Effects of the Glycosylation of the HA Protein of H9N2 Subtype Avian Influenza Virus on the Pathogenicity in Mice and Antigenicity

Abstract

As the H9N2 subtype avian influenza virus (H9N2 AIV) evolves naturally, mutations in the hemagglutinin (HA) protein still occur, which involves some sites with glycosylations. It is widely established that glycosylation of the H9N2 AIV HA protein has a major impact on the antigenicity and pathogenicity of the virus. However, the biological implications of a particular glycosylation modification site (GMS) have not been well investigated. In this study, we generated viruses with different GMSs based on wild-type (WT) viruses. Antigenicity studies revealed that the presence of viruses with a 200G⁺/295G⁻ mutation (with glycosylation at position 200 and deletion of glycosylation at position 295 in the HA protein) combined with a single GMS, such as 87G⁺, 127G⁺, 148G⁺, 178G⁺, or 265G⁺, could significantly affect the antigenicity of the virus. Pathogenicity assays revealed that the addition of GMS, such as 127G⁺, 188G⁺, 148G⁺, 178G⁺, or 54G⁺, decreased the virulence of the virus in mice, except for 87G⁺. The removal of GMS, such as 280G⁻ or 295G⁻, increased the pathogenicity of the virus in mice. Further studies on pathogenicity revealed that 87G⁺/295G⁻ could also enhance the pathogenicity of the virus. Finally, we selected the WT, WT-87G⁺, WT-295G⁻, and WT-87G⁺/295G⁻ strains as our further research targets to investigate the detailed biological properties of the viruses. GMS, which can enhance viral pathogenicity, did not significantly affect replication or viral stability in vitro but significantly promoted the expression of proinflammatory factors to enhance inflammatory responses in mouse lungs. These findings further deepen our understanding of the influence of the glycosylation of the HA protein of H9N2 AIV on the pathogenicity and antigenicity of the virus in mice.

Full text

Research Article
Effects of the Glycosylation of the HA Protein of H9N2
Subtype Avian Influenza Virus on the Pathogenicity in
Mice and Antigenicity
Bing Liang ,
1
Menglu Fan ,
1
Qi Meng ,
1
Yaping Zhang ,
2
Jiayu Jin ,
1
Na Chen ,
1
Yuanlu Lu ,
1
Chenfeng Jiang ,
1
Xingxing Zhang ,
1
Zongyou Zou ,
1
Jihui Ping ,
1
and
Juan Su
1
1
MOE International Joint Collaborative Research Laboratory for Animal Health and Food Safety and
Jiangsu Engineering Laboratory of Animal Immunology, College of Veterinary Medicine, Nanjing Agricultural University,
Nanjing, 210095, China
2
State Key Laboratory of Veterinary Biotechnology, Harbin Veterinary Research Institute in CAAS, Harbin, China
Correspondence should be addressed to Jihui Ping; jihui.ping@njau.edu.cn and Juan Su; sujuan@njau.edu.cn
Received 16 December 2023; Revised 10 April 2024; Accepted 13 April 2024; Published 17 May 2024
Academic Editor: Chunfu Zheng
Copyright ©2024 Bing Liang et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
As the H9N2 subtype avian influenza virus (H9N2 AIV) evolves naturally, mutations in the hemagglutinin (HA) protein still occur,
which involves some sites with glycosylations. It is widely established that glycosylation of the H9N2 AIV HA protein has a major
impact on the antigenicity and pathogenicity of the virus. However, the biological implications of a particular glycosylation
modification site (GMS) have not been well investigated. In this study, we generated viruses with different GMSs based on
wild-type (WT) viruses. Antigenicity studies revealed that the presence of viruses with a 200G
+
/295G
−
mutation (with glycosyla-
tion at position 200 and deletion of glycosylation at position 295 in the HA protein) combined with a single GMS, such as 87G
+
,
127G
+
,148G
+
,178G
+
, or 265G
+
, could significantly affect the antigenicity of the virus. Pathogenicity assays revealed that the
addition of GMS, such as 127G
+
, 188G
+
, 148G
+
,178G
+
, or 54G
+
, decreased the virulence of the virus in mice, except for 87G
+
. The
removal of GMS, such as 280G
−
or 295G
−
, increased the pathogenicity of the virus in mice. Further studies on pathogenicity revealed
that 87G
+
/295G
−
could also enhance the pathogenicity of the virus. Finally, we selected the WT, WT-87G
+
, WT-295G
−
, and WT-
87G
+
/295G
−
strains as our further research targets to investigate the detailed biological properties of the viruses. GMS, which can
enhance viral pathogenicity, did not significantly affect replication or viral stability in vitro but significantly promoted the expression
of proinflammatory factors to enhance inflammatory responses in mouse lungs. These findings further deepen our understanding of
the influence of the glycosylation of the HA protein of H9N2 AIV on the pathogenicity and antigenicity of the virus in mice.
1. Introduction
SARS-CoV-2 has recently become common, and there have
been cases of coinfection with influenza viruses that have
resulted in more serious clinical symptoms, posing a signifi-
cant threat to public health [1]. Moreover, the damage
caused by H9N2 AIV is far-reaching [2]. First, poultry
infected with H9N2 AIV will display clinical symptoms
such as egg reduction. Typically, a substantial amount of
poultry is kept in large and intensive environments on farms,
and due to the high transmissibility of low-pathogenicity
avian influenza virus, the economic impact of an outbreak
of AIV can be incredibly damaging [3–5]. In addition, chick-
ens infected with AIV are more likely to be coinfected with
other pathogens due to the low resistance of the immune
system, resulting in a high fatality rate [6, 7]. People with
high-risk factors, such as farm breeders who come into con-
tact with poultry and those who work in live poultry markets,
have an increased likelihood of being infected with AIVs
with low pathogenicity. In these populations, H9N2 AIV is
often detected in the serum, yet infected individuals usually
do not exhibit any clinical signs [4, 8–11]. In 2013, according
Hindawi
Transboundary and Emerging Diseases
Volume 2024, Article ID 6641285, 18 pages
https://doi.org/10.1155/2024/6641285

to the gene sequences of H7N9 and H10N8 strains isolated
from humans, six genes were identified from H9N2 AIV.
Therefore, H9N2 AIV also acts as a genetic contributor
[12–15]. Vaccines have been used to immunize chickens in
China since 1998. However, with the continuous antigenic
evolution of H9N2 AIV, it has been difficult for vaccines to
provide comprehensive protection for chickens [16, 17].
H9N2 AIV has become the most widespread subtype of
influenza virus in China and has seriously affected the devel-
opment of the poultry industry [18]. H9N2 AIV is continu-
ously mutating under chronic immune stress. One of the key
factors that can evade vaccine immunity is the HA gene [19].
These phenomena are closely related to the glycosylation of
the H9N2 AIV HA protein, as the deletion or addition of
GMS to the HA protein affects antigenicity, pathogenicity,
and even receptor binding [20–22]. Collectively, these find-
ings suggest that we must be mindful of the potential inter-
species transmission ability of H9N2 AIV and investigate the
changes in antigenicity and pathogenicity caused by the gly-
cosylation of the H9N2 AIV HA protein during the evolution
of the virus [23, 24].
Glycosylation of proteins has gradually been recognized
as an important process during viral evolution [25]. AIV
produces two membrane-bound surface glycoproteins, hem-
agglutinin (HA) and neuraminidase (NA), which include N-
linked oligosaccharides. HA is a trimer, and its monomer
consists of the head (HA1) and stem (HA2) regions. N-
glycosylation is a kind of posttranslational modification of
mammalian glycoproteins that involves the attachment of
oligosaccharides to Asn residues of the Asn-X-Ser/Thr-Y
motif, where X/Y can be any amino acid other than proline
[26]. The number of GMSs on the HA protein of the human
seasonal influenza virus has risen with time [27]. Identifying
potential glycosylation sites is an effective strategy for pre-
venting host immunological pressure on influenza viruses
because the glycosylation of HA normally influences the
antigenicity and pathogenicity of the virus [28].
Glycosylation of the HA protein blocks antibodies from
binding, thus protecting against viruses [29]. The GMS on
the HA globular head has been revealed to be particularly
important for the antigenicity and receptor-binding charac-
teristics of influenza virus [22, 30]. GMS differences in H3N2
subtype influenza viruses impact the antibody neutralization
response to influenza vaccine strains, decreasing the efficacy
of seasonal influenza vaccines [31]. The insertion of a novel
N-linked oligosaccharide on the HA of H3N2 subtype influ-
enza viruses resulted in immunological escape from antibody
pressure by modifying their antigenicity [32].
The glycosylation of the influenza virus HA protein is an
essential factor in viral pathogenicity. The introduction of
glycosylation at site 158 (H9: position 152) of the H5N1
subtype influenza virus can promote viral production and
intensify the host response, thus increasing its pathogenicity
in mice [33]. The presence of N127D on the HA of H9N2
AIV suggested that the glycosylation at 127 was eliminated,
which decreased the pathogenicity of the virus in mice [22].
The HA glycosylation of H3N2 subtype influenza viruses has
been shown to increase annually, and the increase in GMS
has resulted in a decrease in its pathogenicity in mice. The
more glycosylated modifications there are on the virus sur-
face, the more sensitive the virus is to lectin in the host body,
and it is easier for lectin to neutralize and inhibit the influ-
enza virus [34].
Distinct GMSs can be found in various strains of differ-
ent subtypes of influenza virus [35]. Since the GMS of HA
has varying effects on viruses and there are different GMSs
on the HA protein that exist naturally, only a few sites have
been identified in existing reports and still need to be better
understood. Thus, it is highly important to investigate the
biological significance of these GMSs comprehensively. In
this study, the HA gene from the A/chicken/Anhui/99/2017
strain was utilized to construct the pHH21-99-HA plasmid as
a site-specific mutation template to rescue these different
GMS viruses. Then, we explored the effects of different
GMSs on the HA of H9N2 AIV on the antigenicity and path-
ogenicity of the virus and screened out the functional GMS.
2. Materials and Methods
2.1. Ethics Statement. All experimental animals used in this
investigation were housed in labs with level 2 biosafety. Mice
were treated humanely in accordance with the People’s
Republic of China’s Animal Ethics Regulations and Guide-
lines as well as the regulations of Nanjing Agricultural Uni-
versity’s Animal Protection and Use Committee (SYXK(Su)
2021-0086).
2.2. Cells and Viruses. This study used Madin–Darby canine
kidney (MDCK) cells, human embryonic kidney cells (293T),
human lung adenocarcinoma epithelial (A549) cells, and Vero
cells. The virus titers in MDCK cells were measured using a
plaque assay. The medium used was high-glucose DMEM
(Gibco) supplemented with 10%FBS, and the cells were incu-
bated at 37°Cand5%CO
2
. This study used a 12-plasmid reverse
genetic technique to create recombinant viruses, with the inter-
nal genes derived from A/chicken/Jiangsu/875/2018 and the HA
and NA genes from A/chicken/Anhui/99/2017 [36].
2.3. Construction of a Three-Dimensional Model of HA
Protein. The amino acid sequence of the A/chicken/Anhui/
LH99/2017 strain was added to the SWISS-MODEL website
(https://swissmodel.expasy.org/interactive) to generate a tri-
meric model of the HA protein, and this model was subse-
quently imported into PyMOL software (DeLano Scientific
LLC) for simulation. Different glycosylation sites and RBS
regions are marked with the indicated colors.
2.4. Generation of GMS HA Plasmids. Using pHH21-99-HA
as the template, all plasmids with different GMSs were con-
structed by site-specific mutation, and Oligo 7.0 software was
used to design specific primers. The detailed primer informa-
tion is provided in Table S3 in Supplementary Materials. The
following materials were added to 0.2 ml PCR tubes: dNTP
mix (10 mM). Next, 1 μlofPhanta
®
Max Buffer, 25 μlof
Phanta
®
Max Super-Fidelity DNA Polymerase, 50 μlof
ddH
2
O, 500 ng of plasmid template, and 1 μl each of upstream
and downstream mutant primers were added. The PCR con-
ditions were as follows: 5 min of predenaturation at 95°C, 30 s
2 Transboundary and Emerging Diseases

of denaturing at 95°C, 30 s of annealing at 58°C, and 25 cycles
of extension at 72°C (1,000 bp/1 min). After an additional
10 min at 72°C, the temperature was maintained at 16°C.
The following reaction system was used to digest the meth-
ylated plasmid template using Dpn I restriction endonucle-
ase: 10x buffer (5 μl), Dpn I enzyme (1 μl), and 50 μlofPCR
products. Reaction conditions: 30 min in a water bath at
37°C. Subsequently, the digested products were trans-
formed into E. coli DH5α, and then, positive monoclonal
colonies were selected. Positive plasmid screening and
sequencing were used to verify the success of the point
mutation.
2.5. Virus Rescue. Several GMS viruses were rescued using
the 12-plasmid reverse genetic method [36]. pCAGGS-
WSN-PB2, pCAGGS-WSN-PB1, pCAGGS-WSN-PA, and
pCAGGS-WSN-NP (1 μg each) were used. pHH21-99-HA (var-
ious GMS HAs), pHH21-99-NA, pHH21-875-PB2, pHH21-
875-PB1, pHH21-875-PA, pHH21-875-NP, pHH21-875-M,
and pHH21-875-NS (0.2 μg each) were used. The plasmids
were transfected into 293T cells at an 80%confluence using
Lipo2000 (2 μl per time, Invitrogen), and the cells were subse-
quently grown in dishes with a diameter of 35 mm. TPCK tryp-
sin was added 24 hr after transfection, and the cell suspension
was inoculated into SPF chicken embryos aged 9–11 days at
48 hr after transfection and incubated in a 37°Cincubatorfor
48 hr for virus proliferation. The allantoic fluid of the chicken
embryo was collected, and then, the HA titers were determined.
The virus titer was calculated by the number of plaques it formed
on MDCK cells and was expressed as PFU/ml.
2.6. Antigenic Analysis. A total of 30 GMS viruses, WT, R-
1–R-29 (different GMS viruses with various glycosylation
sites are shown in Table 1) virus allantoic fluid, were added
to β-propiolactone (at a 1 : 2,000 vol:vol ratio) to inactivate
the virus in this way. Then, the inactivated viruses were made
into oil emulsion-inactivated vaccines by adding Tween 80
and Span, and SPF chickens aged 4–6 weeks were immunized
TABLE 1: Detailed information on different GMS viruses.
Serial number Detailed glycosylated site added or deleted
under WT background Hemagglutinin titer (log2) Virus titer (PFU/ml)
WT (11G
+
, 123G
+
, 280G
+
, 287G
+
, and
295G
+
)
a,b
11 8.5 ×10
8
R-1 WT-54G
+
11 5 ×10
8
R-2 WT-54G
+
/188G
+
10 4.15 ×10
8
R-3 WT-87G
+
9 9.5 ×10
7
R-4 WT-87G
+
/188G
+
/295G
−
11 2.05 ×10
8
R-5 WT-87G
+
/200G
+
8 2.05 ×10
7
R-6 WT-87G
+
/188G
+
/200G
+
/295G
−
12 7.5 ×10
8
R-7 WT-127G
+
/295G
−
10 1.15 ×10
8
R-8 WT-127G
+
10 5.5 ×10
8
R-9 WT-127G
+
/200G
+
/295G
−
10 2.6 ×10
8
R-10 WT-127G
+
/200G
+
10 1.25 ×10
8
R-11 WT-127G
+
/178G
+
10 1.05 ×10
9
R-12 WT-148G
+
10 6.5 ×10
8
R-13 WT-148G
+
/200G
+
/295G
−
9 1.45 ×10
9
R-14 WT-148G
+
/200G
+
10 1.1 ×10
9
R-15 WT-178G
+
11 9.5 ×10
8
R-16 WT-178G
+
/200G
+
/295G
−
12 1.15 ×10
9
R-17 WT-178G
+
/200G
+
11 4.7 ×10
8
R-18 WT-178G
+
/280G
+
11 4.15 ×10
8
R-19 WT-188G
+
11 8 ×10
8
R-20 WT-188G
+
/200G
+
/295G
−
12 1.95 ×10
9
R-21 WT-188G
+
/200G
+
10 2.65 ×10
8
R-22 WT-200G
+
11 8 ×10
8
R-23 WT-265G
+
/200G
+
/295G
−
10 2.5 ×10
8
R-24 WT-265G
+
/200G
+
12 7.5 ×10
8
R-25 WT-267G
+
/200G
+
/295G
−
11 2.65 ×10
8
R-26 WT-265G
+
11 7 ×10
8
R-27 WT-267G
+
11 8 ×10
8
R-28 WT-280G
−
10 1.95 ×10
8
R-29 WT-295G
−
11 2.65 ×10
8
a
Position numbers are according to H9 numbering.
b
The WT protein originally contains four N-glycosylated sites, 11G
+
, 123G
+
, 280G
+
, 287G
+
, and 295G
+
.
Transboundary and Emerging Diseases 3

with multiple subcutaneous and intramuscular injections. At
21 days after immunization, blood samples were collected,
and the serum was separated. Cross-hemagglutination inhi-
bition experiments were subsequently conducted. An antigen
map was drawn according to the hemagglutination inhibi-
tion results (https://acmacs-web.antigenic-cartography.org/).
2.7. Mouse Experiments. The Xipuer-Bikai Experimental
Animal Company supplied pathogen-free 4-week-old female
BALB/c mice (Shanghai, China). Eight mice per group were
intranasally inoculated with a volume of 50 μl containing 4 ×
10
5
PFU of various glycosylated recombination viruses, and
their body weights were monitored until the 14th day post-
infection. Mice in the negative control group received 50 μlof
PBS solution intranasally. On the third day after infection,
lung samples were taken from three mice and used to titrate
the virus. The mice were euthanized when their weights were
less than 75%of their initial body weights. The median lethal
dose test (MLD50) for GMS viruses in mice: Three mice were
infected with recombinant viruses at doses of 10
4
,10
5
, and
10
6
PFU in a 50 μl volume. The mice were fed for 14 days,
after which their daily weight changes and mortality rates
were recorded. The MLD50 was estimated using Reed and
Muench’s approach.
2.8. Western Blotting. To extract the total protein from the
virus, a single layer of MDCK cells at a density of 95%was
infected with the recombinant virus at an MOI of 1. The cells
were then lysed with NP40 18–24 hr after infection. After being
mixed with 4x loading buffer (Solarbio, Beijing), the treated
cells were heated to 100°C for 15 min to denature them. Fol-
lowing separation via 10%polyacrylamide-containing SDS‒
PAGE with 10 μl of each extract, the proteins were transferred
to an acetate membrane (GE Healthcare, Amersham), which
was subsequently blocked with 5%skim milk and treated over-
nightat4°C with a 1 : 1,000 laboratory-made anti-H9N2 HA
monoclonal antibody. The membrane was washed five times
before being incubated for 1 hr at 37°C with a sheep antimouse
secondary antibody conjugated to HRP (1 : 10,000; KPL,
Gaithersburg, MD). After that, the membrane was washed
again, and enhanced chemiluminescence (Vazyme, Nanjing)
was used to detect the target protein bands.
2.9. Receptor-Binding Assay. Two glycopolymers were uti-
lized in solid-phase direct binding studies to assess the
receptor-binding properties of the virus. They were α-2,3-sialyl-
glycopolymer [Neu5Acα2-3Galb1-4GlcNAcb1-pAP (para-ami-
nophenyl)-α-polyglutamic acid (α-PGA)]. The molecule
Neu5Acα2-6Galb1-4GlcNAcb1-pAP (para-aminophenyl)-
alpha-polyglutamic acid (α-PGA) is also present [37, 38]. In
this investigation, a goat anti-chicken antibody (Sigma‒Aldrich,
St. Louis, MO, USA) and horseradish peroxidase (HRP) were
used for ELISAs with chicken serum generated from an
inactivated virus. A microplate reader was used to measure the
absorbance at 490 nm. The dose–response curves of virus
binding to glycopolymers were analyzed using a single-site
binding algorithm and curve fitting with GraphPad Prism to
obtain the constant association values (Ka). The values
displayed are the average Æstandard deviation of three
individual tests, each carried out in triplicate.
2.10. Real-Time RT-PCR. To detect four cytokines, namely,
IL-1β, IL-6, TNF-α, and IFN-β, the ID numbers of each
cytokine were obtained from the National Center for Bio-
technology Information (NCBI) database and then input
into GenBank, with house mice as the species used for
primer retrieval. The top cytokine primer sequences were
selected and are shown in Table S4 in Supplementary Mate-
rials. One milliliter of PBS was added to the mouse lung
sample, the mixture was ground, and the supernatant was
obtained by centrifugation at 12,000 rpm for 10 min at 4°C.
The TRIzol method was used to extract total RNA, which
was subsequently reverse-transcribed into cDNA (Vazyme,
Nanjing). Using a Light Cycler
®
96 system, qRT-PCR was
carried out using a 20 μl system, 2 μl of cDNA, and 1 μlof
each of the downstream and upstream primers. Briefly, 10 μl
of 2x AceQ qPCR Probe Master Mix and 6 μl of ddH
2
O were
added, followed by 40 cycles of 95°C for 5 min, 95°C for 10 s,
and 60°C for 30 s. The relative expression of mRNA was
standardized to the aldehyde-3-phosphate dehydrogenase
(GAPDH) level using the 2
−ΔΔCt
method.
2.11. Viral Growth Kinetics Analysis. Monolayered MDCK
cells and A549 cells were inoculated with WT, WT-87G
+
,
WT-295G
−
, or WT-87G
+
/295G
−
at MOIs of 0.01 and 0.5,
respectively. The cell supernatants were collected at 12, 24,
48, and 60 hr after inoculation, and plaque tests were used to
measure the virus titer and plot the virus growth curve on
several cell lines.
2.12. Thermal Stability Assay. The HA titer of several recom-
binant viruses was determined by incubating them in a water
bath at 56°C for 0, 5, 10, 15, 30, 60, 90, 120, and 150 min. The
thermal stability curves of different recombinant viruses
were constructed according to the HA titer losses.
2.13. pH Stability. Equal amounts of 100 mM phosphate
buffer (pH =6.0), 100 mM acetate buffer (pH =4.0 and pH
=5.0), or neutral phosphate buffer (pH =7.0) were mixed
with the viruses. Following a 10-minute incubation period at
37°C, the mixtures underwent HA titration.
2.14. Cell Fusion Assay. Vero cells were infected with WT,
WT-87G
+
, WT-295G
−
, or WT-87G
+
/295G
−
at an MOI of 3,
cultured for 16–24 hr, treated with 2 μg/ml TPCK trypsin for
15 min, washed with PBS, and subsequently incubated in
PBS at various pH values (pH: 5.0–6.0) for 10 min before
being cultured in DMEM supplemented with 5%FBS for
3 hr. Then, 70%ethanol was added to fix the cells for
10 min, followed by Giemsa staining for 30 min to enable
observation of cell fusion under a microscope.
2.15. Statistical Analyses. All the data were examined with
GraphPad Prism software. Using Student’st-test, significant
differences between various experimental groups were ascer-
tained. At P<0:05, differences were considered to be statis-
tically significant.
4 Transboundary and Emerging Diseases

3. Results
3.1. GMS on HA1 of H9N2 AIV. HA proteins have various
GMSs located in different areas, such as the head or stem
region. In addition, whether there are glycosylations at spe-
cific sites will vary over time, which affects the biological
characteristics of the virus. As a result, we evaluated the
distribution and statistical fraction of all possible GMS on
HA1 of H9N2 AIV. The whole HA gene sequences of all
H9N2 AIV strains from 1994 to 2021 were retrieved from
the GISAID database (https://www.gisaid.org). By analyzing
these sequences, it was discovered that a total of 15 potential
GMSs had appeared on HA1 since the epidemic of H9N2
AIV. These sites were located at amino acid positions 11–13,
54−56, 87–89, 123−125, 127–129, 148−150, 178–180, 188
−190, 200–202, 238−240, 265–267, 267−269, 280–282, 287
−289, and 295–297 (H9 numbering). PyMOL was used to
determine the exact locations of these sites on the HA trimer,
as shown from different perspectives in Figures 1(a) and 1(b).
Three sites (11–13, 280−282, and 295–297) were located at
the stem region of HA, while the remaining nine were
positioned in the head region. Additionally, 87–89, 127
−129, and 178–180 were close to the receptor-binding
region (RBS) of the head of HA.
By analyzing the glycosylation proportions of potential
GMSs from 1994 to 2021, the fluctuation trends of both
global and domestic potential GMS were found to be similar,
with most of them being conservative. Most commonly,
11–13, 123−125, 280–282, and 287–289 were glycosylated,
while 54–56, 87−89, 127–129, 148−150, 178–180, 188−190,
238–240, 265−267, and 267–269 were not glycosylated, as
depicted in Figures 1(c) and 1(d). Detailed information is
provided in Tables S1 and S2. As shown in Figure 1(e),
87–89, 200−202, and 295–297 exhibited considerable fluctua-
tions over time. 87−89: The proportion of glycosylated pro-
ducts increased from 1996 to 2004, followed by a decrease
from 2005 to 2018, but it has increased since 2019. 200
−202: There is a degree of instability, and the overall trend
is decreasing. 295−297: This number has shown an increasing
trend since 2005. However, since 2019, it has been decreasing.
3.2. Generation of GMS Viruses. To explore the influence of
different GMSs on virus antigenicity and pathogenicity, the
initial step was to rescue different GMS viruses. We designed
GMS in its opposite form to determine the impact of the
mutation on the virus. We also speculated that 200G
+
/
295G
−
and other single-point superpositions would further
affect the biological characteristics of the virus. Therefore, we
designed the different GMS mutation forms shown in
Table 1. Here, we used a 12-plasmid reverse genetics system
to rescue different GMS viruses. The HA and NA genes were
from A/Chicken/Jiangsu/875/2018, the internal genes were
from A/Chicken/Jiangsu/875/2018, and the polymerase genes
were from A/WSN/33(H1N1). All these genes were used to
rescue WT. The HA gene of A/Chicken/Anhui/99/2017 was
utilized as a pattern to carry out site-specific mutation to
obtain HAs with diverse glycosylations, which were generated
to rescue different GMS viruses. The glycosylation of recom-
binant viruses with particular mutation sites was successfully
rescued, and the corresponding viral abbreviations are listed
in Table 1. The HA and NA genes of the rescued viruses were
extensively sequenced to confirm that no unexpected muta-
tions occurred. WT was initially composed of a few GMSs,
such as 11G
+
, 123G
+
, 280G
+
, 287G
+
, and 295G
+
. Rescued
different GMS viruses included only 11 of the 15 GMSs, as
viruses containing 11G
+
, 123G
+
,238G
+
, and 287G
+
could
not be saved, or the virus titers were too low to conduct
further experiments.
3.3. Antigenicity Analysis of GMS Viruses. The oligosacchar-
ides present on the surface of the HA protein can impede the
neutralizing effect of neutralizing antibodies on the virus,
thus affecting its antigenicity. To explore the effects of vari-
ous GMSs on antigenicity, we prepared different GMS
viruses as inactivated oil emulsion vaccines to immunize
chickens and subsequently isolated immune serum to per-
form cross-hemagglutination inhibition experiments [39].
Figure 2(b) displays the antigen map, which was created
based on the results of cross-hemagglutination inhibition
experiments. We could assess the impact of various GMSs
on the antigenicity of the virus by comparing the relative
position between the GMS viruses and the WT on the anti-
gen map. Upon analysis of the antigen map, the following
could be observed: (1) All the mutant viruses showed a diver-
gent state compared to WT, with the furthest antigenic dis-
tance from WT being roughly two cells. (2) 127G
+
, 148G
+
,
and 178G
+
, which were near the RBS region, generated
greater antigenic divergence. (3) Most GMSs (e.g., 87G
+
,
127G
+
, 148G
+
, 178G
+
, and 265G
+
) that were combined
with 200G
+
/295G
−
had greater antigenic divergence than
those that were combined with 200G
+
alone. (4) 127G
+
and 178G
+
(as well as their related viruses) caused the great-
est differences on the map. (5) Single GMS viruses such as
54G
+
, 87G
+
, and 188G
+
did not lead to a greater antigen
distance than did WT. However, if 54G
+
or 87G
+
was com-
bined with 188G
+
(e.g., 54G
+
/188G
+
and 87G
+
/188G
+
), the
antigen distances on the map were greater.
3.4. Pathogenicity Analysis of GMS Viruses. Figure S1 thor-
oughly illustrates the comparison between the body weight
changes, mortality, and viral titer of mouse lungs of a partic-
ular GMS and its related combination viruses. Figure S1(a, c,
d, e, f) shows that 54G
+
, 127G
+
, 148G
+
, 178G
+
, and 188G
+
could attenuate the pathogenicity of the WT strain in mice.
However, 87G
+
, 265G
+
, 267G
+
, 280G
−
, and 295G
−
may
increase the pathogenicity of the WT strain in mice (Figure
S1(b, g, h, i, j). Moreover, combining single GMS, such as
87G
+
, 127G
+
, 178G
+
, 188G
+
, 265G
+
, and 267G
+
, with
200G
+
/295G
−
, such as 127G
+
/200G
+
/295G
−
, 178G
+
/
200G
+
/295G
−
, 265G
+
/200G
+
/295G
−
, and 267G
+
/200G
+
/
295G
−
, resulted in greater pathogenicity in mice. Notably,
127G
+
/200G
+
/295G
−
, 148G
+
/200G
+
/295G
−
, 178G
+
/200G
+
/
295G
−
, and 265G
+
/200G
+
/295G
−
can create greater anti-
genic distances on the antigen map. In summary, 127G
+
/
200G
+
/295G
−
, 178G
+
/200G
+
/295G
−
, and 265G
+
/200G
+
/
295G
−
had significant impacts on both pathogenicity and
antigenicity.
Transboundary and Emerging Diseases 5

90°
11–13
54–56
87–89
287–289
280–282
267–269
265–267
295–297
RBS
ðaÞ
188–190
200–202
148–150
123–125
127–129
178–180
238–240
ðbÞ
G
+
G
–
11 54 87 123 127 148 178 188 200 238 265 267 280 287 295
0
10
20
30
40
50
60
70
80
90
100
Domestic
All potential glycosylated modication sites in H9N2 AIV HA1
Percent of glycosylation
modication (%)
ðcÞ
Global
11 54 87 123 127 148 178 188 200 238 265 267 280 287 295
All potential glycosylated modication sites in H9N2 AIV HA1
0
10
20
30
40
50
60
70
80
90
100
Percent of glycosylation
modication (%)
G
+
G
–
ðdÞ
FIGURE 1: Continued.
6 Transboundary and Emerging Diseases

The glycosylation of the influenza virus HA protein is an
important factor in determining viral pathogenicity. To
assess the pathogenicity of different GMS viruses in mice,
mice were intranasally inoculated with different GMS viruses
at a concentration of 4 ×10
5
PFU in 50 μl. Taking into
account the body weight change, mortality, and viral titer
of mouse lungs, the pathogenicity of different GMS strains
was comprehensively ranked. The most powerful GMS
strains that could increase the pathogenicity of the virus to
mice were 295G
−
, 280G
−
, 87G
+
, 265G
+
, 267G
+
, and 200G
+
.
The GMSs that attenuated the pathogenicity of the virus to
mice, from most powerful to least, were 127G
+
, 188G
+
,
148G
+
, 178G
+
, and 54G
+
(Figures 3(c) and 3(d)). It is evi-
dent that the 87G
+
, 87G
+
/200G
+
, 127G
+
/200G
+
/295G
−
,
265G
+
/200G
+
/295G
−
, 265G
+
/200G
−
, and 267G
+
/200G
+
/
295G
−
groups had significantly greater virus titers in mouse
lungs, as illustrated in Figure 3(b).
3.5. The Functional Glycosylation Sites Were at Positions
295–297, 280−282, and 87–89. According to the results of
the above pathogenicity experiments, three GMSs with the
most significant effect on the pathogenicity of mice were
selected: 295G
−
, 280G
−
, and 87G
+
. It is necessary to test
whether these three GMSs are functional because these sites
may not be glycosylated due to steric hindrance, and neigh-
boring amino acids may further hinder glycan attachment
[26]. Consequently, we explored the migration rates of the
WT-295G
−
, WT-280G
−
, and WT-87G
+
strains. As shown in
Figure 4(a), compared with those of the WT, the migration
rates of the 87G
+
strain slightly decreased, while the migra-
tion rates of the 280G
−
and 295G
−
strains slightly increased.
Consequently, the three selected GMSs were confirmed to be
functional glycosylation sites. There are oligosaccharides
attached to these positions that perform various functions.
To investigate whether there were any AIVs with other
combinations of these three glycosylation sites in nature
(87G
+
, 280G
−
, and 295G
−
) and whether viruses with a com-
bination of these sites are more pathogenic to mammals, as
WT-87G
+
, WT-280G
−
, and WT-295G
−
individually were
more pathogenic than WT. Based on these questions, upon
analyzing the HA sequences of H9N2 AIV, it was discovered
that there were three kinds of combinations of these three
GMSs in nature: 87G
+
/295G
−
, 87G
+
/280G
−
/295G
−
, and
280G
−
/295G
−
without 87G
+
/280G
−
. Then, we successfully
rescued the WT-87G
+
/295G
−
, WT-87G
+
/280G
−
/295G
−
,
and WT-280G
−
/295G
−
strains and confirmed that there
were no unwanted mutations. Their specific biological char-
acteristics are listed in Table 2.
3.6. 87G
+
/295G
−
Further Increased Pathogenicity, while 280G
−
/
295G
−
Decreased Pathogenicity. To investigate whether recom-
binant viruses can further increase the pathogenicity of mice,
1994
1995
1996
1997
1998
1999
2000
2001
2002
2004
2003
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
Years
0
10
20
30
40
50
60
70
80
90
100
Percent of glycosylated
modification
(%)
11G
+
54G
+
87G
+
123G
+
127G
+
148G
+
178G
+
188G
+
200G
+
238G
+
265G
+
267G
+
280G
+
287G
+
295G
+
ðeÞ
FIGURE 1: Three-dimensional (3D) structures and analysis of N-glycosylated modification sites in the HA trimer protein of H9N2 AIV. (a, b)
HA trimer proteins are displayed from different angles. Different N-glycosylated modification sites are presented in different colors. The HA
trimeric protein is divided into HA1 heads and HA2 stems. The HA1 head is represented in gray. The HA stem is represented in light blue.
Olive labeling of GMS at positions 11–13, deep olive at 54–56, yellow at 87–89, wheat at 123–125, orange at 127–129, light pink at 148–150,
salmon at 178–180, warm pink at 188–190, purple at 200–202, sky blue at 238–240, marine blue at 265–267, TV blue at 267–269, pale green at
280–282, cyan at 287–289, and green at 295–297. All H9N2 HA amino acid sequences of Chinese or worldwide isolates (as of December
2021) were downloaded from GISAID (https://www.gisaid.org) and aligned using MAFFT. After excluding homologous and duplicate
sequences in BioEdit, the prevalence of each GMS was assessed. (c, d) Percent of all potential GMS from 1994 to 2021 domestically (light
purple and gray) or globally (light yellow and light green). (e) The variation trend of every potential GMS from 1994 to 2021. The color of
each potential GMS is the same as that of the 3D structure.
Transboundary and Emerging Diseases 7

Inoculation
21 days
Serum
Cross HI assay
ðaÞ
A/chicken/Henan/HP/1998
A/chicken/Shandong/6/1996
WT-127G+/178G+
WT-178G+/200G+
WT-127G+/200G+/295G–
A/chicken/Anhui/LH66/2017
A/chicken/Jiangsu/325/2018
WT-148G+
WT-148G+/200G+
WT-178G+/200G+/295G–
WT-87G+/188G+/200G+/295G–
WT
WT-148G
+
/200G
+
/295G
–
WT-178G+
WT-87G+/188G+/295G–
WT-87G+/200G+
WT-127G+/295G–
WT-127G+
ðbÞ
WT
WT-178G+/280G+
WT-54G+/188G+
WT-188G+
WT-188G+/200G+
WT-188G+/200G+/295G–
WT-87G+
WT-127G+/200G+
WT-265G+/200G+/295G–
WT-265G+/200G+
WT-265G+
A/chicken/Guangxi/55/2015
A/chicken/Jiangsu/WJ57/2012
WT-267G
+
/200G
+
/295G
–
WT-280G–
WT-267G+
WT-295G–
WT-200G+
A/chicken/Anhui/LH99/2017
WT-87G+/200G+
WT-87G+/188G+/295G–
WT-178G+
WT-87G+/188G+/200G+/295G–
WT-127G+
WT-127G+/295G–
WT-54G+
ðcÞ
FIGURE 2: Analysis of the antigenicity of different GMS viruses. (a) Different GMS viruses were prepared as inactivated oil emulsion vaccines,
and SPF chickens aged 4–6 weeks were immunized by intramuscular injection. Blood was collected 21 days after immunization, the serum
was separated, and cross-hemagglutination inhibition experiments were subsequently conducted. The plants were created with https://BioRe
nder.com. (b) Antigenic map of all GMS viruses. The scale bar in each map represents 1 antigenic unit (1 antigenic unit corresponds to a
twofold dilution of antiserum in the HI assay). The map was produced by ACMACS (https://acmacs-web.antigenic-cartography.org/).
Different GMS viruses (including site-associated combination viruses) are represented with different colors. WT: red, 54G
+
: pink, 87G
+
:
yellow, 127G
+
: blue, 148G
+
: purple, 178G
+
: orange, 188G
+
: green, 200G
+
: gray, 265G
+
: forest green, 267G
+
: brown, 280G
−
: dark blue, and
295G
−
: dark brown. (c) Detailed antigenic map corresponding to the black-dotted bordered rectangle in Figure 2(b).
8 Transboundary and Emerging Diseases

Infection
Day 0
Euthanasia Day 14
Monitor and
record the body
weight changes
and fatality
Lungs
collected Day 3
ðaÞ
01234567
WT
WT-54G+
WT-54+/188+
WT-87G+
WT-87G+/188G+/295G–
WT-87G+/200G+
WT-87G+/188G+/200G+/295G–
WT-127G+/295G–
WT-127G+
WT-127G+/200G+/295G–
WT-127G+/200G+
WT-127G+/178G+
WT-148G+
WT-148G+/200G+/295G–
WT-148G+/200G+
WT-178G+
WT-178G+/200G+/295G–
WT-178G+/200G+
WT-178G+/280G+
WT-188G+
WT-188G+/200G+/295G–
WT-188G+/200G+
WT-200G+
WT-265G+/200G+/295G–
WT-265G+/200G+
WT-267G+/200G+/295G–
WT-265G+
WT-267G+
WT-280G–
WT-295G–
Viral titers of mice lungs (lgPFU/g)Virus At the time of death
minimum body weigh
t
change (%)
–17.84
–28.79
–26.13
–21.61
–11.56
–22.49
–12.37
–26.32
–25.13
–25.14
–20.83
–27.18
No death∗
No death∗
No death∗
No death∗
No death∗
No death∗
No death∗
No death∗
No death∗
No death∗
No death∗
No death∗
No death∗
No death∗
No death∗
No death∗
No death∗
No death∗
∗∗
∗∗
∗∗
∗
∗∗
∗
ðbÞ
Virulence
GMS that could improve the pathogenicity
of the virus to mice
295G– 280G– 87G+ 265G+ 267G+
ðcÞ
Virulence
GMS that could attenuate the pathogenicity
of the virus to mice
127G+ 188G+ 148G+ 178G+ 54G+
ðdÞ
FIGURE 3: Pathogenicity of all GMS viruses in mice. (a) Mouse infection cartoon pattern diagram. Six-week-old SPF BALB/c mice were
inoculated intranasally with 4 ×10
5
PFU of all GMS viruses, changes in mouse body weight were continuously monitored until 14 days after
challenge, and lungs were collected on day 3 postinfection for virus titration in MDCK cells. The plants were created with https://BioRender.
com. (b) Viral titers in mouse lungs on day 3 postinfection with all GMS viruses and the minimum body weight change at the time of death.
WT is represented in light pink. GMS viruses with stronger pathogenicity than WT viruses are represented in light blue. GMS viruses with
weaker pathogenicity than WT viruses are represented in light yellow. Classification of GMS strains with different degrees of pathogenicity in
mice. No death ∗indicates that no mice died in this group. The values represent the means ÆSDs from three independent experiments
(∗P<0:05, ∗∗P<0:01). (c) GMS could further enhance pathogenicity in mice (represented in light blue overall). The plants were created with https://
BioRender.com. (d) GMS attenuated pathogenicity in mice (represented in light yellow overall). Source: Created with https://BioRender.com.
Transboundary and Emerging Diseases 9

the MLD
50
values of these viruses were determined. According
to the MLD
50
values in Figure 5(a), the pathogenicity of the
viruses to mice decreased in the following order: 295G
−
>
87G
+
/295G
−
=280G
−
>87G
+
>WT =87G
+
/280G
−
/295G
−
>280G
−
/295G
−
. It has been determined that 280G
−
/295G
−
does not increase the pathogenicity of the virus but instead
reduces its pathogenicity. Moreover, compared with 280G
−
/
295G
−
, 87G
+
/280G
−
/295G
−
resulted in increased pathogenic-
ity. Compared with 87G
+
, 87G
+
/295G
−
had greater pathoge-
nicity, while the MLD
50
value of 87G
+
/295G
−
was equal to that
of 280G
−
.
As depicted in Figure 5(b), the body weight changes of
the mice inoculated with these viruses showed that at 10
4
PFU, the weight loss of the 295G
−
group was significantly
greater than that of the other groups. At 10
5
PFU, the weight
loss magnitude decreased in the following order: 295G
−
>
87G
+
/295G
−
>280G
−
>87G
+
>87G
+
/280G
−
/295G
−
>WT
>280G
−
/295G
−
.At10
6
PFU, the weight changes of the
87G
+
/295G
−
, 295G
−
, 87G
+
, and 280G
−
groups were greater
than those of the WT group, and the weight changes of the
four groups could not be distinguished well, followed by
those of the 87G
+
/280G
−
/295G
−
and 280G
−
/295G
−
groups.
This experiment showed not only the effects of increased
pathogenicity in mice by 87G
+
, 280G
−
, and 295G
−
but also
the greater pathogenicity when 87G
+
and 295G
−
were com-
bined. Therefore, WT-87G
+
, WT-295G
−
, and WT-87G
+
/
295G
−
were chosen as the targets for further research.
3.7. 87G
+
Impaired, but 295G
−
Did Not Alter Human
Receptor-Binding Properties. Previous results showed that
87G
+
, 295G
−
, and 87G
+
/295G
−
significantly enhanced the
pathogenicity of the virus in mice, with 87 amino acids of the
HA near the RBS region. We wanted to determine whether
the increased pathogenicity of this virus is due to changes in
its receptor-binding properties. We tested the receptor-binding
properties of viruses with different GMSs. As shown in Figure 6,
A/swine/Jiangxi/261/2016 (H1N1) and A/chicken/Chongqing/
SD001/2021 (H5N6) were used as representative strains that
significantly bind to human-derived α-2,6 sialic acid receptor
and avian-derived α-2,3 sialic acid receptor, respectively. WT
andthreemutantvirusesbindstronglytotheα-2,6 sialic acid
receptor. The binding patterns of WT and WT-295G
−
to the α-
2,6 sialic acid receptor were identical, indicating that 295G
−
did
not affect human receptor-binding capabilities. Compared with
theWTstrains,theWT-87G
+
and WT-87G
+
/295G
−
strains
exhibited decreased affinity for the α-2,6 sialic acid receptor,
demonstrating that 87G
+
can alter the affinity of the virus for
this receptor. Although 87G
+
and 295G
−
dramatically
increased the virus’s pathogenicity in mice, this was not due
to a change in human receptor-binding capabilities. The affin-
ity of the virus for the α-2,6 sialic acid receptor was dramatically
reduced by 87G
+
but not by 295G
−
.
3.8. 87G
+
Decreased the Thermal and Acid Stability of the
Virus, yet 295G
−
Increased. When the HA protein is exposed
to a neutral pH but the temperature continues to increase, its
conformation shifts accordingly. HA protein stability refers
to the maintenance of biological activity under various con-
ditions, and it is a fundamental biological characteristic. A
decrease in the acid and thermal stability of a virus is not
beneficial for its survival in the external environment. How-
ever, after the virus invades the host cell, the decrease in acid
stability (i.e., the increase in cell membrane fusion pH) will
help the virus undergo membrane fusion as soon as possible
and release the virus genome. Therefore, different viruses
have optimal membrane fusion pH values to balance their
WT
70 KD
55 KD
WT-87G+
WT-280G–
WT-295G–
FIGURE 4: Western blotting to confirm N-linked glycosylation at the indicated sites on the mutated virus. WT, WT-87G
+
, WT-280G
−
, and
WT-295G
−
were used to infect MDCK cells at an MOI of 1. The infected cells were lysed to collect viral protein at 16 hr postinfection. Then,
the viral protein was separated by 10%Tris SDS–PAGE, and anti-H9N2 HA mouse monoclonal antibodies were used as antibodies.
TABLE 2: The basis of GMS combination viruses.
Serial
number
Detailed glycosylated site added or deleted
under WT background
Hemagglutinin
titer (log2)
Virus titer
(PFU/ml)
WT (11G
+
, 123G
+
, 280G
+
, 287G
+
, and 295G
+
)
a,b
11 8.5 ×10
8
R-30 WT-87G
+
/295G
−
10 6.05 ×10
8
R-31 WT-87G
+
/280G
−
/295G
−
9 1.88 ×10
8
R-32 WT-280G
−
/295G
−
10 3.8 ×10
8
a
Position numbers are according to H9 numbering.
b
The WT protein originally contains four N-glycosylated sites, 11G
+
, 123G
+
,280G
+
, 287G
+
, and 295G
+
.
10 Transboundary and Emerging Diseases

Days postinfection Days postinfection Days postinfection
0
0
20
40
60
80
100
0
20
40
60
80
100
0
20
40
60
80
100
0
20
40
60
80
100
0
20
40
60
80
100
0
20
40
60
80
100
0
20
40
60
80
100
123456
MLD50: 3.16 × 105 PFU/ml MLD50: 6.92 × 104 PFU/ml MLD50: 3.16 × 104 PFU/ml
MLD50: 3.16 × 104 PFU/ml MLD50: 3.16 × 105 PFU/ml
WT
Survival (%)Survival (%)Survival (%)
WT-87G+WT-280G–
MLD50:1.48 × 104 PFU/ml
MLD50 > 106 PFU/ml
104 PFU
105 PFU
106 PFU
7 8 9 101112131415 0123456789101112131415 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
WT-295G–
Days postinfection
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Days postinfection
0123456789101112131415
Days postinfection
0123456789101112131415
Days postinfection
0123456789101112131415
WT-280G–/295G–
WT-87G+/295G–WT-87G+/280G–/295G–
ðaÞ
Body weight change (%)
Body weight change (%)
Body weight change (%)
104 PFU 105 PFU 106 PFU
WT-280G–
Days postinfection
0
60
70
80
90
100
110
120
60
70
80
90
100
110
120
60
70
80
90
100
110
120
1234567891011121314
Days postinfection
012345678910111213
14
Days postinfection
012345678910111213
14
WT-280G–/295G–
WT-87G+
WT
WT-295G–
WT-87G+/295G–
WT-87G+/280G–/295G–
ðbÞ
FIGURE 5: The effect of recombinant GMS viruses on the pathogenicity of mice. (a) Survival rate of mice infected with different GMS viruses;
mean weight of mice infected with different GMS viruses (10
4
PFU/50 μl, 10
5
PFU/50 μl, or 10
6
PFU/50 μl) (n=3). (b) The body weight
change associated with every dose of the GMS virus. Mice were humanely euthanized when they lost ≥25%of their initial body weight.
Transboundary and Emerging Diseases 11

WT
Sialylglycopolymer (ng)
0.0
0.5
1.0
1.5
2.0
100502512.56.33.21.6
0.8
0.4
0.20.1
Absorbance (490 nm)
α2,3-Sialylglycopolymer
α2,6-Sialylglycopolymer
ðaÞ
WT-87G
+
Sialylglycopolymer (ng)
0.0
0.5
1.0
1.5
2.0
α2,3-Sialylglycopolymer
α2,6-Sialylglycopolymer
100502512.56.33.21.60.80.40.20.1
Absorbance (490 nm)
ðbÞ
WT-295G
–
Sialylglycopolymer (ng)
0.0
0.5
1.0
1.5
2.0
100502512.56.33.21.60.80.40.20.1
Absorbance (490 nm)
α2,3-Sialylglycopolymer
α2,6-Sialylglycopolymer
ðcÞ
WT-87G
+
/295G
–
Sialylglycopolymer (ng)
0.0
0.5
1.0
1.5
2.0
100502512.56.33.21.60.80.40.20.1
Absorbance (490 nm)
α2,3-Sialylglycopolymer
α2,6-Sialylglycopolymer
ðdÞ
Jiangxi/261
Sialylglycopolymer (ng)
0.0
0.5
1.0
1.5
2.0
100502512.56.33.21.60.80.40.20.1
Absorbance (490 nm)
α2,3-Sialylglycopolymer
α2,6-Sialylglycopolymer
ðeÞ
Chongqing/SD
Sialylglycopolymer (ng)
0.0
0.5
1.0
1.5
2.0
100502512.56.33.21.60.80.40.20.1
Absorbance (490 nm)
α2,3-Sialylglycopolymer
α2,6-Sialylglycopolymer
ðfÞ
FIGURE 6: (a–f) Receptor-binding properties of different GMS viruses. The binding ability of different viruses to two different biotinylated
glycans (α-2,3 glycan, blue; α-2,6 glycan, purple) was assessed. The data shown are the means of three repeats. The error bars indicate
standard deviations.
12 Transboundary and Emerging Diseases

activity in acidic extracellular environments and their acid
sensitivity to membrane fusion in acidic endosomes [40].
To investigate whether glycosylation affects thermal sta-
bility, the decreases in the HA titer of different GMS viruses
were analyzed. As shown in Figure 7(a), 87G
+
had the poorest
thermal stability, followed by 295G
−
and 87G
+
/295G
−
, and
WT had the greatest thermal stability. To investigate whether
the recombinant virus affects acid stability, the hemagglutina-
tion titer was then measured, as shownin Figure 7(b), at pH 4,
6, and 7. There was no significant difference in the hemagglu-
tination values. However, at pH 5, the HA titers of 87G
+
and
87G
+
/295G
−
decreased significantly. At a pH of 6, the hem-
agglutination titer of the WT-295G
−
strain was slightly
greater than that of the WT strain. The introduction of
87G
+
or 87G
+
/295G
−
to the HA of the WT strain decreased
the acid stability of the virus. In the absence of glycosylation
at position 295, the acid stability of the virus slightly
increased. Therefore, it was speculated that the decrease
Minutes post in 56°C
HA LOSE (log2)
0
–6
–4
–2
0
5 1015306090120150
WT-87G+WT-87G+/295G–
WT WT-295G–
ðaÞ
∗∗ ∗∗
Hemagglutination (log2)
0
2
4
8
6
7.0 6.0 5.0
pH value
4.0
WT-87G+WT-87G+/295G–
WT WT-295G–
ðbÞ
WT-87G+
WT-87G+/295G+
WT
pH 5.0 pH 5.1 pH 5.2 pH 5.3 pH 5.4 pH 5.5
WT-295G+
ðcÞ
FIGURE 7: Effect of different GMS viruses on stability. (a) Effect of different GMS viruses on thermostability. The thermal stability was assessed
by examining the ability of each virus to hemagglutinate chicken erythrocytes after incubation at 56°C over a time course. (b) Effect of
different GMS viruses on pH stability. Seven log2 hemagglutination units (HAUs) of viruses were incubated in different buffers at 37°C for
10 min, and the viral titers were determined by the HA assay. The results are presented as log
2
HA titers at the indicated pH values. The data
are presented as the means ÆSDs of results from three independent experiments. Statistical significance was determined by two-way
ANOVA ( ∗∗ P<0:01). (c) Effect of different GMS viruses on membrane fusion ability. Syncytium formation in Vero cells infected with
different GMS viruses at different pH values. The cells were fixed and stained with Giemsa solution. The red lines represent the range of pH
values at which fusion was detected microscopically.
Transboundary and Emerging Diseases 13

in acid stability of 87G
+
/295G
−
was largely caused by 87G
+
rather than 295G
−
.
Membrane fusion occurs after the virus infects host cells
due to a conformational shift in the HA protein produced by
the acidity of the capsule membrane. When the syncytium
formed, multiple Vero cells underwent membrane fusion
and fusion, resulting in nuclear aggregation and fusion,
and large cell clusters with blue Giemsa staining could be
observed. If a syncytium cannot form, Vero cells cannot
undergo membrane fusion, thus maintaining their individual
dispersed cell state. After staining, the dispersed single nuclei
were evenly distributed [41, 42]. To determine the effect of
the four recombinant viruses on cell fusion, syncytial forma-
tion experiments were conducted. The results are shown in
Figure 7(c). At pH =5.3, syncytium formation in the WT,
87G
+
, and 295G
−
strains was observed; however, the degree
to which the syncytium formed in the WT and 87G
+
strains
was greater than that in the 295G
−
strain, while the pH at
which the syncytium formed in the 87G
+
/295G
−
strain was
5.4, which increased the pH of cell membrane fusion. The
lower the pH of membrane fusion, the better the acid stabil-
ity. Therefore, the acid stability of the WT-295G
−
mutant
was slightly better than that of WT, while the acid stability of
the WT-87G
+
/295G
−
mutant was worse than that of WT.
Overall, 295G
−
slightly improved the effect on viral acid
stability. Moreover, 87G
+
significantly weakened the stability
of the virus.
3.9. 87G
+
, 295G
−
, and 87G
+
/295G
−
Could Further Enhance
the Inflammatory Response of Infected Mouse Lungs. To ana-
lyze the virus replication efficiency in vitro, MDCK and A549
cells were infected with the modified virus, and the cell
supernatants were collected for titration. The results are
shown in Figure 8(a). The replication titers of the three
modified viruses were greater in the MDCK cells 12 hr after
infection than in the WT cells. However, except for the WT-
87G
+
strain, whose replication titer was lower than that of
the WT strain, there was no difference in replication titer
between the WT-295G
−
, WT-87G
+
/295G
−
, and WT strains.
In the A549 cells, the replication titers of the three mutated
viruses were lower than those of the WT strain at any time
point. The results indicated that the replication titers of the
four mutated viruses in MDCK cells were indistinguishable,
while those in A549 cells were low.
To detect the replication titers of the virus in mice and
the expression levels of cytokines in the lungs, on the third
day after inoculation, the mice were killed, and their lungs
were collected for viral titration of MDCK cells. Figure 8(b)
shows that the replication titers of the three mutated viruses
were much greater than those of the WT virus, demonstrat-
ing that the mutated viruses increased replication in the
mouse lungs. Next, we examined cytokine levels in the lungs
of the mice. IL-1β, IL-6, and TNF-αare proinflammatory
cytokines, while IFN-βis an anti-inflammatory cytokine.
The expression levels of cytokines in the lungs were detected
by RT-qPCR, as shown in Figure 8(c). The three mutant
viruses showed considerably greater expression levels of IL-
1β, IL-6, and TNF-αthan did the wildtype. The expression of
IFN-βin the WT-87G
+
strain was similar to that in the WT
strain, yet the expression of IFN-βin the other two viruses
was significantly decreased. Therefore, infecting mice with
three different mutated viruses, 87G
+
, 295G
−
and 87G
+
/
295G
−
, could result in a greater inflammatory response
than that in WT mice, thus leading to more extensive lung
damage.
4. Discussion
The glycosylation of the influenza virus HA protein is
thought to be a method of masking or covering antibody-
binding sites, allowing immunological escape. As a result,
certain GMSs in the head of the HA protein are bound to
impact the antigenicity of the virus [29, 43, 44]. Analysis of
the antigen map revealed that the combination of 200G
+
/
295G
−
could increase the antigen distance on the antigen
map. It has also been reported that immunization with a
200G
+
/295G
−
combined vaccine in chickens can provide
100%protection against 200G
+
/295G
+
and 200G
−
/295G
+
virus reinfection [45]. However, the 200G
+
/295G
−
epidemic
proportion has decreased in the last few years. When two or
more GMSs were combined, such as 54G
+
/188G
+
and 87G
+
/
188G
+
, the virus was able to cause greater antigenic diver-
gence simultaneously because of the masking effect of glyco-
sylation on the HA protein, making it more difficult for
antibodies to neutralize. It has been documented that anti-
bodies derived from vaccines prepared with viruses modified
with more glycosylated sites have a wider range of neutraliz-
ing activity than antibodies derived from viruses modified
with fewer glycosylated sites [30], suggesting that viruses
with more glycosylations of HA as vaccine candidates can
provide cross-protective advantages for different viral strains
[29]. The HA protein of seasonal H1N1 subtype influenza
viruses acquired two additional glycosylation sites at posi-
tions 129 and 163 (H9: positions 127 and 156), providing
viruses with neutralizing activity against antisera produced
by any of the wild-type viruses [27]. The introduction of new
oligosaccharides at positions 63, 122, 126, 133, and 246 (H9:
positions 56, 115, 121, 127, and 236) of the HA protein of the
H3N2 subtype influenza virus leads to immune escape by
altering its antigenicity [32]. Antigenicity research revealed
that 127G
+
and its related viruses could generate greater
antigenic divergence. In addition, 127G
+
and 178G
+
(as
well as their related viruses) caused the greatest distance on
the map. It has been shown that H9N2 AIV demonstrated
the most prominent antigen escape due to 127 and 183
amino acid changes in the HA gene, resulting in an oligosac-
charide connected at residue 127 [22]. During the evolution
of human seasonal influenza viruses, the number of GMS on
the HA protein has been increasing annually [27], showing a
selective advantage, and an increase in GMS can effectively
prevent the binding of neutralizing antibodies to antigenic
epitopes [46]. As a result, the acquisition of potential GMS is
a successful approach for influenza viruses to avoid host
immune pressure [35].
In this study, the sites with enhanced pathogenicity in
mice were mainly located at the edge of HA, for example,
14 Transboundary and Emerging Diseases

295G
−
, 280G
−
, 87G
+
, 265G
+
, and 267G
+
. The areas of
decreased pathogenicity in the H1N1 subtype influenza virus
were also localized at the margin of the HA head at positions
71G
+
and 286G
+
(H9: positions 71 and 280) [21]. We also
found that viruses with 127G
+
, 188G
+
, 148G
+
, 178G
+
,or
54G
+
could attenuate pathogenicity in mice. The addition
of GMS at position 127 of the HA protein of H9N2 AIV
reportedly diminishes the pathogenicity of the virus in
mice, which is consistent with our research results [22].
Although 87G
+
, 295G
−
, and 87G
+
/295G
−
could increase
pathogenicity in mice, the replication titers of the three
recombinant mutated viruses were all greater than those of
WT in the first 12 hr in MDCK; however, after 12 hr, the
replication titers of the four viruses were nearly indistin-
guishable. In A549 cells, the WT replication titer was always
greater than that of the three recombinant mutant viruses. It
∗∗
∗
∗∗∗
WT-87G+WT-87G+/295G–
WT WT-295G–
Hours postinfection
0 1224364860
Hours postinfection
Virus titer lgPFU/ml
Virus titer lgPFU/ml
0
4
5
6
7
8
9
1
2
3
4
12 24
MDCK A549
36 48 60
ðaÞ
∗∗∗ ∗∗∗∗
WT-87G+
WT-87G+
WT-87G+/295G–
WT-87G+/295G–
WT
WT
WT-295G–
WT-295G–
lgPFU/g
4
3
5
6
ðbÞ
∗∗∗∗
∗∗∗
∗∗∗∗
∗∗∗
∗∗∗
∗∗∗∗
∗∗∗
∗∗∗
∗
∗∗∗∗
∗∗∗
WT-87G+
WT-87G+/295G–
WT
WT-295G–
WT-87G+
WT-87G+/295G–
WT
WT-295G–
WT-87G+
WT-87G+/295G–
WT
WT-295G–
WT-87G+
WT-87G+/295G–
WT
Fold change
WT-295G–
WT-87G+WT-87G+/295G–
WT WT-295G–
0
2
4
6
0
2
4
6
8
0
50
100
150
0.0
0.5
1.0
1.5
20010
IL-1βTNF-αIFN-βIL-6
ðcÞ
FIGURE 8: The viral replication of different GMS viruses in vitro and in vivo. (a) Growth kinetics of different GMS viruses in MDCK and A549
cells. Growth curves of different cell lines infected with WT, WT-87G
+
, WT-280G
−
, or WT-295G were generated at multiplicities of infection
(MOIs) of 0.01 and 0.5. Then, cell supernatants were collected at 12, 24, 36, and 48 hr postinfection, and viral titers were determined via
plaque assays. The values represent the means ÆSDs from three independent experiments ( ∗P<0:05, ∗∗ P<0:01, and ∗∗∗ P<0:001). (b) Viral
titers in mouse lungs on day 3 postinfection. Six-week-old SPF BALB/c mice were inoculated intranasally with 1 ×10
5
PFU of WT, WT-
87G
+
, WT-280G
−
, or WT-295G
−
, and lungs were collected on day 3 postinfection for virus titration in MDCK cells by plaque assay
(∗P<0:05, ∗∗P<0:01, ∗∗∗ P<0:001, and ∗∗∗∗P<0:0001). (c) Cytokine production in the lungs of mice. Lungs were collected on day 3
postinfection. Relative mRNA levels of IL-1β, IL-6, TNF-α, and IFN-βin infected mouse lungs were measured by RT-qPCR. All values were
normalized to GAPDH and are expressed as the fold change compared with controls. The values represent the means ÆSDs from three
independent experiments ( ∗P<0:05, ∗∗ P<0:01, ∗∗∗P<0:001, and ∗∗∗∗ P<0:0001).
Transboundary and Emerging Diseases 15

has been reported that when GMS at positions 10, 23, and
286 (H9: positions 10, 23, and 280) on HA of the H5N1
subtype influenza virus is removed, the cleavage of HA is
almost completely blocked, leading to a significant decrease
in the growth rates of the mutant viruses in MDCK and CEF
cells [47]. In comparison to the wild-type virus of the H5N1
subtype influenza virus, viruses that lack glycosylation at
either position 158 or 169 (H9: positions 152 and 163) had
a significantly lower growth titer in both MDCK and A549
cells but displayed increased pathogenicity in mice [21]. The
pathogenicity of the virus is sometimes not positively corre-
lated with its replication efficiency in specific cells.
IL-1β, IL-6, and TNF-αare proinflammatory factors. In
this study, infection of mice with WT-87G
+
, WT-295G
−
,or
WT-87G
+
/295G
−
significantly increased the expression
levels of proinflammatory factors in their lungs, resulting
in severe inflammatory reactions. Severe inflammatory reac-
tions in mouse lungs are often associated with high pathoge-
nicity of the virus. It was reported that the removal of
glycosylation at positions 158 or 169 (H9: positions 152 and
163) of the H5N1 subtype of avian influenza virus could
increase the pathogenicity of the virus to mice, and the expres-
sion levels of cytokines such as IL-6, IL-8,and MX-1 in mouse
lungs are significantly increased [48]. The addition of glyco-
sylation at position 158 (H9: position 152) of the H5N6 sub-
type of avian influenza virus can enhance the pathogenicity of
the virus to mice, resulting in increased levels of inflammatory
factors in mouse lungs (e.g., HMGB1, IL-10, and TNF-α) [49].
Mice infected with WT-295G or WT-87G
+
/295G showed
decreased expression of the antiviral cytokine IFN-βin their
lungs. It has been reported that the expression level of IFN-β
detected in some elderly individuals infected with the influ-
enza virus is significantly reduced compared to that in young
people to prevent the immune system from detecting the
virus’s genes, allowing the virus to replicate unchecked and
ultimately causing severe inflammatory reactions [50].
SP-D is a mouse lung surface protein belonging to one of the
lectins in the collectin family [51]. It tends to bind oligosacchar-
ides that end in mannose [52]. SP-D binds with the high man-
nose−oligosaccharide chains of the HA and NA proteins of the
influenza virus, thus inhibiting their activity, resulting in viral
aggregation and decreased infectivity [53, 54]. Several studies
have shown that the absence of glycosylation at sites of the
influenza virus HA protein, such as residue 158 (H9: position
152) of the H5N1 subtype, can cause increased pathogenicity in
mice [21]. The greater pathogenicity observed due to the deletion
of GMS is attributed to the reduced sensitivity to SP-D. A/Hong
Kong 1/68 (H3N2) HA was changed with several GMSs, result-
ing in greater susceptibility to SP-D and lower pathogenicity in
mice [34]. The removal of glycosylated HA1 from the H3 sub-
type (A/Beijing/353/89) and H1 subtype (A/Brazil/11/78)
reduced the neutralization of the virus by SP-D, resulting in
enhanced pathogenicity in mice [55, 56]. Our findings imply
that 295G
−
and 280G
−
can increase viral pathogenicity in
mice.Therefore,webelievethatthe greater pathogenicity in
mice is likely due to the loss of GMS, which causes decreased
sensitivity to SP-D; however, this hypothesis was not examined
in detail in this research and requires further evidence.
5. Conclusions
We generated viruses modified with various GMSs to explore
the effects of GMS on antigenicity and pathogenicity. Anti-
genic analysis revealed that viruses with 200G
+
/295G
−
com-
bined with single GMS, such as 87G
+
, 127G
+
, 148G
+
, 178G
+
,
and 265G
+
, could significantly affect the antigenicity of the
virus. Pathogenicity analysis revealed that the addition of
GMS, such as 127G
+
, 188G
+
, 148G
+
, 178G
+
, or 54G
+
,
decreased the virulence of the virus, except for 87G
+
. The
removal of GMS, such as 280G
−
or 295G
−
, increased the
virulence of the virus in mice. Further pathogenicity studies
revealed that 87G
+
/295G
−
could also enhance the pathoge-
nicity of the virus, while 280G
−
/295G
−
weakened it. This
study can guide epidemiologic surveillance results, provide
early warning of cross-host adaptation of H9N2 AIV, and
reveal the potential threat to public health security.
Data Availability
The data supporting the study’sfindings are available from
the corresponding author upon request.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
This work was supported by the Major Science and Technology
Projects in Xinjiang Autonomous Region—Xinjiang Livestock
and Poultry Disease Prevention and Control System Quality
Improvement Project (2023A02007), and National Natural Sci-
ence Foundation of China (32272992, 31772775), Fundamental
Research Funds for the Central Universities (YDZX2023005).
We thank Dr. Yoshihiro Kawaoka (University of Wisconsin-
Madison) for the gifts of the pHH21 and pCAGGS vectors.
Supplementary Materials
Table S1: detailed information about the proportions of gly-
cosylations of different GMSs globally. Table S2: detailed
information about the proportions of domestic glycosyla-
tions of different GMSs. Table S3: detailed primers sequences
information about specific GMS. Table S4: primers of quan-
titative real-time RT-PCR assay. Figure S1: body weight
changes, viral titers in the lungs, and fatality rates of each
group of GMS viruses. (Supplementary Materials)
References
[1] J. Shi, Z. Wen, G. Zhong et al., “Susceptibility of ferrets, cats,
dogs, and other domesticated animals to SARS–coronavirus
2,”Science, vol. 368, no. 6494, pp. 1016–1020, 2020.
[2] J. Guo, Y. Wang, C. Zhao et al., “Molecular characterization,
receptor binding property, and replication in chickens and
mice of H9N2 avian influenza viruses isolated from chickens,
peafowls, and wild birds in eastern China,”Emerging Microbes
&Infections, vol. 10, no. 1, pp. 2098–2112, 2021.
[3] B. Wang, Y. Huang, B. Hu et al., “Characterization of a
reassortant H11N9 subtype avian influenza virus isolated from
16 Transboundary and Emerging Diseases

spot-billed duck in China,”Virus Genes, vol. 59, pp. 604–612,
2023.
[4] T. M. Uyeki, D. C. Nguyen, T. Rowe et al., “Seroprevalence of
antibodies to avian influenza A (H5) and A (H9) viruses
among market poultry workers, Hanoi, Vietnam, 2001,”PLoS
ONE, vol. 7, no. 8, Article ID e43948, 2012.
[5] Y. Ge, Q. Yao, X. Wang et al., “Detection of reassortant avian
influenza A (H11N9) virus in wild birds in China,”
Transboundary and Emerging Diseases, vol. 66, no. 3,
pp. 1142–1157, 2019.
[6] Y. Sun and J. Liu, “H9N2 influenza virus in China: a cause of
concern,”Protein &Cell, vol. 6, no. 1, pp. 18–25, 2015.
[7] X. Li, J. Shi, J. Guo et al., “Correction: genetics, receptor
binding property, and transmissibility in mammals of
naturally isolated H9N2 avian influenza vIruses,”PLoS
Pathogens, vol. 16, no. 1, Article ID e1008284, 2020.
[8] Q. Yu, L. Liu, J. Pu et al., “Risk perceptions for avian influenza
virus infection among poultry workers, China,”Emerging
Infectious Diseases, vol. 19, no. 2, pp. 313–316, 2013.
[9] P. J. Blair, S. D. Putnam, W. S. Krueger et al., “Evidence for
avian H9N2 influenza virus infections among rural villagers in
Cambodia,”Journal of Infection and Public Health, vol. 6,
no. 2, pp. 69–79, 2013.
[10] A. Coman, D. N. Maftei, W. S. Krueger et al., “Serological
evidence for avian H9N2 influenza virus infections among
Romanian agriculture workers,”Journal of Infection and
Public Health, vol. 6, no. 6, pp. 438–447, 2013.
[11] J. Okoye, D. Eze, W. S. Krueger, G. L. Heil, J. A. Friary, and
G. C. Gray, “Serologic evidence of avian influenza virus
infections among Nigerian agricultural workers,”Journal of
Medical Virology,vol. 85, no. 4, pp. 670–676, 2013.
[12] J. Pu, S. Wang, Y. Yin et al., “Evolution of the H9N2 influenza
genotype that facilitated the genesis of the novel H7N9 virus,”
Proceedings of the National Academy of Sciences of the United
States of America, vol. 112, no. 2, pp. 548–553, 2015.
[13] Y. Zhang, J. Shi, P. Cui et al., “Genetic analysis and biological
characterization of H10N3 influenza A viruses isolated in
China from 2014 to 2021,”Journal of Medical Virology,
vol. 95, no. 2, Article ID e28476, 2023.
[14] P. Cui, J. Shi, C. Yan et al., “Analysis of avian influenza A
(H3N8) viruses in poultry and their zoonotic potential, China,
September 2021 to May 2022,”Eurosurveillance, vol. 28,
no. 41, Article ID 2200871, 2023.
[15] Y. Zhang, S.-M. Zou, X.-D. Li et al., “Detection of reassortant
avian influenza A (H11N9) virus in environmental samples
from live poultry markets in China,”Infectious Diseases of
Poverty, vol. 5, Article ID 59, 2016.
[16] P. Zhang, Y. Tang, X. Liu et al., “Characterization of H9N2
influenza viruses isolated from vaccinated flocks in an
integrated broiler chicken operation in eastern China during a
5 year period (1998–2002),”Journal of General Virology,
vol. 89, no. 12, pp. 3102–3112, 2008.
[17] K. J. Park, H.-I Kwon, M.-S. Song et al., “Rapid evolution of
low-pathogenic H9N2 avian influenza viruses following
poultry vaccination programmes,”Journal of General Virology,
vol. 92, no. 1, pp. 36–50, 2010.
[18] Y. Bi, J. Li, S. Li et al., “Dominant subtype switch in avian
influenza viruses during 2016–2019 in China,”Nature
Communications, vol. 11, Article ID 5909, 2020.
[19] N. C. Wu and I. A. Wilson, “Influenza hemagglutinin
structures and antibody recognition,”Cold Spring Harbor
Perspectives in Medicine, vol. 10, no. 8, Article ID a038778,
2020.
[20] X. Yin, G. Deng, X. Zeng et al., “Genetic and biological
properties of H7N9 avian influenza viruses detected after
application of the H7N9 poultry vaccine in China,”PLoS
Pathogens, vol. 17, no. 4, Article ID e1009561, 2021.
[21] X. Zhang, S. Chen, Y. Jiang et al., “Hemagglutinin
glycosylation modulates the pathogenicity and antigenicity
of the H5N1 avian influenza virus,”Veterinary Microbiology,
vol. 175, no. 2–4, pp. 244–256, 2015.
[22] M. Fan, B. Liang, Y. Zhao et al., “Mutations of 127, 183 and
212 residues on the HA globular head affect the antigenicity,
replication and pathogenicity of H9N2 avian influenza virus,”
Transboundary and Emerging Diseases, vol. 69, no. 4,
pp. e659–e670, 2022.
[23] N. Chen, B. Zhang, L. Deng, B. Liang, and J. Ping, “Virus-host
interaction networks as new antiviral drug targets for IAV and
SARS-CoV-2,”Emerging Microbes &Infections, vol. 11, no. 1,
pp. 1371–1389, 2022.
[24] J. Shi, X. Zeng, P. Cui, C. Yan, and H. Chen, “Alarming
situation of emerging H5 and H7 avian influenza and effective
control strategies,”Emerging Microbes &Infections, vol. 12,
no. 1, Article ID 2155072, 2023.
[25] D. J. Vigerust and V. L. Shepherd, “Virus glycosylation: role in
virulence and immune interactions,”Trends in Microbiology,
vol. 15, no. 5, pp. 211–218, 2007.
[26] M. D. Tate, E. R. Job, Y.-M. Deng, V. Gunalan, S. Maurer-
Stroh, and P. Reading, “Playing hide and seek: how
glycosylation of the influenza virus hemagglutinin can
modulate the immune response to infection,”Viruses, vol. 6,
no. 3, pp. 1294–1316, 2014.
[27] S. Sun, Q. Wang, F. Zhao, W. Chen, and Z. Li, “Glycosylation
site alteration in the evolution of influenza A (H1N1) viruses,”
PLoS ONE, vol. 6, no. 7, Article ID e22844, 2011.
[28] S. Sun, Q. Wang, F. Zhao, W. Chen, and Z. Li, “Prediction of
biological functions on glycosylation site migrations in human
influenza H1N1 viruses,”PLoS ONE, vol. 7, no. 2, Article ID
e32119, 2012.
[29] C.-J. Wei, J. C. Boyington, K. Dai et al., “Cross-neutralization
of 1918 and 2009 influenza viruses: role of glycans in viral
evolution and vaccine design,”Science Translational Medicine,
vol. 2, no. 24, pp. 24ra21–24ra21, 2010.
[30] X. Sun, A. Jayaraman, P. Maniprasad et al., “N-linked
glycosylation of the hemagglutinin protein influences
virulence and antigenicity of the 1918 pandemic and seasonal
H1N1 influenza A viruses,”Journal of Virology, vol. 87,
no. 15, pp. 8756–8766, 2013.
[31] S. J. Zost, K. Parkhouse, M. E. Gumina et al., “Contemporary
H3N2 influenza viruses have a glycosylation site that alters
binding of antibodies elicited by egg-adapted vaccine strains,”
Proceedings of the National Academy of Sciences of the United
States of America, vol. 114, no. 47, pp. 12578–12583, 2017.
[32] Y. Abe, E. Takashita, K. Sugawara, Y. Matsuzaki, Y. Muraki,
and S. Hongo, “Effect of the addition of oligosaccharides on
the biological activities and antigenicity of influenza A/H3N2
virus hemagglutinin,”Journal of Virology, vol. 78, no. 18,
pp. 9605–9611, 2004.
[33] D. Zhao, L. Liang, S. Wang et al., “Glycosylation of the
hemagglutinin protein of H5N1 influenza virus increases its
virulence in mice by exacerbating the host immune response,”
Journal of Virology, vol. 91, no. 7, 2017.
[34] D. J. Vigerust, K. B. Ulett, K. L. Boyd, J. Madsen, S. Hawgood,
and J. A. McCullers, “N-linked glycosylation attenuates H3N2
influenza viruses,”Journal of Virology, vol. 81, no. 16,
pp. 8593–8600, 2007.
Transboundary and Emerging Diseases 17

[35] J. L. Cherry, D. J. Lipman, A. Nikolskaya, and Y. I. Wolf,
“Evolutionary dynamics of N-glycosylation sites of influenza
virus hemagglutinin,”PLoS Currents: Disasters, vol. 1,
Article ID RRN1001, 2009.
[36] G. Neumann, T. Watanabe, H. Ito et al., “Generation of
influenza A viruses entirely from cloned cDNAs,”Proceedings
of the National Academy of Sciences of the United States of
America, vol. 96, no. 16, pp. 9345–9350, 1999.
[37] X. Li, J. Shi, J. Guo et al., “Genetics, receptor binding property,
and transmissibility in mammals of naturally isolated H9N2
Avian Influenza viruses,”PLoS Pathogens, vol. 10, no. 11,
Article ID e1004508, 2014.
[38] G. Wang, G. Deng, J. Shi et al., “H6 influenza viruses pose a
potential threat to human health,”Journal of Virology, vol. 88,
no. 8, pp. 3953–3964, 2014.
[39] D. J. Smith, A. S. Lapedes, J. C. de Jong et al., “Mapping the
antigenic and genetic evolution of influenza virus,”Science,
vol. 305, no. 5682, pp. 371–376, 2004.
[40] C. M. Carr, C. Chaudhry, and P. S. Kim, “Influenza hemaggluti-
nin is spring-loaded by a metastable native conformation,”
Proceedings of the National Academy of Sciences, vol. 94, no. 26,
pp. 14306–14313, 1997.
[41] T. P. Peacock, D. J. Benton, J. James et al., “Immune escape
variants of H9N2 influenza viruses containing deletions at the
hemagglutinin receptor binding site retain fitness in vivo and
display enhanced zoonotic characteristics,”Journal of
Virology, vol. 91, no. 14, 2017.
[42] T. Daidoji, Y. Watanabe, Y. Arai, J. Kajikawa, R. Hirose, and
T. Nakaya, “Unique infectious strategy of H5N1 avian influenza
virus is governed by the acid-destabilized property of
hemagglutinin,”Viral Immunology, vol. 30, no. 6, pp. 398–407,
2017.
[43] B. Manicassamy, R. A. Medina, R. Hai et al., “Protection of
mice against lethal challenge with 2009 H1N1 influenza A
virus by 1918-like and classical swine H1N1 based vaccines,”
PLoS Pathogens, vol. 6, no. 1, Article ID e1000745, 2010.
[44] J. J. Skehel and D. C. Wiley, “Receptor binding and membrane
fusion in virus entry: the influenza hemagglutinin,”Annual
Review of Biochemistry, vol. 69, pp. 531–569, 2000.
[45] Q. Peng, R. Zhu, X. Wang et al., “Impact of the variations in
potential glycosylation sites of the hemagglutinin of H9N2
influenza virus,”Virus Genes, vol. 55, pp. 182–190, 2019.
[46] M. Zhang, B. Gaschen, W. Blay et al., “Tracking global
patterns of N-linked glycosylation site variation in highly
variable viral glycoproteins: HIV, SIV, and HCV envelopes
and influenza hemagglutinin,”Glycobiology, vol. 14, no. 12,
pp. 1229–1246, 2004.
[47] X. Zhang, S. Chen, D. Yang et al., “Role of stem glycans
attached to haemagglutinin in the biological characteristics of
H5N1 avian influenza virus,”Journal of General Virology,
vol. 96, no. 6, pp. 1248–1257, 2015.
[48] R. Gao, M. Gu, L. Shi et al., “N-linked glycosylation at site 158
of the HA protein of H5N6 highly pathogenic avian influenza
virus is important for viral biological properties and host
immune responses,”Veterinary Research, vol. 52, Article ID 8,
2021.
[49] P. S. Pillai, R. D. Molony, K. Martinod et al., “Mx1 reveals
innate pathways to antiviral resistance and lethal influenza
disease,”Science, vol. 352, no. 6284, pp. 463–466, 2016.
[50] P. C. Reading, M. D. Tate, D. L. Pickett, and A. G. Brooks,
“Glycosylation as a target for recognition of influenza viruses
by the innate immune system,”in Advances in Experimental
Medicine and Biology, vol. 598, pp. 279–292, Springer, New
York, NY, 2007.
[51] L. E. Carreto-Binaghi, E. M. Aliouat, and M. L. Taylor,
“Surfactant proteins, SP-A and SP-D, in respiratory fungal
infections: their role in the inflammatory response,”Respiratory
Research, vol. 17, Article ID 66, 2016.
[52] C. Casals, M. A. Campanero-Rhodes, B. García-Fojeda, and
D. Solís, “The role of collectins and galectins in lung innate
immune defense,”Frontiers in Immunology, vol. 9, Article ID
1998, 2018.
[53] K. L. Hartshorn, R. Webby, M. R. White et al., “Role of viral
hemagglutinin glycosylation in anti-influenza activities of
recombinant surfactant protein D,”Respiratory Research,
vol. 9, Article ID 65, 2008.
[54] L. M. Parsons, O. Zoueva, G. Grubbs et al., “Glycosylation of H4
influenza strains with pandemic potential and susceptibilities to
lung surfactant SP-D,”Frontiers In Molecular Biosciences,vol. 10,
Article ID 1207670, 2023.
[55] M. D. Tate, A. G. Brooks, and P. C. Reading, “Specific sites of
N-linked glycosylation on the hemagglutinin of H1N1 subtype
influenza A virus determine sensitivity to inhibitors of the
innate immune system and virulence in mice,”The Journal of
Immunology, vol. 187, no. 4, pp. 1884–1894, 2011.
[56] M. D. Tate, E. R. Job, A. G. Brooks, and P. C. Reading,
“Glycosylation of the hemagglutinin modulates the sensitivity
of H3N2 influenza viruses to innate proteins in airway
secretions and virulence in mice,”Virology, vol. 413, no. 1,
pp. 84–92, 2011.
18 Transboundary and Emerging Diseases