Mechanism of the Antiviral Effect of Hydroxytyrosol on Influenza Virus

Article from www.elsevier.com/locate/antiviral


1. INTRODUCTION

Hydroxytyrosol (HT), a small-molecule phenolic compound, inactivated influenza A viruses including H1N1, H3N2, H5N1, and H9N2 subtypes. HT also inactivated Newcastle disease virus but not bovine rotavirus, and fowl adenovirus, suggesting that the mechanism of the antiviral effect of HT might require the presence of a viral envelope. Pretreatment of MDCK cells with HT did not affect the propagation of H9N2 virus subsequently inoculated onto the cells, implying that HT targets the virus but not the host cell. H9N2 virus inactivated with HT retained unaltered hemagglutinating activity and bound to MDCK cells in a manner similar to untreated virus. Neuraminidase activity in the HT-treated virus also remained unchanged. However, in the cells inoculated with HT-inactivated H9N2 virus, neither viral mRNA nor viral protein was detected. Electron microscopic analysis revealed morphological abnormalities in the HT-treated H9N2 virus. Most structures found in the HT-treated virus were atypical of influenza virions, and localization of hemagglutinin was not necessarily confined on the virion surface. These observations suggest that the structure of H9N2 virus could be disrupted by HT.

Hydroxytyrosol (HT) is a small-molecule phenolic compound
found in the leaves and fruits of olive (Olea eurolaea) as a metabolite
of oleuropein (Ole), which is one of the major polyphenolic components of olive products (Fig. 1). HT can be purified from olive products

(Ciafardini et al., 1994; Lee-Huang et al., 2007a;
Visioli et al., 1998a), and can also be prepared by chemical synthesis
(Capasso et al., 1999; Espín et al., 2001; Lee-Huang et al.,
2007a; Tuck et al., 2000). The polyphenolic compounds contained
in olive leaf extracts and olive oils have been reported to exert various
bioactivities, including antioxidant (Coni et al., 2000; Visioli
et al., 1995, 1998a,b, 2001), anti-inflammatory (Beauchamp et al.,
2005; Bitler et al., 2005; Pacheco et al., 2007), and antimicrobial
activities against bacteria, fungi, andmycoplasma (Aziz et al., 1998;
Bisignano et al., 1999; Furneri et al., 2002). Among them, the beneficial
effects of olive phenols on coronary heart diseases have been
extensively studied (Covas, 2007).
! Corresponding author. Tel.: +81 155 495893; fax: +81 155 495893.
E-mail address: hogawa@obihiro.ac.jp (H. Ogawa).
1 Present address: Central Research Laboratories, Nippon Zenyaku Kogyo Co., Ltd.,
Koriyama, Fukushima 963-0196, Japan.
It has also been reported that olive leaf extracts exhibit antiviral
activities against human immunodeficiency virus type 1 (HIV-1)
(Bao et al., 2007; Lee-Huang et al., 2003, 2007a,b) and viral haemorrhagic
septicaemia rhabdovirus (Micol et al., 2005). Lee-Huang et
al. (2003) have demonstrated the strong activity of olive leaf extract
against HIV-1, in which the compound mixture inhibited acute
infection and cell-to-cell transmission of HIV-1, and ultimately suppressed
viral replication in the infected cells. Their recent studies
identified Ole and HT as the HIV-1 inhibitors contained in the olive
leaf extracts. In these studies, Ole and HT inhibited not only the
fusion between viral and cellular membranes but also the integrase
activity of the virus. A molecular modeling study revealed that Ole
and HT bound to the conserved hydrophobic pocket on the surface
of the HIV-1 envelope protein gp41, which is a transmembrane
subunit of HIV-1 envelope glycoprotein. The results of a computational
study suggested that HT is the predicted main moiety for
the binding to gp41. Binding of HT to gp41 appears to cause a conformational
change in the glycoprotein, which could result in the
inhibition of viral entry into the target cells (Bao et al., 2007; Lee-
Huang et al., 2007a,b). These studies suggest that Ole and HT might
be useful against other viruses with type I transmembrane envelope
glycoproteins.
Ma et al. (2001) have previously reported that Ole isolated from
the fruits of Ligustrum lucidum exhibits significant antiviral activ-
0166-3542/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.antiviral.2009.03.002
36 K. Yamada et al. / Antiviral Research 83 (2009) 35–44
Fig. 1. Chemical structure of hydroxytyrosol and oleuropein.
ities against respiratory syncytial virus and parainfluenza type 3
virus, but not against herpes simplex type 1 virus and influenza
A virus, although all these viruses possess type I transmembrane
envelope glycoproteins. These results might suggest that the antiviral
effect of Ole and HT is applicable to a limited number of viruses
that have the relevant binding sites for these compounds.
In the present study, the antiviral effect of HT was investigated
using influenza A viruses, Newcastle disease virus (NDV), group
A bovine rotavirus (BRV), and fowl adenovirus (FAV). Among the
viruses tested, HT was effective against the enveloped viruses, but
not against the non-enveloped viruses. We attempted to elucidate
the mechanism of the antiviral effect of HT against influenza virus,
and found morphological abnormalities in the HT-treated influenza
virus.
2. Experimental/materials and methods
2.1. Cells and viruses
Avian influenza viruses A/chicken/Yokohama/aq55/01 (H9N2)
(Eto and Mase, 2003) and A/chicken/Yamaguchi/7/04 (H5N1),
and NDV Ibaraki/85 strain were provided by the National
Institute of Animal Health, Japan. Human influenza viruses
A/Hokkaido/30/2000 (H1N1) and A/Hokkaido/52/98 (H3N2) were
from the Hokkaido Institute of Public Health, Japan. BRV Lincoln
strain was from the Hokkaido Livestock Research Institute, Japan,
and serotype 1 FAV 99ZH strain fromthe Zennoh Institute of Animal
Health, Japan.
The H9N2 viruswas propagated in the allantoic cavity of 10-dayold
embryonated chicken eggs, and purified by ultracentrifugation
using 30% and 60% sucrose solutions. The protein content of the
purified viruswas measured by using a Micro BCA protein assay kit
(Pierce, Rockford, IL). H5N1 virus and NDV were also propagated
in embryonated chicken eggs, and the obtained allantoic fluid was
used as virus stock solution. Human influenza viruses were grown
inMDCKcells, and the culture supernatant served as the virus stock
solution. The infectivity of influenza viruses was tested in MDCK
cells.Upon the inoculation of influenza viruses other than the H5N1
virus, the mediumwas replaced with virus growth medium (VGM),
which was prepared according to the WHO Manual on Animal
Influenza Diagnosis and Surveillance (WHO Manual). The infectivities
of NDV, BRV, and FAV were tested in MDBK, MA104, and LMH,
respectively. Dulbecco’s modified Eagle’s medium (Nissui Pharmaceutical
Co., Ltd., Tokyo, Japan) was used for culturing the MDCK,
MDBK, and MA104 cells. LMH cells were cultured in Waymouth’s
medium (Invitrogen, Carlsbad, CA). Each cell medium was supplemented
with 10% fetal bovine serum and 2mM l-glutamine. Prior
to the inoculation ofMA104 cells with BRV, the viruswas pretreated
with 10!g/ml trypsin at 37 "C for 30 min, then the trypsinized virus
was added to the cells in which themedium had been replaced with
VGM.
2.2. Treatment of viruses and cells with HT
HT (98.3% purity), purified from olive leaf extracts, was provided
by Eisai Food & Chemical Co., Ltd. (Tokyo, Japan). HT was
dissolved in phosphate-buffered saline (PBS) and diluted to concentrations
ranging from 0.1 to 1000!g/ml (0.65–6535.9!M). The
viruses (105.2–106.6 TCID50/ml) were added to HT solutions of different
concentration and maintained at room temperature for a
predetermined period of time. Following the treatment, the virus
titer in the mixturewas measured by inoculating serial 10-fold dilutions
(beginning with a 1:10 dilution) of the mixture into the host
cells. The TCID50 was calculated by the Behrens-Karber method
based on the cytopathic effect (CPE) observed on day 5 postinoculation
(p.i.). The effect of benzalkonium chloride (BC) on the
H9N2 virus was also tested as a control compound with virucidal
activity. The BC solutions at the concentrations between 0.002% and
0.016%were prepared using commercially available 10% BC solution
(Osvan, Nihon Pharmaceutical Co., Ltd., Tokyo, Japan).
In order to test the effect of HT on the cells, the cells were first
cultured with a medium containing HT for 24 h. The cell medium
was then replaced with fresh medium lacking HT, the cells having
been initially washed with PBS. The serially diluted HT-untreated
virus was inoculated into the cells, and the virus titers measured in
the HT-treated and untreated cells were compared.
The cytotoxicity of HT on the cells was analyzed by measuring
the lactate dehydrogenase released from the cells using
a Cytotoxicity Detection kit (Roche, Mannheim, Germany). The
percentage cytotoxicity was calculated by the following equation
using the obtained absorbance values, from which the absorbance
values in the corresponding background control were subtracted:
[(Experimental release−spontaneous release)/(maximum
release−spontaneous release)]×100%.
The hemagglutination titer of the H9N2 virus was measured in
96-well microplates with V-shaped bottom. The virus was serially
diluted in a twofold dilution with PBS. Into each well containing
50!l of the virus solution, an equal volume of 0.8% chicken
erythrocytes suspended in PBS was added. Following mechanical
vibration, the plates containing the mixture of virus and erythrocyteswere
kept at roomtemperature, and the resultswere recorded
after 30 min. The titer was expressed as the reciprocal of the highest
dilution of the virus showing complete hemagglutination. The
assaywas triplicated for each virus dilution, and the HA titer determined
represents the titer identically recorded with all of the three
or two out of the three tests. We considered the difference greater
than 2 times to be a significant difference in hemagglutination titer.
The neuraminidase (NA) activity of the virus was measured
according to theWHOManual. Briefly, the viruswas serially diluted
in a twofold dilution with PBS and incubated with Fetuin (Sigma, St.
Louis, MO) at 37 "C for 18 h. The amount of sialic acid liberated was
determined chemically with the thiobarbituric acid that develops
color in proportion to the concentration of free sialic acid. The color
developed was measured at a wavelength of 549 nm.
2.3. Real-time reverse transcription-polymerase chain reaction
(RRT-PCR)
Total RNA was extracted from the MDCK cells inoculated
with either the HT-treated or untreated H9N2 virus using Isogen
(Nippon Gene, Tokyo, Japan). The RNA thus obtained was
K. Yamada et al. / Antiviral Research 83 (2009) 35–44 37
transcribed into cDNA by SuperScript III Reverse Transcriptase
(Invitrogen) under the following conditions: 50 "C for 60 min
and 70 "C for 10 min. Real-time reverse transcription-polymerase
chain reaction (RRT-PCR) was performed in order to detect
the viral RNA (vRNA) and viral mRNA (vmRNA) by using the
LightCycler-FastStart DNA Master SYBR Green I kit (Roche). In
order to detect vRNA, the extracted RNA was reverse transcribed
using the Uni12 primer 5%-AGCAAAAGCAGG-3%. For vmRNA detection,
the oligo-dT primer was used in the RT reaction. Partial
segments of the H9N2 virus genes were amplified using the
following primer sets (respective product sizes are in parentheses):
PB2, 5%-CTGGGAGCAGATGTACACTC-3% and 5%-ACAGCTTGTTCCTCAGTTGG-
3% (207 bp); PB1, 5%-GTGTCAGATGGAGGGCCAAA-
3% and 5%-TCGCAACGGCATCATACTCC-3% (214 bp); PA, 5%-AAGGCTCCATCGGAAAGGTG-
3% and 5%-AGCCCTCCAAGATCGAAGGT-3%
(169 bp); HA, 5%-AGTGCATGGAGACAATTCGG-3% and 5%-CATTGGACATGGCCCAGAAC-
3% (200 bp); NP, 5%-GCAACTGCTGGTCTTACCCA-
3% and 5%-CCGAATCAGCTCCATCACCA-3% (201 bp); NA, 5%-GCATAGCATGGTCCAGCTCA-
3% and 5%-CCTGATGCACTTCCATCCGT-3%
(220 bp); M1, 5%-GACGTTCCATGGAGCAAAGG-3% and 5%-GCCTGATTAGTGGGTTGGTG-
3% (204 bp); NS2, 5%-TAGTGGGCGAAATCTCACCA-
3% and 5%-ATCCTCATCGCTGCTTCTCC-3% (161 bp). The
thermal profile included the following cycles: denaturation
at 95 "C for 10 min followed by 45 cycles at 95"C for 10 s,
55 or 56 "C for 5 s, and 72 "C for 6–9 s. RRT-PCR for detecting
the housekeeping gene GAPDH was also performed using the
following primer set: 5%-GGGGCCATCCACAGTCTTCT-3% and 5%-
GCCAAAAGGGTCATCATCTC-3%. Following the amplification, a
melting curve analysis was performed to confirm the specificity
of the reaction. The product size was confirmed by 1% agarose gel
electrophoresis. The threshold cycle (CT) value for each sample
was calculated by determining the point at which the fluorescence
exceeded the threshold limit.
2.4. Flow cytometry analysis
The virus bound to the cell surfacewas analyzed by flowcytometry.
MDCK cells were inoculated with the HT-treated or untreated
H9N2 virus at a concentration of 10!g/ml and incubated at 4 "C
for 1 h. The cells were then stained with the monoclonal antibody
(mAb) against hemagglutinin (HA) labeled with Alexa Fluor 488
by using a Zenon Mouse IgG Labeling kit (Invitrogen) according to
the Manufacturer’s instructions. Stained cells were fixed with 1%
paraformaldehyde solution.
The presence of H9N2 virus within the cells was also analyzed
by flowcytometry using amAbagainst the viral nucleoprotein (NP).
MDCKcellswere inoculated with either theHT-treated or untreated
virus at a concentration of 1!g/ml and incubated at 37 "C for 3
or 24 h. The cells were then harvested, followed by fixation and
permealization using a BD Cytofix/Cytoperm kit (BD Biosciences,
San Jose, CA). Subsequently, the cells were intracellulary stained
with anti-NP mAb, which was labeled with Alexa Fluor 488.
Stained cells were analyzed using a FACSCanto flow cytometer
(BD Biosciences). The mAbs used for cell stainingwere produced in
our laboratory as previously described (J. Virol.Methods, submitted
for publication).
2.5. Electron microscopic analysis
The HT-treated or untreated H9N2 virus was attached to a
carbon-coated collodion grid (Nisshin EM Co. Ltd., Tokyo, Japan),
and negatively stained with 2% phosphotungstic acid, pH 6.8 (PTA).
In order to detect the HA protein using a specific antibody, the virus
on the grid was first blocked with 1% bovine serum albumin in PBS
for 30 min, stained with anti-HAmAbbiotinylated using Biotin-OSu
(Dojindo, Kumamoto, Japan) for 30 min, and then incubated with
streptavidin immunogold conjugate (BBInternational, Cardiff, UK)
for 60 min. The viruswas stained with the PTA, and examined using
a Hitachi H7500 electron microscope (Tokyo, Japan).
3. Results
3.1. Antiviral effect of HT on enveloped and non-enveloped viruses
The antiviral effect of HT on the H9N2 virus, NDV, BRV, and FAV
was investigated by measuring the virus titers 24 h after mixing
Fig. 2. Comparison of the antiviral effects of HT on a variety of viruses. The viruses were treated with the HT solution at room temperature for 24 h. The treated virus was
serially diluted in a 10-fold dilution beginning with a 1:10 dilution, and inoculated into the cells to evaluate the virus titer. The cells for titration of each virus are described
in Section 2. Asterisks represent the results obtained at a HT concentration slightly cytotoxic to the cells. Black bars represent the titer of the control virus that had been kept
at room temperature for 24 h in the absence of HT.
38 K. Yamada et al. / Antiviral Research 83 (2009) 35–44
Fig. 3. Antiviral effect of HT on influenza viruses. (A) Effect of HT on H1N1, H3N2 and
H5N1 viruses. The viruses were treated with HT (1000!g/ml) at room temperature
for 24 h. The treated virus was serially diluted in a 10-fold dilution beginning with
a 1:10 dilution, and inoculated into MDCK cells to evaluate the virus titer. Black
bars represent the titer of the control virus that had been kept at room temperature
for 24 h in the absence of HT. (B) Propagation of H9N2 virus inoculated into MDCK
cells concurrently with the addition of HT to the cell. The virus was inoculated into
MDCK cells immediately after mixing with HT solution. Final concentrations of HT
in the cell medium are shown. Asterisk represents the results obtained at the HT
concentration slightly cytotoxic to the cells. (C) Propagation of H9N2 virus in MDCK
cells pretreated with HT. The cells had been cultured in the presence of HT for 24 h.
Then the cell medium was replaced with the one lacking HT, and inoculated with
untreated H9N2 virus.
the virus with different concentrations of HT. HT concentrationdependently
lowered the titer of the two enveloped viruses at
concentrations between 0.1 and 10!g/ml (0.65–65!M) for the
H9N2 virus and 0.1 and 100!g/ml (0.65–650!M) for NDV (Fig. 2).
HTalso lowered the titers of other influenza viruses includingH1N1,
H3N2, and H5N1 subtypes (Fig. 3A). In contrast, the virus titers
of the non-enveloped BRV and FAV were virtually unchanged by
treatment with the HT solution, although the virus titers of BRV
and FAV appeared to be slightly decreased at the concentration of
1000!g/mlHT(Fig. 2). The 1:10 dilution of the inoculum containing
1000!g/mlHTyielded the highest concentration ofHT(100!g/ml)
within the cells. The concentration of 100!g/ml HTwas the threshold
dose that did not cause a cytotoxic effect against MDCK, LMH,
and MA104 cells, but was slightly cytotoxic (7.2% cytotoxicity) to
MDBK cells (data not shown).
3.2. Kinetics of the antiviral effect of HT on the H9N2 virus
In HT solutions with concentrations greater than 12.5!g/ml,
the titer of the H9N2 virus was lowered by >1.5 log in 1 h, and further
decreased in a time-dependent manner (Fig. 4). The virus titer
decreased by >4 log and reached undetectable levels at 12 h following
the inoculation of the virus into 200!g/ml HT. At 12.5 and
50!g/ml HT, the virus titer decreased by >4 log at 24 h. In contrast,
the titer of the H9N2 virus declined in BC solutions at a much faster
speed. The titer decreased nearly 4 log in 0.016% and 0.008% BC
solutions within 10 min and 3 h, respectively (Fig. 5).
If the H9N2 virus was inoculated into MDCK cells concurrently
with the addition of HT to the cells, the virus titer decreased by
>1.5 log at a concentration higher than 50!g/ml HT, but only a
minor changewas observed at lower concentrations of HT (Fig. 3B).
Since the HT concentrations indicated in Fig. 3B present the final
concentrations of HT contained in the cell medium, the final concentrations
of HT are 10 times higher than those in Figs. 2 and 4. The
results presented in Fig. 3B indicate that if HT gets into contact with
the H9N2 virus in the presence of cells it is unable to significantly
alter the virus titer, even at the semi-cytotoxic dose.
The H9N2 virus was able to propagate in the cells that had been
cultured in a medium containing HT for 24 h as in the untreated
cells. Thus, pretreatment of the cells with HT did not affect the titer
of the H9N2 virus subsequently inoculated onto the cells (Fig. 3C).
3.3. HA activity and NA activity of the HT-treated H9N2 virus
HA and NA activities were compared between the HT-treated
and untreated H9N2 virus. The HT-treated virus retained an hemagglutination
titer at an equivalent level to the untreated virus
(Table 1). NA activity also did not differ between the HT-treated
and untreated virus (Fig. 6). In flow cytometric analysis, HA was
detected on the surface of cells inoculated with the HT-treated virus
at a level similar to the cells inoculated with untreated virus (Fig. 7).
3.4. Suppression of viral RNA synthesis in the cells inoculated
with the HT-treated H9N2 virus
The synthesis ofvmRNAandvRNAinMDCKcells inoculatedwith
the HT-treated or untreated H9N2 virus was analyzed at 0.2, 1, 3, 6,
and 24 h after the inoculation of the virus. A portion of 8 segments
of the influenza virus RNA was amplified by RRT-PCR. In the cells
inoculated with untreated H9N2 virus, the CT values for each of the
segments of vmRNA continued to decline between 1 and 6 or 24 h
p.i. (Fig. 8). The CT values for vRNA in the untreated virus-inoculated
cells underwent a similar transition (data not shown). These results
indicated active replication of the virus in the cells following the
viral inoculation. On the other hand, in the cells inoculated with
the HT-treated H9N2 virus, neither the CT values for vmRNA (Fig. 8)
nor those for vRNA (data not shown) changed notably until 24 h p.i.
Table 1
Hemagglutination titer of H9N2 virus treated with HT.
Treatment with HT (!g/ml) Hemagglutination titer/treatment time
1 h 3 h 6 h 12 h 24 h
200 64 64 64 64 64
50 64 64 64 64 64
12.5 64 32 64 64 64
3.13 64 64 64 64 64
0 64 64 64 64 32
The H9N2 virus was treated with the HT solution as in Fig. 1, and hemagglutination
titers were measured using chicken erythrocytes.
K. Yamada et al. / Antiviral Research 83 (2009) 35–44 39
Fig. 4. Antiviral effect of HT on H9N2 virus at various concentrations and different treatment times. The H9N2 virus was treated with the HT solution at room temperature
for the defined period of time. The treated virus was serially diluted in a 10-fold dilution beginning with a 1:10 dilution, and inoculated into MDCK cells. Black bars represent
the titer of the control virus that had been kept at room temperature for the corresponding period of time in the absence of HT.
3.5. Lack of viral protein synthesis in the cells inoculated with the
HT-treated H9N2 virus
In order to detect viral protein synthesis in the cells following the
inoculation of the H9N2 virus, viral NP was intracellularly stained
with anti-NP mAb. In the cells inoculated with untreated H9N2
virus, NP was detected at 3 h p.i., but only at a low level. However,
the amount of NP increased dramatically at 24 h p.i., indicating the
vigorous synthesis of viral protein. In contrast,NPproteinwas undetectable
in the cells inoculated with the HT-treated H9N2 virus at
24 h p.i., suggesting a lack of viral protein synthesis (Fig. 9).
3.6. Morphological abnormality in the HT-treated virus
In the electron microscopic analysis of negatively stained virus,
the virion shapes of the untreated H9N2 virus were found to be
typical of influenza virus. These included spherical, elliptical, and
filamentous virions. Each particle was surrounded by a layer, from
the surface of which projected numerous spikes; these spikes are
assumed to be HA and NA spikes (Fig. 10A). Binding of colloidal gold
confirmed that HA proteins are localized on the spikes (Fig. 11A). In
contrast, few particles were detected in the HT-treated H9N2 virus
as typical influenza virions, and many ambiguous structures were
found instead. These structures appeared to lack the layer of spikes
on the surface. Some virion-shaped particles with spiked surfaces
were observed; however, the outline of these particles was not as
distinct as the untreated virus (Fig. 10B). Localization of HAwas not
necessarily confined to the virion surface, rather it was found to be
diffuse (Fig. 11B).
4. Discussion
The present study demonstrated the antiviral effect of HT on
influenza virus and NDV. HT effectively lowered the virus titer of
the two enveloped viruses in a dose-dependent manner. In contrast,
the virus titer of all the non-enveloped viruses tested, which
included BRV and FAV, was not affected by HT (Figs. 2 and 3). These
results suggest that the viral envelope is likely to be involved in the
mechanism of the HT antiviral effect.
Our observations appear to be similar to those of previous studies
on the antiviral effect of HT on HIV-1, in which HT and Ole
(a precursor substance of HT) dose-dependently inhibited HIV-1
infection and replication. It has been reported that HT and Ole bind
to the HIV–gp41 fusion domain, thereby interfering with the formation
of the gp41 fusion-active core and resulting in the inhibition of
fusion, based on the results obtained by molecular modeling studies
(Bao et al., 2007; Lee-Huang et al., 2003, 2007a, b). HT (mol
wt: 153) is a main metabolite of Ole (mol wt: 539). It is interesting
to note that such a small molecule can effectively block the
protein–protein interaction in the HIV–gp41 fusion domain, crucially
disturbing viral cell entry. If the influenza virus protein has a
binding site for HT in the domain essential for cell entry, HT might
40 K. Yamada et al. / Antiviral Research 83 (2009) 35–44
Fig. 5. Antiviral effect of BC on H9N2 virus at various concentrations and different treatment times. The experiment was performed as in Fig. 4. Black bars represent the titer
of the control virus that had been kept at room temperature for the corresponding period of time in the absence of BC. Broken lines represent the level of virus titer detectable
in the assay. Because the addition of the solution containing BC was toxic to MDCK cells even at the concentration of 0.0008% (10 times dilution of the virus solution treated
with 0.008% BC solution), the virus titer detectable in the assay was higher than with HT.
act on influenza virus. In the present study, the HT-treated H9N2
virus retained unchanged HA activity and bound to MDCK cells in a
mannersimilar to untreatedvirus (Table 1 and Fig. 7); however, viral
replicationwas suppressed in these cells (Figs. 8 and 9). It should be
explored whether the HT-treated H9N2 virus was internalized into
the cells following its binding to the cell. In addition, investigations
should be undertaken in order to determine whether the binding
of HT to the viral proteins occurs at the sites involved in the fusion
process. We are currently trying to determine if the envelope of
influenza virus has a binding site for HT like HIV-1 gp41 does, and,
further, if the HT binding indeed affects the viral cell entry.
On the other hand, the antiviral mechanism of HT on the H9N2
virus and HIV-1 should not necessarily be the same, since the cell
entry mechanism of these viruses is not identical. In fact, the kinetics
of the antiviral effects of HT on the H9N2 virus seemed to be
different to those on HIV-1. Itwas reported thatHTand Ole inhibited
the syncytial formation in cell-to-cell transmission of HIV-1 and the
synthesis of the viral core protein p24 in the assay co-culturing HIVFig.
6. NA activity of HT-treated H9N2 virus. Following a 24-h treatment of the H9N2
virus with 200!g/ml HT, the virus solutions were serially diluted to 1:2, and the NA
activity of the solutions was measured. The data shown are representative of two
separate experiments with similar results. NA activities of untreated virus (&) and
virus treated with HT (!).
infected H9 cells and uninfected MT-2 target cells (Lee-Huang et
al., 2007a). These results suggest that the antiviral activity of HT on
HIV-1 is effective under circumstances in which the target cells have
already been infectedwith the virus. In the present study, the antiviral
effect of HT on the H9N2 viruswas found to be a time-dependent
but relatively slowprocess. HT required a period between 1 and 3 h
to lower the H9N2 virus titer by >1.5 log, and between 12 and 24 h
to decrease the titer by >4 log in MDCK cells (Fig. 4). The addition
of HT to the cells simultaneously with the viral inoculation caused
only a marginal antiviral effect (Fig. 3B). Thus, it is necessary for
HT to treat H9N2 virus, unlike HIV-1, in the absence of target cells
for a certain period of time in order to facilitate its inactivation.
Interestingly, we found that Ole does not exert any antiviral effect
to the H9N2 virus (data not shown). This further suggests that the
mechanism of antiviral effect of HT against influenza virus would
be different from that against HIV-1. The results obtained in this
study also suggest the following: (1) the antiviral effect of HT on
the H9N2 virus is attributable to the direct effect of HT on the virus
and not via an indirect effect on the cells (Fig. 2C); (2) HT appears to
interfere with the early stage of viral infection rather than the later
stage when proliferated virus are released from the cells. Indeed,
the NA activity of the H9N2 virus was not affected by HT (Fig. 6).
It has been reported that CYSTUS052, a polyphenol rich plant
extract from a special variety of Cistus incanus, exerts a potent antiinfluenza
virus activity (Ehrhardt et al., 2007). In these studies,
treatment of the cells with CYSTUS052 resulted in a reduction of
progeny virus titers up to two logs if cellswere either pre-incubated
with CYSTUS052 or the extract was added simultaneously with the
virus. The titer reduction was also observed if the virus was preincubated
with CYSTUS052. The results suggested that CYSTUS052
appears to inhibit the viral entry process, and indeed CYSTUS052
was found to interfere the binding of the viral HA to cellular receptors.
Although the active moiety of CYSTUS052 has not been known,
the antiviral mechanism of HT seems to be different from that of
CYSTUS052.
Since HT requires pre-incubation time in order to be effective
to influenza virus and the time needed for the viral inactivation
K. Yamada et al. / Antiviral Research 83 (2009) 35–44 41
Fig. 7. Presence of viral HA protein on the surface of the cells inoculated with HT-treated H9N2 virus. The surface of MDCK cells was stained with anti-HA mAb labeled with
Alexa Fluor 488 1 h after the inoculation of untreated virus or the virus treated with 200!g/ml HT for 24 h.
Fig. 8. Suppression of vmRNA synthesis in the cells inoculated with HT-treated H9N2 virus. The vmRNAs for each of 8 segments of the virus were amplified by RRT-PCR using
total RNA extracted from the MDCK cells at 0.2, 1, 3, 6, and 24 h after the inoculation of either HT-treated or untreated H9N2 virus and oligo-dT primer. Representative data
from two separate experiments are shown in the Figure. Cycle numbers for threshold (CT) in the RRT-PCR are presented. The black and white bars represent results obtained
with untreated virus and the virus treated with 200!g/ml HT for 24 h, respectively.
42 K. Yamada et al. / Antiviral Research 83 (2009) 35–44
Fig. 9. Lack of viral NP protein in the cells inoculated with HT-treated H9N2 virus. MDCK cells were harvested 3 or 24 h after the inoculation of untreated virus or the virus
treated with 200!g/ml HT for 24 h. The cells were permeabilized and stained with anti-NP mAb labeled with Alexa Fluor 488.
Fig. 10. Electron microscopic analysis of HT-treated H9N2 virus. The negatively stained virus was analyzed. The bar represents 100 nm. (A) Untreated virus; (B) the virus
treated with 200!g/ml of HT for 24 h.
K. Yamada et al. / Antiviral Research 83 (2009) 35–44 43
Fig. 11. Immunoelectron microscopic analysis of HT-treatedH9N2virus. The viruswas stained with biotinylated anti-HAmAbfollowed by streptavidin immunogold conjugate.
The bar represents 100 nm. (A) Untreated virus; (B) the virus treated with 200!g/ml of HT for 24 h.
seems to be longer than those of the currently available disinfectants
such as BC (Figs. 4 and 5), it will not be applicable for
clinical use as it is. Bioavailability of HT has been studied because
of increased interests to its antioxidant effect. It was reported
that HT underwent extensive metabolism and urinary excretion
in vivo. The highest concentration of HT in urine was observed in
the first 4 h after the ingestion of olive oil in humans (Caruso et
al., 2001; Miró-Casas et al., 2001). Such a high rate metabolism
should also hamper the clinical usage of HT. However, clarifying
the antiviral mechanism of HT might open the way to utilize
this chemical available from natural products in influenza control.
Alternatively, the information obtained in the study may
enable HT to become a platform for developing novel antiviral
agents.
The HT-treated H9N2 viruswas morphologically analyzed under
the electron microscope. It was found that most of the structures
observed in the HT-treated H9N2 virus had been ill-defined.
These structures appeared to be lacking surface spikes. HA protein
seemed not be confined only to the surface of virion-shaped
particles. These results suggest that the structure of viral envelope
could be disrupted by HT (Figs. 10 and 11). Suppression of
vmRNA synthesis and lack of viral NP protein observed within the
cells inoculated with HT-treated H9N2 virus (Figs. 8 and 9) might
be the results of an insufficiency in viral binding, viral uncoating
or other steps in viral infection. Although we could detect
the unchanged HA and NA activities in the HT-treated H9N2 virus
(Table 1 and Figs. 6 and 7), the results may not necessarily represent
the activities of intact HA and NA that are attached to the
virion surface. We are currently investigating whether the activities
of HA and NA detected in the HT-treated virus are caused by
the envelope-bound HA and NA or the envelope-unbound free HA
and NA, and whether further the disruption of viral envelope could
result in the generation of HA and NA detached from the virion
surface.
This study first demonstrated the antiviral effect of HT on
influenza virus. The morphological changes observed under the
electron microscope might suggest that the effect of HT to influenza
virus is virucidal rather than antiviral. Additional studies are clearly
required in order to clarify the crucial events that occur during the
interaction of HT with the influenza virus. Although the results of
the current and previous studies suggest that HT seems to act on a
molecule on the viral envelope, the antiviral effect of HT may not
be equivalent to all enveloped viruses. Not only the extent but also
the mechanism of the antiviral effect of HT would depend on how
many viral sites HT interacts with and, furthermore, howimportant
the target sites are to the virus.




Acknowledgements
We thank Sachiko Matsuda for technical assistance. This work
was partially supported by grants from the Program of Founding
Research Centers for Emerging and Reemerging Infectious Diseases,
and by a Grant-in-Aid for Exploratory Research (19659115) from
MEXT (Japan).
References
Aziz, N.H., Farag, S.E., Mousa, L.A., Abo-Zaid, M.A., 1998. Comparative antibacterial
and antifungal effects of some phenolic compounds. Microbios 93, 43–54.
Bao, J., Zhang, D.W., Zhang, J.Z., Huang, P.L., Huang, P.L., Lee-Huang, S., 2007. Computational
study of bindings of olive leaf extract (OLE) to HIV-1 fusion protein
gp41. FEBS Lett. 581, 2737–2742.
44 K. Yamada et al. / Antiviral Research 83 (2009) 35–44
Beauchamp, G.K., Keast, R.S., Morel, D., Lin, J., Pika, J., Han, Q., Lee, C.H., Smith, A.B.,
Breslin, P.A., 2005. Phytochemistry: ibuprofen-like activity in extra-virgin olive
oil. Nature 437, 45–46.
Bisignano, G., Tomaino, A., Lo Cascio, R., Crisafi, G., Uccella, N., Saija, A., 1999. On
the in vitro antimicrobial activity of oleuropein and hydroxytyrosol. J. Pharm.
Pharmacol. 51, 971–974.
Bitler, C.M., Viale, T.M., Damaj, B., Crea, R., 2005. Hydrolyzed olive vegetation water
in mice has anti-inflammatory activity. J. Nutr. 135, 1475–1479.
Capasso, R., Evidente, A., Avolio, S., Solla, F., 1999. A highly convenient synthesis of
hydroxytyrosol and its recovery from agricultural waste waters. J. Agric. Food
Chem. 47, 1745–1748.
Caruso, D., Visioli, F., Patelli, R., Galli, C., Galli, G., 2001. Urinary excretion
of olive oil phenols and their metabolites in humans. Metabolism 50,
1426–1428.
Ciafardini, G., Marsilio, V., Lanza, B., Pozzi, N., 1994. Hydrolysis of oleuropein by Lactobacillus
plantarum strains associated with olive fermentation. Appl. Environ.
Microbiol. 60, 4142–4147.
Coni, E., Di Benedetto, R., Di Pasquale, M., Masella, R., Modesti, D., Mattei, R., Carlini,
E.A., 2000. Protective effect of oleuropein, an olive oil biophenol, on low density
lipoprotein oxidizability in rabbits. Lipids 35, 45–54.
Covas, M.I., 2007. Olive oil and the cardiovascular system. Pharmacol. Res. 55,
175–186.
Ehrhardt, C., Hrincius, E.R., Korte, V., Mazur, I., Droebner, K., Poetter, A., Dreschers,
S., Schmolke, M., Planz, O., Ludwig, S., 2007. A polyphenol rich plant
extract, CYSTUS052, exerts anti influenza virus activity in cell culture without
toxic side effects or the tendency to induce viral resistance. Antiviral Res. 76,
38–47.
Espín, J.C., Soler-Rivas, C., Cantos, E., Tomás-Barberán, F.A., Wichers, H.J., 2001. Synthesis
of the antioxidant hydroxytyrosol using tyrosinase as biocatalyst. J. Agric.
Food Chem. 49, 1187–1193.
Eto, M., Mase, M., 2003. Isolation of the Newcastle disease virus and the H9N2
influenza A virus from chicken imported from China. J. Jpn. Vet. Med. Assoc.
56, 333–339.
Furneri, P.M., Marino, A., Saija, A., Uccella, N., Bisignano, G., 2002. In vitro
antimycoplasmal activity of oleuropein. Int. J. Antimicrob. Agents 20,
293–296.
Lee-Huang, S., Zhang, L., Huang, P.L., Chang, Y.T., Huang, P.L., 2003. Anti-HIV activity
of olive leaf extract (OLE) and modulation of host cell gene expression by HIV-1
infection and OLE treatment. Biochem. Biophys. Res. Commun. 307, 1029–1037.
Lee-Huang, S., Huang, P.L., Zhang, D., Lee, J.W., Bao, J., Sun, Y., Chang, Y.T., Zhang,
J., Huang, P.L., 2007a. Discovery of small-molecule HIV-1 fusion and integrase
inhibitors oleuropein and hydroxytyrosol: Part I. Fusion inhibition. Biochem.
Biophys. Res. Commun. 354, 872–878.
Lee-Huang, S., Huang, P.L., Zhang, D., Lee, J.W., Bao, J., Sun, Y., Chang, Y.T., Zhang,
J., Huang, P.L., 2007b. Discovery of small-molecule HIV-1 fusion and integrase
inhibitors oleuropein and hydroxytyrosol. Part II. Integrase inhibition. Biochem.
Biophys. Res. Commun. 354, 879–884.
Ma, S.C., He, Z.D., Deng, X.L., But, P.P., Ooi, V.E., Xu, H.X., Lee, S.H., Lee, S.F., 2001. In
vitro evaluation of secoiridoid glucosides from the fruits of Ligustrum lucidum as
antiviral agents. Chem. Pharm. Bull. 49, 1471–1473.
Micol, V., Caturla, N., Pérez-Fons, L., Más, V., Pérez, L., Estepa, A., 2005. The olive
leaf extract exhibits antiviral activity against viral haemorrhagic septicaemia
rhabdovirus (VHSV). Antiviral Res. 66, 129–136.
Miró-Casas, E., Farré Albaladejo, M., Covas, M.I., Rodriguez, J.O., Menoyo Colomer,
E., Lamuela Raventós, R.M., de la Torre, R., 2001. Capillary gas chromatographymass
spectrometry quantitative determination of hydroxytyrosol and tyrosol in
human urine after olive oil intake. Anal. Biochem. 294, 63–72.
Pacheco, Y.M., Bermúdez, B., López, S., Abia, R., Villar, J., Muriana, F.J., 2007. Minor
compounds of olive oil have postprandial anti-inflammatory effects. Br. J. Nutr.
29, 1–4.
Tuck, K.L., Tan, H.W., Hayball, P.J., 2000. Synthesis of tritium-labeled hydroxytyrosol,
a phenolic compound found in olive oil. J. Agric. Food Chem. 48, 4087–4090.
Visioli, F., Bellomo, G., Montedoro, G., Galli, C., 1995. Low density lipoprotein oxidation
is inhibited in vitro by olive oil constituents. Atherosclerosis 117, 25–32.
Visioli, F., Bellomo, G., Galli, C., 1998a. Free radical-scavenging properties of olive oil
polyphenols. Biochem. Biophys. Res. Commun. 247, 60–64.
Visioli, F., Bellosta, S., Galli, C., 1998b. Oleuropein, the bitter principle of olives,
enhances nitric oxide production by mouse macrophages. Life Sci. 62, 541–546.
Visioli, F., Caruso, D., Plasmati, E., Patelli, R., Mulinacci, N., Romani, A., Galli, G., Galli,
C., 2001. Hydroxytyrosol, as a component of olive mill waste water, is dosedependently
absorbed and increases the antioxidant capacity of rat plasma. Free
Radic. Res. 34, 301–305.




Kentaro Yamadaa,1, Haruko Ogawaa,!, Ayako Haraa, Yukio Yoshidaa, Yutaka Yonezawaa,
Kazuji Karibea, Vuong Bui Nghiaa, Hiroyuki Yoshimurab, Yu Yamamotoc,
Manabu Yamadac, Kuniyasu Nakamurac, Kunitoshi Imaia

A Research Center, Obihiro University of Agriculture and Veterinary Medicine, Obihiro, Hokkaido 080-8555, Japan
b Eisai Food & Chemical Co., Ltd., Nihonbashi, Tokyo 103-0027, Japan
c Research Team of Viral Diseases, National Institute of Animal Health, Tsukuba, Ibaraki 305-0856, Japan

Article history:
Received 29 September 2008
Received in revised form 28 February 2009
Accepted 11 March 2009

Mechanism of the antiviral effect of hydroxytyrosol on influenza virus appears to involve morphological change of the virus.

Yamada K, Ogawa H, Hara A, Yoshida Y, Yonezawa Y, Karibe K, Nghia VB, Yoshimura H, Yamamoto Y, Yamada M, Nakamura K, Imai K.

Research Center for Animal Hygiene and Food Safety, Obihiro University of Agriculture and Veterinary Medicine, Obihiro, Hokkaido, Japan.