 20). In brief, by using the protein preparation tools from the Schrodinger platform, at first, we assessed the quality of the crystal structure of hACE2 and SARS-CoV-2 spike S1 followed by addition of hydrogens to the hydrogen bond orientation, charges, missing atoms, and side chains of the different residues of both proteins. The three-dimensional structure of prenol was attained from Zinc database. Finally, the complex structure was subjected to energy minimization in the Optimized Potential for Liquid Simulations (OPLS3) force field to make it torsion free. We applied the dynamic hydrogen bonding module for finding potential hydrogen bonds among the structures. Hydrophobic interactions were also evaluated.

Thermal shift assay was conducted using the SYBR green real-time melting strategy following the thermal shift dye kit (Life Technologies) in Applied Biosystems 7500 standard real-time thermal cycler. Each reaction contained 250 ng of either rSARS-CoV-2 spike S1 or hACE2 protein in the presence of 10 µM prenol or vitamin C, 10 μl thermal shift buffer, and 2 μl of dye as described earlier (20–22). Thermal shift reaction was performed in the dark. The 96-well PCR plate was loaded in the thermal cycler for the following two-stage program (1 cycle: 25°C for 2 min; 70 cycles: 27°C for 15 s, 26°C for 1 min; autoincrement 1°C for both stages). The filter was set at ROX with no quencher filter and no passive filter. ΔTm was calculated in GraphPad Prism 8.4.3 (GraphPad Software, La Jolla, CA).

hACE2-expressing HEK293 cells were plated in 24-well plates at 70–80% confluence followed by infection with lentiviral particles pseudotyped with either SARS-CoV-2 spike S1 or VSV-G at a multiplicity of infection of 0.5 as described previously (23). These pseudoviruses (GeneCopoeia) contained Luciferase and GFP reporter genes that helped us to quantify virus entry into ACE2-expressing 293T cells just by monitoring either luciferase activity or GFP fluorescence. Cells were also infected with empty lentiviral (lenti-naked) particles as control. After 48 h of infection, the entry of pseudovirus into HEK293 cells was checked by either luciferase activity or GFP fluorescence. Luciferase activity was assayed in total cell extracts using the Luciferase Assay System (Promega) in a TD-20/20 Luminometer (Turner Designs, Sunnyvale, CA) as described previously (24, 25). GFP fluorescence was monitored in an Olympus BX-41 fluorescence microscope as described previously (16, 18, 26).

Microarray analysis was performed as described earlier (21, 27). In brief, RNA samples were collected from A549 cells using Qiagen RNeasy kit (catalog number [Cat #] 74104; Germantown, MD). Quantity and purity of RNA were determined using the NanoDrop LTE (NanoDrop Technologies, Wilmington, DE). The mRNA of each sample was converted into cDNA with SuperScript III First-Strand synthesis Kit (Cat # 18080-051; ThermoFisher) followed by microarray analysis using ExProfile Human Inflammatory Response and Autoimmunity Related Gene qPCR Array (GeneCopoeia). Once PCR is done, cycle threshold (Ct) values were imported from the PCR console, and we used online software modules to proceed with further calculations. Data normalization was performed by correcting all Ct values with the average Ct values of constantly expressed housekeeping genes present on the array. PCR-array results were displayed by clustergram analyses with three-color presentation from red (high expression) to white to blue (low expression).

Total RNA was isolated from human A549 lung cells and mouse lung tissues using RNeasy Mini kit (Qiagen) and Ultraspec-II RNA reagent (Biotecx Laboratories, Houston, TX), respectively. To remove any contaminating genomic DNA, we digested total RNA with DNase. Real-time PCR was performed in the ABI-Prism7700 sequence detection system (Applied Biosystems, Foster City, CA) using the SYBR green real-time kit obtained from QuantaBio (Beverly, MA) as described before (28–30). The following primer sequences were used: TNF-α: sense, 5′-TTCTGTCTACTGAACTTCGGGGTGATCGGTCC-3′, antisense, 5′-GTATGAGATAGCAAATCGGCTGACGGTGTGGG-3′; IL-1β: sense, 5′-GGATATGGAGCAACAAGTGG-3′, antisense, 5′-ATGTACCAGTTGGGGAACT-3′; IL-6, sense, 5′-GACAACTTTGGCATTGTGG-3′; antisense, 5′-ATGCAGGGATGATGTTCTG-3′; and GAPDH: sense, 5′-GGTGAAGGTCGGTGTGAACG-3′, antisense, 5′-TTGGCTCCACCCTTCAAGGTG-3′.

A549 cells plated at 60–70% confluence were transfected with different reporter constructs (e.g., pNF-κB-Luc, pIL-1b promoter-Luc, pIL-6 promoter-Luc) using Lipofectamine Plus (Life Technologies). After 24 h of transfection, cells were incubated with different concentrations of prenol and vitamin C for 10 min and then exposed to SARS-CoV-2 spike S1 under serum-free conditions for 4 h. Luciferase activities were analyzed in cell extracts using the Luciferase Assay System kit (Promega) in a TD-20/20 Luminometer (Turner Designs) as described previously (31, 32).

Western blotting was performed as previously described (33, 34). Equal amounts of proteins were electrophoresed in 8% (for monitoring τ level) or 10% SDS-PAGE and transferred onto the nitrocellulose membrane. The blot was probed with primary Abs against IL-1β (dilution 1:200) and Actin (1:1000) overnight at 4°C. The next day, primary Abs were removed, and the blots were washed with PBS containing 0.1% Tween 20 and then probed with corresponding infrared fluorophore-tagged secondary Abs (1:10,000; Jackson Immunoresearch) and kept at room temperature for 1 h. Finally, blots were scanned with an Odyssey infrared scanner (LI-COR Biosciences, Lincoln, NE). Band intensities were quantified using ImageJ software (35, 36).

Nuclear extracts were prepared, and EMSA was performed as described previously (19, 25, 37) with minor modifications. In brief, IRDye infrared dye end-labeled oligonucleotides containing the consensus binding sequence for NF-κB were purchased from LI-COR Biosciences. Nuclear extract (6 µg) was incubated with binding buffer and with IRDye-labeled NF-κB probe for 20 min followed by separation on a 6% polyacrylamide gel in 0.25× Tris borate-EDTA buffer and analysis by the Odyssey Infrared Imaging System (LI-COR Biosciences).

Mice were maintained and experiments conducted in accordance with National Institutes of Health guidelines and were approved by the Rush University Medical Center Institutional Animal Care and Use Committee. C57/BL6 mice (8–10 wk old; Envigo) of both sexes were intoxicated with rSARS-CoV-2 spike S1 protein (50 ng/mouse/d) intranasally as described by us recently (16–18). Concisely, recombinant spike S1 was dissolved in 2 μl normal saline, and mice were held in supine position to deliver 1 µl volume into each nostril using a Pipetman. Control mice received only 2 µl saline.

Prenol was mixed in 0.1% methylcellulose starting from 5 d of SARS-CoV-2 spike S1 intoxication, and C57/BL6 mice (8–10 wk old; Envigo) of both sexes were treated orally with prenol (50 mg/kg body weight/d) via gavage for 10 d.

Electrocardiographic (ECG) recording was performed as described by us recently (17, 18). In brief, mice were acclimatized to the ECG pulse transducer pad (AD instruments TN 012/ST) and the experimental housing conditions before ECG recording. ECG pulse transducer pad was placed around the heart of each animal, and ECG recording was carried out for 120s. For ECG analysis, ECG data were exported from the Labchart pro, version 8.0 (Power Lab 4/35 model) as raw data format, and the digital signal processing was performed using this software. The recording was conducted for 120 s, the ECG signals were recorded at a sampling range of 20 mV with 4 beats/s sampling rate, and different ECG parameters were calculated as described previously (17). The recording and analysis settings were kept the same for all experimental mice included in this study.

TNF-α ELISA was carried out in mouse serum as described earlier (24, 38, 39) using an assay kit (eBioscience) according to the manufacturer’s instructions.

Statistical analyses were performed using GraphPad Prism 8.0. Mouse behavioral measures were examined by an independent one-way ANOVA using SPSS. Homogeneity of variance between test groups was examined using Levene's test. Post hoc analyses were conducted using Tukey’s tests. Other data were expressed as means ± SD of three independent experiments. Statistical differences between means were calculated by Student t test (two-tailed). A p value <0.05 was considered statistically significant.

Fruit consumption is known to boost immunity and protect humans from different infections. Because SARS-CoV-2 binds to its receptor ACE2 to infect host cells, we screened major components of fruit based on the inhibition of interaction between ACE2 and the RBD of SARS-CoV-2 spike S1 (18, 40). We used an AlphaScreen-based ACE2:SARS-CoV-2 spike S1 binding assay (Cat #79936; BPS Bioscience) as described by us recently (16, 17). Expectedly, SARS-CoV-2 spike S1 binds to immobilized ACE2, which was inhibited by different doses of prenol (Fig. 1A). To understand whether prenol interferes with S1-to-ACE2 binding through competitive binding or through irreversible conformational changes, we washed out prenol from the AlphaScreen assay. Interestingly, prenol washout was capable of recovering S1-to-ACE2 binding (Supplemental Fig. 1A), indicating competitive inhibition by prenol. In contrast, vitamin C and other fruit components (cyanidin and rutin) could not decrease the binding between SARS-CoV-2 spike S1 and ACE2 (Fig. 1A), indicating the specificity of the effect. Because prenol is known as isopentenol, to understand the structural requirement, we also tested the effect of isopentane or 2-methyl butane (Supplemental Fig. 1B–D). However, in contrast with isopentenol, isopentane had no effect on the association between spike S1 and ACE2 (Fig. 1A). Again, HDMB, a compound with structural similarity with prenol, but containing the hydroxyl group at a different position, two methyl substitutions at a different position, and a carboxylic acid group at the end (Supplemental Fig. 1C, 1D), also did not hinder the association between SARS-CoV-2 spike S1 and ACE2 (Fig. 1A), suggesting that a particular structure is necessary to detach the binding of spike S1 with ACE2.

To dissociate the binding between spike S1 and ACE2, prenol should bind to spike S1, ACE2, or both. To delineate whether prenol binds to spike S1 and/or ACE2, we used thermal shift assay, an important tool for monitoring the interaction between ligand and its receptor. The melting pattern of full-length spike S1 and hACE2 was checked using SYBR Green reaction at 27–94°C as described previously (20, 21). It is clearly evident from characteristic sigmoidal melting curves that recombinant spike S1 (Fig. 1B) and hACE2 (Fig. 1C) were conformationally stable. However, thermal shift assay demonstrated that prenol was capable of strongly shifting the melting curve of spike S1 by 5°C (Fig. 1B). In contrast, prenol displayed a thermal shift of only 1.13°C with hACE2 protein (Fig. 1C). Similar to the inability of vitamin C to inhibit the binding between spike S1 and ACE2, this molecule also exhibited a thermal shift of only 0.9°C with spike S1 (Fig. 1D). These results suggest that prenol, but not vitamin C, binds to SARS-CoV-2 spike S1.

To confirm again that prenol binds to SARS-CoV-2 spike S1, not ACE2, we used an in silico analysis. To model the interaction between RBD of spike S1 and ACE2 in the absence or presence of prenol, we applied a rigid-body protein–protein interaction tool. In the absence of prenol, several residues (Asp30, Asp38, Gln42, and Lys353) of ACE2 networked with various residues of spike S1 (Lys417, Tyr449, Gly496, Asn501, and Tyr505) (Fig. 1E). However, similar to that found in thermal shift assay, prenol exhibited association with spike S1, not ACE2 (Fig. 1F). Interestingly, prenol cooperated with Lys417 residue of spike S1 that basically associated with Asp30 of ACE2 by a salt bridge (ionic bond) in the absence of prenol (Fig. 1F). Therefore, in the absence of prenol, that ionic bond was destroyed and Lys417 displayed an altered rotameric pose to result in dissociation between spike S1 and ACE2 (Fig. 1F).

Because prenol, not vitamin C, suppressed the binding between SARS-CoV-2 spike S1 and ACE2, next we investigated the effect of both prenol and vitamin C on viral entry. Pseudoviruses are appropriate for monitoring viral entry, because they allow viral entry to be distinguished from other virus life-cycle events. Therefore, we used lentiviral particles pseudotyped with the SARS-CoV-2 spike S1 protein. Because human embryonic kidney 293 (HEK293) cells do not have any detectable ACE2 receptors, we used HEK293 cells expressing hACE2 to monitor the entry of pseudo–SARS-CoV-2. In pseudovirus luciferase assay, viral entry into cells correlates to the levels of luciferase signals in the cells. As expected, lenti-naked infection did not increase luciferase signals in hACE2-expressing HEK293 cells (Fig. 2A). In contrast, marked increase in luciferase activity was found in pseudo–SARS-CoV-2–infected cells (Fig. 2A), representing the entry of pseudo–SARS-CoV-2 into hACE2-HEK293 cells. However, prenol treatment markedly repressed pseudo–SARS-CoV-2–induced luciferase activity (Fig. 2A). In contrast, vitamin C had no inhibitory effect on luciferase activity driven by pseudo–SARS-CoV-2 (Fig. 2G), suggesting that unlike prenol, vitamin C is unable to inhibit the entry of pseudo–SARS-CoV-2 into hACE2-HEK293 cells.

To further confirm prenol-mediated suppression of viral entry, we monitored GFP fluorescence, which is also a measure of infection in this assay. We observed marked GFP expression in hACE2-expressing HEK293 cells that were infected with SARS-CoV-2 pseudovirus (Fig. 2C, 2I), but not lenti-naked viral particles (Fig. 2C, 2I). However, prenol strongly inhibited pseudo–SARS-CoV-2–induced GFP expression (Fig. 2C, 2E). In contrast, vitamin C had no effect on GFP expression in pseudo–SARS-CoV-2–infected hACE2-HEK293 cells (Fig. 2I, 2K). Together, these results suggest that prenol, but not vitamin C, is capable of inhibiting the entry of pseudo–SARS-CoV-2 in human cells.

Although VSV, a zoonotic arbovirus of the Rhabdoviridae family, primarily affects animals (e.g., horses, cattle, swine, etc.), humans can be also infected with VSV while handling infected animals. Therefore, to understand whether the inhibitory effect of prenol on the entry of pseudo–SARS-CoV-2 into hACE2-HEK293 cells is specific, we examined the effect of both vitamin C and prenol on the entry of pseudo-VSV. Similar to pseudo–SARS-CoV-2, infection with pseudo-VSV also led to a marked increase in luciferase activity in hACE2-HEK293 cells (Fig. 2B). However, in contrast with the inhibition of pseudo–SARS-CoV-2–induced luciferase activity, prenol remained unable to suppress pseudo-VSV–induced luciferase activity in hACE2-HEK293 cells (Fig. 2B). Although vitamin C did not inhibit pseudo–SARS-CoV-2–induced luciferase activity, interestingly, we observed strong inhibition of luciferase activity in pseudo-VSV–infected hACE2-HEK293 cells (Fig. 2H). We also monitored GFP fluorescence and found marked GFP expression in cells infected with VSV pseudovirus, but not lenti-naked viral particles (Fig. 2J). Although prenol had no effect on VSV pseudovirus–induced GFP expression in hACE2-HEK293 cells (Fig. 2D, 2F), consistent with the inhibition of luciferase activity, vitamin C inhibited GFP expression in VSV pseudovirus–infected cells (Fig. 2J, 2L). These results suggest that vitamin C, but not prenol, is capable of inhibiting the entry of pseudo-VSV into hACE2-HEK293 cells.

Cytokine storm is associated with a severe symptom of acute respiratory distress syndrome found in some COVID-19 patients (41). Recently, we have demonstrated that SARS-CoV-2 spike S1 induces the expression of proinflammatory cytokines in A549 lung cells and in vivo in mouse lungs (17, 18). However, prenol strongly inhibited spike S1–mediated mRNA expression of IL-1β and TNF-α in A549 cells at different concentrations (5 and 10 µM) tested (Fig. 3A, 3B) without altering the viability of cells as monitored by lactate dehydrogenase (LDH) release (Supplemental Fig. 2A), suggesting that prenol-mediated inhibition of IL-1β and TNF-α mRNAs in spike S1–stimulated A549 cells is not dependent on any change in cell viability. Moreover, prenol may inhibit the expression of proinflammatory cytokines via suppression of ACE2, not disrupting the interaction between spike S1 and ACE2 in A549 cells. Therefore, we examined the effect of prenol on the expression of ACE2 in A549 cells and found no effect of prenol on the level of ACE2 (Supplemental Fig. 2B, 2C), indicating that prenol-mediated anti-inflammatory effect in spike S1–stimulated A549 cells is not due to any change in ACE2.

In contrast with strong inhibition of spike S1–mediated mRNA expression of IL-1β and TNF-α by prenol at both 5 and 10 µM concentrations, vitamin C at the same doses remained unable to inhibit the level of IL-1β and TNF-α mRNAs in A549 cells (Fig. 3A, 3B).

NF-κB is a proinflammatory transcription factor (42), and activation of NF-κB is necessary for the expression of most of the proinflammatory molecules in various cell types. Recently, we have demonstrated that rSARS-CoV-2 spike S1 induces the activation of NF-κB in A549 cells (17, 18). Therefore, to understand the mechanism of prenol- and vitamin C–mediated modulation of proinflammatory molecules in spike S1–challenged A549 cells, we monitored DNA-binding and transcriptional activities of NF-κB. As obvious by EMSA, recombinant spike S1 stimulated the DNA-binding activity of NF-κB in A549 cells (Fig. 3C). However, prenol, but not vitamin C, decreased spike S1–induced DNA-binding activity of NF-κB (Fig. 3C). To confirm these results, we also monitored NF-κB–dependent luciferase reporter activation and found induction of NF-κB–dependent luciferase activity by SARS-CoV-2 spike S1 in A549 cells (Fig. 3D). Whereas prenol treatment strongly inhibited NF-κB–dependent reporter activity, mild stimulation of spike S1–induced NF-κB activation was observed after vitamin C treatment (Fig. 3D). These results suggest that vitamin C is unable to suppress SARS-CoV-2 spike S1–induced inflammatory events.

Because prenol inhibited SARS-CoV-2 spike S1–induced expression of TNF-α and IL-1β, we examined the anti-inflammatory effect of prenol in more detail in spike S1–challenged A549 lung cells. Therefore, we performed a targeted gene array (GeneCopoeia) to monitor human inflammatory response and autoimmunity-related gene expression profile. Microarray followed by heatmap analyses (Fig. 4A) clearly demonstrated that in addition to the upregulation of TNF-α and IL-1β, spike S1 treatment increased the expression of a number other genes (e.g., TLR5, TLR2, IL-6, IL-18, IL-22, CD40, CXCL1, CXCL3, CCL16, CCL7, CCR1, CXCR4, NOS2) in A549 cells that are involved in inflammation and immunomodulation. However, prenol treatment strongly suppressed the expression of these and other proinflammatory molecules in spike S1–challenged A549 cells (Fig. 4A, 4B). Individual real-time PCR also confirmed the increase in TNF-α (Fig. 4C), IL-1β (Fig. 4D), and IL-6 (Fig. 4E) mRNAs in spike S1–stimulated A549 cells, which were inhibited by prenol. To further confirm, we also monitored promoter activation and found an increase in luciferase activities driven by TNF-α (Fig. 4F), IL-1β (Fig. 4G), and IL-6 (Fig. 4H) gene promoters in spike S1–stimulated A549 cells. However, prenol suppressed TNF-α–, IL-1β–, and IL-6–driven reporter activation in spike S1–challenged cells (Fig. 4F–H). Immunoblot analysis also revealed an increase in IL-1β protein by SARS-CoV-2 spike S1 and its reversal by prenol treatment (Fig. 4I, 4J). Expectedly, bacterial LPS also induced the activation of IL-6 promoter (Supplemental Fig. 3A) and IL-1β promoter (Supplemental Fig. 3B) in A549 cells. However, prenol remained unable to inhibit luciferase activity driven by IL-6 promoter (Supplemental Fig. 3A) and IL-1β promoter (Supplemental Fig. 3B) in LPS-stimulated A549 cells, indicating that the anti-inflammatory effect of prenol is specific for spike S1.

Whereas CCL13 is an antimicrobial protein with bactericidal activity (43), CCL21 is a chemokine ligand capable of reversing tumor-mediated immune suppression (44). Interestingly, we found that prenol was capable of increasing the expression of both CCL13 and CCL21 in spike S1–stimulated A549 cells (Fig. 4A, 4B), indicating a broad-spectrum immunomodulatory effect of prenol in SARS-CoV-2 spike S1–challenged cells.

It is known that mutations in the viral genome are capable of altering the virus's pathogenic potential. Even an exchange of a single amino acid may affect a virus's ability to escape the immune system and confuse the vaccine development progress against the virus. Consequently, since the beginning of SARS-CoV-2 in December 2019, variants have been emerging at a regular interval. Most of the SARS-CoV-2 variants have mutations in spike glycoprotein, which is the primary focus of different vaccines (45, 46). Therefore, it is seen that vaccines are not equally efficient against different SARS-CoV-2 variants (46). Thus, it is of highest importance to study these variants and associated mutations. Until now, 12 different important variants have been identified: B.1.1.7 (Alpha), B.1.351 (Beta), B.1.525 (Eta), B.1.427/B.1.429 (Epsilon), B.1.526 (Iota), B.1.617.1 (Kappa), B.1.617.2 (Delta), C.37 (Lambda), P.1 (Gamma), P.2 (Zeta), P.3 (Theta), and the recently discovered B.1.1.529 (Omicron). Because many variants, including Delta and Omicron, evade the protection provided by available COVID-19 vaccines, we wanted to delineate whether prenol exhibited any efficacy against Delta variant SARS-CoV-2 spike S1, Alpha (N501Y) variant SARS-CoV-2 spike S1, Beta (E484K) variant SARS-CoV-2 spike S1, and Omicron variant SARS-CoV-2 spike S1 in human A549 lung cells. We found that regular SARS-CoV-2 spike S1 and Delta variant SARS-CoV-2 spike S1 were almost equally effective in inducing the expression of IL-1β mRNA and human IL-1β promoter-driven luciferase activity in A549 cells (Supplemental Fig. 3C, 3D). However, N501Y variant SARS-CoV-2 spike S1, E484K variant SARS-CoV-2 spike S1, and Omicron variant SARS-CoV-2 spike S1 were less efficient than the Delta variant SARS-CoV-2 spike S1 in upregulating the expression of IL-1β mRNA and increasing the transcriptional activity of human IL-1β promoter in A549 cells (Supplemental Fig. 3C, 3D).

Next, we examined whether prenol could suppress the mRNA expression of proinflammatory cytokines induced by spike S1 of different SARS-CoV-2 variants. As described earlier, Delta variant SARS-CoV-2 spike S1 (Fig. 5A, 5B), N501Y variant SARS-CoV-2 spike S1 (Fig. 5C, 5D), and E484K variant SARS-CoV-2 spike S1 (Fig. 5E, 5F) induced the mRNA expression of IL-6 (Fig. 5A, 5C, 5E) and IL-1β (Fig. 5B, 5D, 5F) in A549 cells. Similarly, Omicron variant SARS-CoV-2 spike S1 also increased the mRNA expression of TNF-α (Fig. 5M) and IL-1β (Fig. 5N) in A549 cells. However, prenol treatment markedly inhibited the mRNA expression of IL-6 (Fig. 5A, 5C, 5E) and IL-1β (Fig. 5B, 5D, 5F) in A549 cells induced by Delta variant SARS-CoV-2 spike S1 (Fig. 5A, 5B), N501Y variant SARS-CoV-2 spike S1 (Fig. 5C, 5D), and E484K variant SARS-CoV-2 spike S1 (Fig. 5E, 5F). Prenol was also capable of inhibiting the mRNA expression of TNF-α (Fig. 5M) and IL-1β (Fig. 5N) in Omicron variant SARS-CoV-2 spike S1–stimulated A549 cells. Recently, we have demonstrated that spike S1–interacting domain of ACE2 receptor (SPIDAR) peptide inhibits SARS-CoV-2 spike S1–induced expression of proinflammatory cytokines in A549 cells (18). Therefore, in this study, for comparison, we also included wild type spike S1-interacting domain of ACE2 receptor (wtSPIDAR) peptide. Similar to prenol, wtSPIDAR was also capable of suppressing the mRNA expression of proinflammatory molecules in A549 cells stimulated by Delta variant SARS-CoV-2 spike S1 (Fig. 5A, 5B), N501Y variant SARS-CoV-2 spike S1 (Fig. 5C, 5D), E484K variant SARS-CoV-2 spike S1 (Fig. 5E, 5F), and Omicron variant SARS-CoV-2 spike S1 (Fig. 5M, 5N). Next, we investigated whether prenol and wtSPIDAR were capable of inhibiting the activation of the proinflammatory cytokine gene promoter in A549 cells stimulated with spike S1 of a different SARS-CoV-2. Similar to the inhibition of mRNA expression, both prenol and wtSPIDAR reduced luciferase activities driven by IL-6 promoter and IL-1β promoter in A549 cells that were stimulated by Delta variant SARS-CoV-2 spike S1 (Fig. 5G, 5H), N501Y variant SARS-CoV-2 spike S1 (Fig. 5I, 5J), and E484K variant SARS-CoV-2 spike S1 (Fig. 5K, 5L). Accordingly, prenol and wtSPIDAR also inhibited Omicron variant spike S1–mediated activation of TNF-α promoter (Fig. 5O) and IL-1β promoter (Fig. 5P) in A549 cells. These results suggest that spike S1 of different SARS-CoV-2 variants induces the expression of proinflammatory cytokines in human A549 lung cells, and that both prenol and wtSPIDAR are capable of inhibiting this proinflammatory event.

Although SARS-CoV-2 does not easily interact to ACE2 and infect normal mice, we demonstrated fever and important cardiac and respiratory symptoms of COVID-19 in normal C57/BL6 mice by intranasal intoxication of SARS-CoV-2 spike S1 (17, 18). However, boiled rSARS-CoV-2 spike S1 remains unable to induce fever and other symptoms of COVID-19 in mice (18). Moreover, anti–SARS-CoV-2 spike S1 Ab neutralizes spike S1–mediated increase in body temperature and other symptoms in mice (18), indicating the involvement of SARS-CoV-2 spike S1 protein in causing these COVID-19 symptoms in mice. Hence we examined whether oral prenol could reduce these symptoms in mice. Usually, COVID-19 patients receive drugs after diagnosis of the disease. Therefore, we studied whether prenol being administered from 5 d after initiation of the disease (Fig. 6A) was still capable of guarding mice from COVID-19–related complications. We used the 5-d window because all SARS-CoV-2 spike S1–insulted mice displayed a body temperature of around 100°F on 5 d of intoxication (Fig. 6B). As reported earlier (17, 18), intranasal challenge with rSARS-CoV-2 spike S1 (Fig. 6A) resulted in marked upregulation of the expression of TNF-α (Fig. 6C), IL-1β (Fig. 6D), and IL-6 (Fig. 6E) in vivo in the lung of C57/BL6 mice. Consistent with that found in A549 cells, orally administered prenol strongly suppressed the levels of TNF-α (Fig. 6C), IL-1β (Fig. 6D), and IL-6 (Fig. 6E) mRNAs in the lungs of SARS-CoV-2 spike S1–intoxicated mice. Accordingly, prenol also decreased the level of TNF-α in serum of SARS-CoV-2 spike S1–insulted mice (Fig. 6F). Fever is possibly one of the noticeable symptoms of COVID-19 (12, 13), and orally administered prenol also decreased body temperature of SARS-CoV-2 spike S1–intoxicated mice (Fig. 6G).

Some of the cardiac-related problems of COVID-19 could be modeled in SARS-CoV-2 spike S1–intoxicated mice (17, 18). To characterize SARS-CoV-2 spike S1–mediated abnormal heart function in mice, we performed the following: first, we examined the efficacy of heat-denatured SARS-CoV-2 spike S1 in modulating ECG in C57/BL6 mice. SARS-CoV-2 spike S1 intoxication led to cardiac arrhythmias in mice as apparent from abnormal ECG (Supplemental Fig. 4A, 4B), increase in heart rate (Supplemental Fig. 4F), decrease in QRS interval (Supplemental Fig. 4G), heart rate variability (Supplemental Fig. 4H) and QT interval (Supplemental Fig. 4I), and increase in RR interval (Supplemental Fig. 4J). However, boiled or heat-inactivated rSARS-CoV-2 spike S1 remained unable to do that (Supplemental Fig. 4A–C, 4F–I), suggesting that the induction of arrhythmias in mice is due to native SARS-CoV-2 spike S1 protein. Second, anti-spike S1 Ab neutralized cardiac arrhythmias in SARS-CoV-2 spike S1–challenged C57/BL6 mice (Supplemental Fig. 4A–C, 4F–I). This result was specific because control IgG remained unable to reduce cardiac arrhythmias in spike S1–challenged mice (Supplemental Fig. 4A–C, 4F–I). Therefore, we examined whether oral prenol inhibited arrhythmias in these mice. Interestingly, prenol treatment regularized electrical activity of the heart (Fig. 7A–D) and stabilized heart rate (Fig. 7E), heart rate variability (Fig. 7F), QRS interval (Fig. 7G), QT interval (Fig. 7H), and RR interval (Fig. 7I) in SARS-CoV-2 spike S1–intoxicated mice. As reported earlier (17, 18), the level of serum LDH was also markedly higher in SARS-CoV-2 spike S1–intoxicated mice than in normal mice (Fig. 7J). Nonetheless, oral prenol stabilized serum LDH in SARS-CoV-2 spike S1–insulted mice (Fig. 7J).

Recently, we have demonstrated that SARS-CoV-2 spike S1 insult causes functional discrepancies in C57/BL6 mice (16–18). Similarly, we found an overall decrease in locomotor activities in SARS-CoV-2 spike S1–challenged mice (Fig. 8A–G). However, prenol treatment improved and/or normalized overall locomotor activities as apparent by heatmap (Fig. 8A), distance traveled (Fig. 8B), velocity (Fig. 8C), cumulative duration (Fig. 8D), center zone frequency (Fig. 8E), center zone cumulative duration (Fig. 8F), and rotarod latency (Fig. 8G). This is a natural alcohol present in fruit and accordingly, we did not observe any drug-related side effect (e.g., hair loss, appetite loss, weight loss, untoward infection, irritation) in any mouse on treatment with oral gavage with prenol.

It has been shown that the vitamin C found in different fruits makes our immune system strong to help us fight against cold, flu, and all sorts of infection (47). Intake of fruits and vitamin C is also known to accelerate recovery from different diseases and contribute to general well-being. However, in a systematic review and meta-analysis of randomized controlled trials involving 572 patients, Rawat et al. (48) did not find any significant benefit with vitamin C for COVID-19. In another randomized, controlled, clinical trial with 56 critical COVID-19 patients involving three different hospitals in Hubei, China, high-dose i.v. vitamin C also remained unsuccessful (5). Because fruits are the richest sources of vitamin C, these results suggest that vitamin C and fruit may not have much beneficial effect in COVID-19. To address this issue mechanistically, we examined whether vitamin C and other major components of fruit inhibit the binding of SARS-CoV-2 spike S1 with ACE2. Interestingly, we found that prenol, but not other components (vitamin C, cyanidin, and rutin) of fruit tested, decreased the binding between spike S1 and ACE2. Other compounds with structures similar to prenol also did not inhibit the association between spike S1 and ACE2, indicating the specificity of the effect. The binding of SARS-CoV-2 spike S1 with ACE2 is needed for the virus to enter into host cells. Consistent with the modulation of spike S1-to-ACE2 association, prenol, but not vitamin C, inhibited the entry of pseudotyped SARS2 to hACE2-expressing HEK293 cells. Prenol known as fruit alcohol is naturally available in different fruits, and berries are particularly rich in prenol. These results suggest prenol may have therapeutic importance for COVID-19. Moreover, our studies also suggest that vitamin C may not have a direct effect on COVID-19, and that consumption of whole fruit may be beneficial for fighting against COVID-19 because it contains prenol.

To inhibit the interaction between ACE2 and SARS-CoV-2 spike S1, a molecule should bind to either ACE2 or spike S1. ACE2 is an important molecule for the modulation of blood pressure and hypertension because the main role of ACE2 is to convert a vasoconstrictor (angiotensin II) to a vasodilator (angiotensin 1–7) (9, 10). Because ACE2 is predominant in lung, heart, and kidney (10), SARS-CoV-2 easily infects lung, heart, and kidney (10), causing multiorgan failure in severe COVID-19 cases because ACE2 is predominant in these organs. Although inhibition of ACE2 would reduce the SARS-CoV-2 infection and associated complications, ACE2 is a beneficial molecule; therefore, inhibition of ACE2 should not be a valid therapeutic option for COVID-19. Accordingly, from a therapeutic approach, we concentrated on finding a molecule that will bind to SARS-CoV-2 spike S1, not ACE2, to inhibit the association between spike S1 and ACE2. Our biophysical assays and in silico structural analysis clearly demonstrated that prenol bound to spike S1, but not ACE2, to inhibit the association between ACE2 and spike S1. It is important because without disturbing the function of ACE2, prenol should not cause any problem for COVID-19 patients with pre-existing pulmonary, cardiovascular, and kidney issues. In this context, it is important to reiterate that vitamin C neither associated to spike S1 nor inhibited the binding between spike S1 and ACE2, at least providing a rationale why vitamin C did not exhibit any significant benefit for COVID-19 in different clinical trials.

Another important aspect of our approach is the focus on identifying a molecule that can target specifically the ACE2-interacting domain of spike S1. This is because SARS-CoV-2 variants are not expected to make many mutations at the ACE2-binding region. If a variant causes several mutations at the ACE2 binding site, it could be a blessing for humans because the resultant variant might not be able to bind to ACE2 and infect ACE2-expressing human cells. Alternatively, there is another possibility where a variant can cause mutation(s) in such a way that binding would be stronger, resulting in greater infection. Probably, this happened with the N501Y variant. However, prenol is able to suppress the expression of proinflammatory molecules in A549 cells stimulated by N501Y and other variants (N501Y, E484K, Omicron, and Delta) of SARS-CoV-2. Under this circumstance, it is important to mention although available vaccines are targeting SARS-CoV-2 spike S1, fruit alcohol prenol is pointing toward the ACE2-binding domain of SARS-CoV-2 spike S1. Because the variants of SARS-CoV-2 mostly made mutations at other regions of the spike protein, resulting in the modulation of structure and/or biophysical properties of spike S1, variants try to evade the protection provided by vaccine. Therefore, although available vaccines do not protect against different variants, prenol having a different mode of action may be effective against different variants of SARS-CoV-2.

Until now, a few drugs (e.g., remdesivir, dexamethasone, baricitinib, and tocilizumab) have been approved by the Food and Drug Administration for COVID-19 (49–53). However, prenol has several advantages over these. For example, remdesivir that is being repurposed from HIV to COVID-19 for emergency use has many negative side effects, including hypotension, hypertension, tachycardia, bradycardia, hypoxia, fever, dyspnea, wheezing, angioedema, rash, and nausea, among others. Remdesivir is also not recommended for severe COVID-19 patients suffering from kidney problems. Tocilizumab is known to cause chest pain, cough, difficulty breathing, and formation of bloody urine, among others. Similarly, baricitinib also causes chest pain, cough, and difficulty breathing, among others (54). Although until now no study has been performed on prenol, it is naturally present in fruits, and there are no indications that fruit consumption causes any health issues in humans. Moreover, prenol should be economical in comparison with costly remdesivir. Dexamethasone being a corticosteroid is a broad-spectrum anti-inflammatory drug. Therefore, it often causes immunosuppression, which may not be beneficial for COVID-19. In contrast, prenol specifically inhibits the entry of pseudo–SARS-CoV-2 into cells and the expression of proinflammatory molecules induced by SARS-CoV-2 spike S1, suggesting that prenol should not cause immunosuppression.

In summary, to our knowledge, we present the first evidence that prenol, a fruit alcohol, inhibits the binding of SARS-CoV-2 spike S1 with ACE2, stops the entry of pseudo–SARS-CoV-2 into HEK293 cells, and attenuates the expression of proinflammatory molecules in human A549 cells induced by spike S1 of different variants of SARS-CoV-2. In contrast, vitamin C, the celebrity component of fruit, remains unable to exhibit these properties. Moreover, orally administered prenol inhibits lung inflammation, normalizes heart functions, reduces fever, decreases serum LDH, and improves locomotor activities in SARS-CoV-2 spike S1–intoxicated mice. These results suggest that prenol and prenol-containing fruits may be more beneficial than vitamin C for COVID-19.

The authors have no financial conflicts of interest.