Natural extracts, honey, and propolis as human norovirus inhibitors

Norovirus is the most important cause of acute gastroenteritis, yet there are still no antivirals, vaccines, or treatments available. Several studies have shown that norovirus-specific monoclonal antibodies, Nanobodies, and natural extracts might function as inhibitors. Therefore, the objective of this study was to determine the antiviral potential of additional natural extracts, honeys, and propolis samples. Norovirus GII.4 and GII.10 virus-like particles (VLPs) were treated with different natural samples and analyzed for their ability to block VLP binding to histo-blood group antigens (HBGAs), which are important norovirus co-factors. Of the 21 natural samples screened, date syrup and one propolis sample showed promising blocking potential. Dynamic light scattering indicated that VLPs treated with the date syrup and propolis caused particle aggregation, which was confirmed using electron microscopy. Several honey samples also showed weaker HBGA blocking potential. Taken together, our results found that natural samples might function as norovirus inhibitors.

In order to evaluate how the natural extracts alter the stability of intact norovirus VLPs, the diameters of natural extract treated VLPs were measured using DLS. The untreated GII.4 and GII.10 VLPs exhibited a single peak, indicating a homogenous sample with VLP diameters of approximately 40-46 nm (Fig. 2). For GII.4, incubation with apple sweetener, agave syrup, and maple syrup indicated that the size of the VLPs were comparable with untreated VLPs. Coconut blossom, date syrup, royal jelly, and barley malt dramatically amplified the heterogeneity of GII.4 VLPs and led to an increased peak shift, indicating particle aggregation. Similarly for GII.10, apple sweetener, agave syrup, and maple syrup treatments were comparable to the untreated VLPs. Coconut blossom caused a slight increase in VLP heterogeneity, whereas date syrup, royal jelly, and barley malt amplified the heterogeneity of GII.10 VLPs and lead to an increased peak shift.
To directly observe the effects and compare with DLS measurements, the natural extracts were incubated with VLPs and examined using EM (Fig. 3). The morphology of the untreated GII.4 VLPs were mainly T = 4 particles, whereas the untreated GII.10 VLPs were a mixture of T = 3 and T = 1 icosahedral particles. The morphology of the GII.4 VLPs treated with apple sweetener, agave syrup and maple syrup were comparable to untreated VLPs. With coconut blossom, date syrup, and barley malt, the GII.4 VLPs were found in small and large clumps. For royal jelly, single VLPs were observed, but a large part of the EM grid contained long thin rods, which originated from the natural extract sample. Taken together, the EM results matched the DLS measurements and confirmed that several natural extracts caused VLP aggregation and clumping.
Honey treatment. Ten different honey samples were examined for their ability to block norovirus GII.4 and GII.10 VLPs from binding to HBGAs (Fig. 4). The honey samples inhibited GII.4 and GII.10 VLPs at IC 50 values ranging between 4.72-24.77% and 3.38-15.26%, respectively. For GII.4 VLPs, the honey samples with the lowest IC 50 concentrations were raspberry honey (IC 50 = 4.72%), Alpine forest honey (IC 50 = 4.78%), Oak tree honey (IC 50 = 5.85%), and eucalyptus honey (IC 50 = 6.22%). For GII.10 VLPs, a similar finding was observed, i.e., Alpine forest honey (IC 50 = 3.38%), raspberry honey (IC 50 = 5.60%), eucalyptus honey (IC 50 = 5.96%), and Oak tree honey (IC 50 = 6.11%). Experiments with only solvent produced results similar to the PBS controls (no treatment), which indicated the solvents did not affect the inhibition. Overall, these findings showed that honey samples from different herbal and regional origins could block norovirus VLPs at varied concentrations and similarly between the two genotypes. We also evaluated the treatment with honey on the stability of VLPs using DLS (Fig. 5). All honey samples caused noticeable changes to the GII.4 and GII.10 VLP diameters, where an increase in the heterogeneity and a peak shift was observed. Treatment with Robinia honey was slightly different between the GII.4 and GII.10 VLPs, where the GII.10 had a greater increase in heterogeneity compared to GII.4 VLPs. To directly observe these effects, the honey-treated VLPs were examined using EM (Fig. 6). Most of the honey samples caused the GII.4 and GII.10 VLPs to clump into small aggregates, although single particles were also observed. This result suggested that VLPs were prone to sticking together and remaining for the most part intact. VLPs from binding to HBGAs (Fig. 7). The propolis samples showed a dose-dependent inhibition and the GII.4 and GII.10 VLPs were blocked at low concentrations with IC 50 values ranging between 0.44-5.38%. The highest inhibition was observed with the 96% ethanol propolis, i.e., GII.4 (IC 50 = 0.44%) and GII.10 (IC 50 = 0.57%). A small difference in inhibition levels was observed with DMSO propolis i.e., GII.4 (IC 50 = 1.74%) and GII.10 (IC 50 = 5.38%). We also evaluated propolis treatment on the stability of GII.4 and GII.10 VLPs using DLS (Fig. 8). Almost all propolis samples increased the heterogeneity and led to a peak shifts ranging between 100-1000 nm, even at short (10 min) incubation times. One propolis sample, 20% PEG200 propolis had little effect on the GII.10 VLPs, whereas for the GII.4 VLPs a shift in diameter and heterogeneity was observed. To directly observe the effects, the propolis-treated GII.4 and GII.10 VLPs were examined using EM (Fig. 9). Three of the propolis samples (DMSO, PEG200, and 70% EtOH) caused the VLPs to aggregate into clumps, although intact particles were also observed. For the 96% ethanol propolis sample, large aggregates of disassembled VLPs and very few intact particles were observed.

Discussion
A recent study using a norovirus cell culture system, where human enteroids are used to propagate the virus, showed a norovirus-specific monoclonal antibody inhibited norovirus virions in cell culture 31 . They also observed norovirus VLP aggregation (measured using DLS) occurred when treated with the monoclonal antibody and suggested that the VLPs might cross-link which in turn sterically blocked the HBGA binding site. However, the monoclonal antibody did not bind directly to the HBGA pocket and the footprint was located on the side of the P domain. A similar observation was made with several norovirus-specific Nanobodies 5,6 . Inhibition studies using natural extracts, such as pomegranate juice, cranberry juice, and green tea extracts have indicated varied levels of HBGA inhibition, although the precise binding site on the capsid remains unknown [26][27][28][29][30] . The structural  www.nature.com/scientificreports/ basis other natural samples, i.e., citrate and HMOs showed that these molecules bound precisely at the HBGA pocket on the capsid, which in turn blocked VLP attachment to HBGAs [16][17][18][19] . Presently, there is little available for treating a norovirus infection, while vaccines and antivirals are still at the clinical trial phase and development stage, respectively. From a therapeutic point of view, most infected individuals would favor the possibility of reducing the severe symptoms, which can include profuse vomiting, nausea, fever, and diarrhea. In an outbreak setting, norovirus is highly contagious, and virions are stable in the environment for long periods. Disinfection methods include chlorine bleach and hydrogen peroxide. Alcohol based disinfectants were found to have mixed conclusions but are generally regarded as a last line for defense.  www.nature.com/scientificreports/ For practical purposes, chlorine bleach and hydrogen peroxide generally require the closure of the contaminated areas, and this can include hospital wards, schools, cruise ships, and aged-care homes. Another important area of norovirus infections is related to food contamination, which can include oysters, clams, ice, and fruit. Contaminated food is becoming an increasing problem worldwide and decontaminating food is a high priority for the agriculture and seafood sectors. The discovery of an inhibiting, safe, and natural treatment would be of great interest and benefit. So far, decontamination methods have included UV radiation, chemical disinfectants, and high pressure [32][33][34][35][36][37][38][39][40] .
In the current study, we showed date syrup was the strongest inhibitor of GII.4 (IC 50 = 0.06%) and GII.10 (IC 50 = 0.11%) norovirus VLPs using ELISA, which was also confirmed using DLS and EM. Barley malt showed a  www.nature.com/scientificreports/ weaker inhibition potential with GII.4 norovirus VLPs (IC 50 = 0.90%), which was confirmed using DLS and EM, whereas for GII.10 norovirus barely malt had little effect at the tested concentrations. Honey samples were less effective at blocking GII.4 and GII.10 VLP attachment to HBGAs compared to date syrup, although treatment caused VLP aggregation. All propolis samples showed inhibition potential for GII.4 and GII.10 VLPs, especially 96% EtOH propolis (IC 50 = 0.44% and IC 50 = 0.57%, respectively), which was confirmed using DLS and EM. Our results were comparable to citrate and Nanobody treatments that showed the integrity of norovirus VLPs were compromised and changes in morphology included particle disassembly and aggregation 6,18 . Likewise, treatments with other natural extracts observed similar findings. For example, grape seed extract treatment resulted in particle deformation and enlargement 28 . Several studies using surrogate noroviruses (feline calicivirus and murine norovirus) also found that natural fruits or their components could inhibit or reduce infectivity [41][42][43][44][45] . In summary, the data developed in this study have produced some promising findings. The idea of natural therapy against viral infections is not entirely new and is just becoming an interesting topic of noroviruses inhibition. Further studies that examine norovirus inhibition in cell culture are expected as well as testing other natural extracts are planned.

Methods
Natural extracts, honey, and propolis. Seven natural extracts (date syrup, barely malt, apple sweetener, agave nectar, coconut blossom syrup, royal jelly and maple syrup), ten different honey samples (Alpine forest, coriander, eucalyptus, fir I, fFir II, Mexico, Oak tree, Orange blossom, and Robinia), and two propolis samples (MF and tincture) were purchased from various sources (Table 1). Natural extracts and honey samples were diluted in PBS (pH 7.4), filtered (0.45 µm), and then stored at 4 °C. Propolis MF was mashed and then incubated in three different solvents, i.e., either 20% PEG200, 100% DMSO, or 70% ethanol. These propolis samples were left for 30 days at room temperature and then clarified by centrifugation and filtering (pore size 0.45 µm), and the final solution was stored at 4 °C. Propolis tincture was purchased ready to use in 96% ethanol. Controls with solvents (i.e., PEG200, DMSO, and ethanol) at concentrations comparable with the diluted propolis samples were performed, to ensure that inhibition or aggregation effects were not caused by solvents.   www.nature.com/scientificreports/ VLP expression. Norovirus VP1 of two different GII genotypes, GII.4 (JX459908, Sydney2012) and GII.10 (AF504671, Vietnam 026), were expressed in insect cells as described previously 46 . Briefly, VLPs were clarified from insect cell medium, pelleted by ultracentrifugation, and then further purified using CsCl gradient ultracentrifugation.
HBGA blocking assay. Blocking assays were performed using a method as descried earlier for HMOs 16,17 .
Briefly, 96-well plates were coated with pig gastric mucin type III (PGM), washed three times with PBS containing 0.1% Tween 20 (PBS-T), and subsequently blocked with 5% skimmed milk in PBS. The concentration of untreated GII.4 and GII.10 VLPs were optimized for binding to PGM as previously described 17,47 . The GII.4 VLPs (1 μg/ml) or GII.10 VLPs (10 μg/ml) were mixed (1:1) with serially diluted extracts (starting at a concentration from 25%) for 3 h at room temperature. Plates were washed three times with PBS-T, then 100 µl of each VLP/ natural extract mixture was added to triplicate wells for 2 h at room temperature. After washing, 100 µl of GII.4 or GII.10 genotype-specific polyclonal rabbit antibody was added as a primary detection antibody for 1 h at room temperature. Following a wash step, horseradish peroxidase (HRP) conjugated polyclonal anti-rabbit antibody or HRP-conjugated streptavidin was added to the wells and incubated for 1 h at room temperature. Electron microscopy. The VLPs (treated and untreated) were analyzed using negative stain electron microscopy (EM) as described 6 . A 25% stock of natural extracts or honey were mixed (1:1) with 1 mg/ml VLPs and incubated for 1 h at room temperature (final concentration of natural extract and honey = 12.5%). Propolis samples were prepared as described for DLS. Samples were diluted 1:40 in distilled water and immediately loaded on carbon coated EM grids. Grids were washed with distilled water, stained with 0.75% uranyl acetate, and then examined on a Zeiss 900 electron microscope. Numerous EM images were analyzed for each sample and one representative was shown. 5%, PEG200 2.5%, 70%-ethanol 8.75%, and 90%-ethanol 12%) at 10, 60, and 120 min (represented by light, medium and dark colored lines, respectively). Untreated VLPs showed a single peak (black lines). VLPs treated with 100% DMSO propolis, 70% ethanol propolis, and 96% ethanol propolis showed similar peak shifts for (A) GII.4 and (B) GII.10. The 20% PEG200 propolis treatment also showed a peak shift for GII.4 VLPs, although was less pronounced than with the other propolis samples, while for GII.10 VLPs, the 20% PEG200 propolis showed only a minor peak shift. Each experiment was performed in triplicate and representative measurements were presented.

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.