Juice of Citrullus lanatus var. citroides (wild watermelon) inhibits the entry and propagation of influenza viruses in vitro and in vivo

Abstract Vaccines and various anti‐influenza drugs are clinically used to prevent and treat influenza infections. However, with the antigenic mismatch of vaccines and the emergence of drug‐resistant viral strains, new approaches for treating influenza are warranted. This study focused on natural foods as potential candidates for the development of new treatment options for influenza infections. The screening of plants from the Cucurbitaceae family revealed that the juice of Citrullus lanatus var. citroides (wild watermelon) had the strongest ability to inhibit the replication of influenza virus in Madin–Darby canine kidney cells. The results of a time‐of‐addition assay indicated that wild watermelon juice (WWMJ) inhibits the adsorption and late stages of viral replication, suggesting that WWMJ contains multiple constituents with effective anti‐influenza activity. A viral adsorption analysis showed that WWMJ reduces the amount of viral RNA in the cells at 37°C but not at 4°C, confirming that WWMJ inhibits viral entry into the host cells at 37°C. These results suggest that a mechanism other than the inhibition of viral attachment is involved in the anti‐influenza action of WWMJ, which is perhaps responsible for a reduction in internalization of the virus. Administration of WWMJ into the nasal mucosa of BALB/c mice infected with the A/PR/8/34 mouse‐adapted influenza virus was seen to significantly improve the survival rate. The findings of this study, therefore, demonstrate the anti‐influenza potential of WWMJ in vitro and in vivo, thereby suggesting the candidature of WWMJ as a functional food product that can be used to develop anti‐influenza agents and drugs.


Influenza virus infections spread at epidemic levels worldwide every
year, and sometimes become pandemics. In 2009, a strain of the influenza A virus (H1N1) rapidly spread worldwide, and the global pandemic alert level was raised to phase 5, indicating the sustained human-to-human transmission of a novel influenza strain of animal origin (Fraser, 2009).
Influenza viruses are RNA viruses of the family Orthomyxoviridae.
Type A, B, and C influenza viruses can infect humans, among which types A and B cause seasonal epidemics. Type A influenza viruses have eight single-stranded RNAs that encode 13 proteins (Jagger, 2012), and are classified into subtypes based on the antigenicity of the hemagglutinin (HA) and neuraminidase proteins on the viral membrane (Webster, 1992). Type A influenza causes epidemics every year, resulting in severe respiratory disease symptoms, such as sneezing, sore throat, fever, headache, muscle fatigue, and malaise in infected individuals (García, 2006).
The declining birth rate and aging population of Japan has caused influenza virus infection to be recognized as an increasingly serious health risk. The appearance of resistant viruses and the infection of humans with avian influenza viruses are also considered as noteworthy risks (Dortmans, 2013). In recent years, the emergence of new viruses has been reported; type D Influenza virus, for example, infects not only bovine and porcine bronchial epithelial cells, but also human bronchial epithelial cells, and therefore has the possibility of zoonotic infection (Song, 2016).
Several drugs are used to treat influenza virus infections, including M2 channel blockers (amantadine and rimantadine) and neuraminidase inhibitors (zanamivir and oseltamivir). These drugs inhibit viral growth after the virus has been adsorbed into the host cells (and viral RNA has been released into the cell) and therefore restricts the release of viruses into the extracellular space. However, recent reports have indicated that some influenza A viruses are resistant to M2 channel blockers (Belshe, 1989) and neuraminidase inhibitors (Bloom, 2010;Le, 2005;Seibert, 2012); therefore, new prevention and treatment regimens for influenza virus infections are required.
Based on such previous findings, the present study focused on the anti-influenza effects of functional foods, such as wild watermelon juice. It has been previously reported that adlay tea is effective against influenza virus (Nagai, 2018), and thus we continued to search for more food extracts that are effective against this virus in the present study (Horio, 2020;Nagai, 2019).
Wild watermelon (WWM), Citrullus lanatus var. citroides, which is native to the Kalahari desert in southern Africa, can adapt to and grow under severely dry and high-ultraviolet-light conditions. WWM is used as a dietary source of hydrogen and water in its native region, and its seeds are also known to contain many essential amino acids (Umar, 2013). WWM has a high citrulline content, which protects the plant from the stresses of its native environment (Akashi, 2004;Takahara, 2005;Yokota, 2002). Although there have been several reports of the benefits of WWM, its food functionality is still a relatively new field of research.
In the present study, Madin-Darby canine kidney (MDCK) cells and BALB/c mice infected with the A/PR/8/34 mouse-adapted influenza virus were used to demonstrate the in vitro and in vivo anti-influenza activity of WWM juice (WWMJ). Time-of-addition assay and viral adsorption analysis were used to investigate the inhibition of viral replication and to assess the amount of viral RNA in cells under different temperature conditions, respectively. The focus-forming reduction assay and cell viability tests enabled the determination of viral activity. The findings of this study have emphasized the potential candidature of natural foods, such as WWMJ, as alternate therapeutic options for severe viral infections, such as that caused by the influenza virus.

| Cell lines and viruses
Madin-Darby canine kidney cells were grown in Eagle's minimum es-

| Preparation of watermelon extracts and other sample extracts
Wild watermelon juice and the juice from commercially available watermelon (WMJ) were tested for their anti-influenza activity in this study. WWMJ was provided by Euglena, Co., Ltd. WWMJ was treated at 80°C for 30 min, and the proteins were removed to establish if lectins participated in the antiviral activity. WMJ was squeezed from watermelons obtained from a Hitorijime cultivar produced in Kumamoto and purchased from a supermarket. WWMJ and WMJ were centrifuged at 1,600 ×g and the supernatants freeze-dried. The dry weight of each resulting powder was measured and then dissolved in 20 mg/ ml of ultrapure water. The samples were then sterilized by filtration through a Millex GX membrane with a 24 mm diameter and a pore size of 0.45 µm (Merck Millipore) and stored at − 30°C until analysis. Other sample foods of Cucurbitaceae, winter squash (Cucurbita maxima) and zucchini (Cucurbita pepo L.), were purchased from a supermarket. The samples were cut and freeze-dried. Next, the dried samples (5 g) were powdered, mixed with water (50 ml), and extracted in a hot water bath for 60 min at 80°C. Subsequent steps after the centrifugation were the same as described above.

| Viral yield determination in the presence of watermelon samples
The effects of the addition of food samples on viral yield were determined using a modified version of the previously described procedure (Nagai, 2018). MDCK cells were cultured in a 24-well plate (Thermo Fisher Scientific) at 1 × 10 5 cells/well in 500 µl/well MEM containing 7% FBS and incubated for 24 hr at 37°C. The confluent monolayers of cells were then rinsed twice with serum-free MEM.
Diluted virus culture was incubated with the cells at a MOI of 0.001 for 1 hr at 37°C. The infected cells were rinsed once with serum-free MEM after 1 hr and then cultured in DMEM containing watermelon extract (500 µl/well). Supernatants were collected after 24 hr as the influenza virus samples and used in a focus-forming assay.

| Focus-forming reduction assay (FFRA) of viral activity
The FFRA was performed according to a slightly modified version of the previously described method (Nagai, 2018 andOkuno, 1990). The ethanol was then removed completely. Cells that were not stained immediately were stored at −80°C until staining.
Focus staining was performed by adding 50 µl of murine monoclonal anti-HA antibody [ C179 for A (H1N1) viruses, Okuno, 1993; F49 for A (H3N2) viruses, Ueda et al., 1998; and 7B11 for B viruses, Nakagawa, 1999 ] and a goat antimouse IgG antibody conjugated to horseradish peroxidase (Merck KGaA). The peroxidase reaction was developed for 30 min according to the procedure given by Graham and Karnovsky (1966), using 0.1% H 2 O 2 and 0.3 mg/ml 3,3′-diaminobenzidine tetrahydrochloride (Wako) in phosphate-buffered saline (PBS). Cells were rinsed with water and dried with a hair dryer after the reaction. The numbers of foci in immunostained infected cells were determined under an inverted light microscope.

| Time-of-addition assay
A time-of-addition assay was performed using a modified version of the previously described procedure (Nagai, 2018). MDCK cells were plated in 24-well plates as described above, rinsed twice with serumfree MEM, and then inoculated with A/PR/8/34 (MOI = 0.01). Cells were rinsed twice with serum-free MEM and incubated in DMEM after 1 hr as described above. DMEM containing 1 mg/ml WWMJ, which is approximately ten times the median inhibitory concentration (IC 50 ) (Table 1), was added at different time points: within the 12 hr period prior to infection (-12--1 hr, pretreatment), between 1 hr prior to infection and time of infection (-1-0 hr, adsorption), and between 0-2 hr, 2-4 hr, 4-6 hr, 6-8 hr, or 0-8 hr after infection (replication), as shown in Figure 2a. Cell monolayers were rinsed twice with serum-free MEM after each incubation period and the medium was replaced with fresh medium. Cells were cultured for 8 hr after infection. Infected cells were then frozen at −80°C and subjected to two freeze-thaw cycles before the viral yield was determined with the focus-forming assay.

| Cell viability determination
Cell viability was determined with a Cell Proliferation Kit I (MTT) (F. Hoffmann-La Roche Ltd). The cytopathic effects in the virusinfected cells to which various concentrations of WWMJ had been added were observed under a microscope.

| Hemagglutination inhibition (HI) test
The HI test was conducted using receptor-destroying enzymetreated guinea-pig red blood cells in 96-well U-bottom plates (Thermo Fisher Scientific) with the standard microtiter assay as described previously (Dowdle, 1979).

| Viral adsorption inhibition assay
The amount of virus attached to the cells was determined by measuring the viral RNA encoding the M protein (MP). Viral RNA bound to cells was extracted, and cDNA synthesized and viral RNA quantified as described previously (Nagai, 2018).

| Cell fusion inhibition test
Cell-cell fusion was analyzed using the method previously described (Nagai, 2018). Briefly, the influenza virus solution (A/ PR/8/34: MOI of 0.001) was added to CV-1 cells and cultured for 24 hr. Infected cells were washed twice and incubated for 15 min in DMEM containing 10 µg/ml trypsin. Cells were washed twice and incubated for 30 min in DMEM containing 1 mg/ml WWMJ.
Next, cells were washed and treated with fusion medium to adjust the pH to 5.0 and incubated for 2 min. Thereafter, cells were washed twice and incubated for 3 hr with DMEM containing 2% FBS. Cells were then stained with Giemsa and the number of fused cells counted.   means ± standard deviations (SD). A p-value (p) < .05 was considered statistically significant.

| WWMJ inhibits influenza viral growth
The viral growth in infected MDCK cells treated with WWMJ and WMJ (squeezed from a commercially available WM) was compared.
WWMJ inhibited the proliferation of the influenza virus within cells in a concentration-dependent manner, but WMJ did not cause viral inhibition (Figure 1a). The IC 50 of WWMJ was 0.10 mg/ml. The MTT test revealed that WWMJ was not cytotoxic to MDCK cells ( Figure 1b). Notably, 1 mg/ml WWMJ reduced the virus by 80%-90% as compared to control cells lacking WWMJ. The IC 50s of winter squash and zucchini were 1.07 and 0.74 mg/ml, respectively.
WWMJ had the highest viral growth inhibition activity among plants of Cucurbitaceae that we assayed.

| Antiviral effects of WWMJ against various type A and B influenza viruses
The antiviral effects of WWMJ on various influenza strains were then investigated.

| Influenza growth stage inhibited by WWMJ
The stage at which viral growth is inhibited by WWMJ was determined by performing a time-of-addition assay. The time points at which WWMJ was added to the incubation mixture are shown in  (Figure 2b), the inhibitory mechanism at this step was then elucidated. Type I highmannose-specific antiviral algal lectins have been reported to bind with high affinity to the viral envelope HA (Mu, 2017). The presence of similar lectins in WWMJ was determined by performing a viral yield assay in MDCK cells using heat-treated WWMJ, which showed that the virus-inhibiting activity of WWMJ was not affected even after heat treatment.

| WWMJ inhibits the cellular entry of influenza virus
Wild watermelon juice inhibits viral adsorption onto cells as shown in Figure 2b ternalization, such as by endocytosis. However, the viral adsorption assay showed that the amount of bound virus at 37°C was 10 times higher than that at 4°C. WWMJ did not suppress cell-cell fusion using the A/PR/8/34 virus and CV-1 cell at any of the concentrations tested (up to 4 mg/ml), and the fusion index was found to be more than 0.85 at every concentration (Figure 3b). These results suggested that WWMJ inhibited viral entry by endocytosis but did not inhibit viral fusion ability.

| Effects of WWMJ treatment on influenza virus-infected mice
The therapeutic effects of These WWMJ in vivo were investigated using a mouse model of influenza viral infection. The experimental schedule is shown in Figure 5a. Mice were treated with 20 mg/kg/day WWMJ (test group) or PBS (control group), and their survival rates were evaluated. The body weights of the WWMJ-treated and control mice did not show any significant difference (unpublished data).
As shown in Figure 5b, the nasal administration of WWMJ slightly improved the survival rate of the infected mice monitored for 14-day postinoculation. WWMJ-treated mice were found to die 6-day postinoculation, as compared to death 4-day postinoculation observed with control mice. The survival rates of the WWMJ-treated mice were 0.90 and 0.50 at 6-and 7-day postinoculation, respectively, as compared to 0.20 and 0, respectively, in the control mice. Thus, WWMJ treatment increased the lifespans of influenza virus-infected mice by at least 2 to 3 days. Kaplan-Meier method and log-rank test showed that this difference in survival was significant (p < .01).

| D ISCUSS I ON
Influenza is an acute respiratory disease that affects many people throughout the world. Several anti-influenza drugs are used to treat patients infected with various influenza viruses. However, viral strains resistant to these drugs have been reported (Barr, 2007;Seibert, 2012). Viral infections are inhibited by carbohydrates, such as the marine-microalga-derived sulfated polysaccharide p-KG03 (Kim, 2012).
The sugar contents of WWMJ and WMJ were determined and found to be 2°Bx and 8°Bx, respectively, on the Brix scale; however, neither contained sulfated polysaccharides (unpublished data). WMJ was found to have four times the sugar content of WWMJ, which indicates that the antiviral effect of WWMJ is not attributable to the influence of sugar. Figure 3 shows that WWMJ did not inhibit the recognition of sialic acid on the cell membrane by HA protein. As a conclusion, Figure 4 suggests that WWMJ inhibits energy-dependent entry of viruses, in other words, endocytosis.
The finding that WWMJ has a strong anti-influenza activity, whereas the commercially available WMJ has little anti-influenza activity, implies that watermelons have lost their effective antiviral components over the course of breeding to suit our taste. Consistent with this idea, it has been reported that wild plants have anti-influenza properties (Haasbach, 2014). Therefore, wild plants such as WWM and various other spices may have strong antiviral activities against influenza viruses.
In this study, further experiments were carried out to clarify the mechanism and stage of the anti-influenza activity of WWMJ.  of Figures 2 and 4 suggested that clathrin-dependent and/or -independent endocytosis (Fujioka, 2018) might be inhibited by WWMJ.
Inhibition of the late phase of influenza virus propagation might involve inhibition of viral assembly, and the mechanism might be similar to the inhibition mechanism of daidzein (Horio, 2020 (Nagai, 2018).
Therefore, the viral replication-inhibition activity of WWMJ does not show virus-type specificity, unlike that of amantadine (Barr, 2007

S TU D I E S I N VO LV I N G A N I M A L O R H U M A N S U BJ EC T S
This study was performed with the approval of the Animal