Virus isolation and transmission electron microscope
The oropharyngeal and cloacal samples from sick chickens were identified as positive for influenza virus by both hemagglutination assay (HA) and RT-PCR. Further characterization revealed the novel influenza virus as H3N3 AIV, as confirmed by Sanger sequencing of the HA and NA genes. Moreover, Transmission electron microscopy (TEM) analysis of the allantoic fluid samples from infected SPF chicken embryos revealed the presence of influenza virus-like particles, characterized by their typical vesicular and spiked morphology. (Fig. 1).
In this study, we comprehensively analyzed all available influenza virus strain information from the GenBank and GISAID EpiFlu™ database. Despite a relatively small total number (n = 67) of H3N3 influenza virus isolates recorded since its initial detection in 1972, the virus exhibited a broad host range, including swine, humans, seals, and avian species (Fig. 2A). Furthermore, avian species, particularly chickens, ducks, geese and other birds were the primary hosts for H3N3 viruses. Notably, ducks constituted the largest host group, representing over half (49.3%) of all the H3N3 virus hosts, followed by the environment isolates (13.4%). In contrast, the isolation rate from chickens was relatively low (9.0%) (Fig. 2B).
Genomic structure and comparative analysis
The complete genome of A/chicken/China/16/2023 (H3N3) was sequenced by next-generation sequencing (NGS). It comprised eight fragments encoding PB2 (2280 nucleotides, nt), PB1 (2274 nt), PA (2151 nt), HA (1707 nt), NP (1497 nt), NA (1410 nt), M (982 nt), and NS (876 nt). The genome organization was consistent with other members of the Influenzavirus A genus (Table 1). The amino acid sequences of the encoded proteins were similar in length and number to those of other AIVs. Additionally, the complete coding sequence (CDS) of this novel H3N3 AIV isolate was annotated (Table 1).
Table 1
Genomic features of A/chicken/China/16/2023(H3N3) strain.
Segment number
|
Gene
|
Abbreviation
|
Coding sequence (from-to)a
|
Total length/nt
|
Protein length/aab
|
1
|
Polymerase PB2
|
PB2
|
1-2280
|
2280
|
759
|
2
|
Polymerase PB1 and PB1-F2 protein
|
PB1
|
1-2274; PB1-F2: 95–367
|
2274
|
847
|
3
|
Polymerase PA and PA-X protein
|
PA
|
1-2151; PA-X: 1-570, 572–760
|
2151
|
968
|
4
|
Hemagglutinin
|
HA
|
1-1701
|
1707
|
566
|
5
|
Nucleocapsid protein
|
NP
|
1-1497
|
1497
|
498
|
6
|
Neuraminidase
|
NA
|
1-1410
|
1410
|
469
|
7
|
Matrix protein 2 and matrix protein 1
|
M
|
M1: 1-759; M2: 1–26, 715–982
|
982
|
97
|
8
|
Nuclear export protein and nonstructural protein 1
|
NS
|
NS1: 21–674; NS2: 21–50, 523–858
|
876
|
338
|
aIncluding stop codon; bWithout stop codon. |
Comparative nucleotide sequence analysis of each gene was performed using BLAST against the GenBank database. As shown in Table 2, comparative nucleotide sequence analysis revealed high identity between the H3N3 AIV genome segments and their closest influenza virus counterparts: 98.03% (PB2), 98.24% (PB1), 98.19% (PA), 98.77% (HA), 98.33% (NP), 98.72% (NA), 99.19% (M), and 98.40% (NS), respectively. Intriguingly, the H3N3 strain exhibited a complex reassortment pattern. Three segments (PB2, PB1, and NP) shared the highest nucleotide similarity with H9N2 subtype AIVs, while another three segments (HA, M, and NS) were most closely related to H3N8 subtype AIVs. The NA gene was most similar to H10N3, and the PA gene showed high homology to both H9N2 and H3N8 subtypes (Table 2). These findings strongly suggest that the H3N3 AIV is a novel triple reassortant derived from H9N2, H3N8, and H10N3 influenza viruses.
Table 2
Sequence identity of novel H3N3 AIV with closest homologous Influenza Viruses.
Gene
|
Viruses with greatest homology
|
Nucleotide identity (%)
|
GenBank
accession No.
|
PB2
|
A/chicken/Shanghai/11/2018(H9N2)
|
98.03
|
MK053869
|
A/chicken/China/71/2019(H9N2)
|
97.98
|
MN263225
|
PB1
|
A/environment-air/Kunshan/NIOSH-BL20/2018(H9N2)
|
98.24
|
MN606218
|
A/swine/Shandong/TA009/2019(H9N2)
|
98.01
|
MT265024
|
PA
|
A/chicken/China/2-3-1/2022(H9N2)
|
98.19
|
OR528343
|
A/China/ZMD-22-2/2022(H3N8)
|
98.19
|
ON342808
|
HA
|
A/China/ZMD-22-2/2022(H3N8)
|
98.77
|
ON342803
|
A/chicken/China/NXFWB5/2022(H3N8)
|
98.70
|
OP897716
|
NP
|
A/chicken/China/HP-1/2018(H9N2)
|
98.33
|
OR528592
|
A/chicken/Anhui/12.25_YHZGS004-O/2018(H9N2)
|
98.33
|
MW102497
|
NA
|
A/Environment/China/04940NA/2021(H10N3)
|
98.72
|
OM802522
|
A/chicken/China/189-6/2021(H10N3)
|
98.58
|
OL638128
|
M
|
A/chicken/Huizhou/104/2022(H3N8)
|
99.19
|
OQ291992
|
A/chicken/Dongguan/364/2022(H3N8)
|
99.19
|
OQ291664
|
NS
|
A/Duck/Jiangxi/447/2022(H3N8)
|
98.40
|
OQ826115
|
A/chicken/Shandong/3424/2016(H9N2)
|
97.94
|
MH667576
|
Notably, the HA and PA gene of the H3N3 isolate exhibited the highest nucleotide homology with A/China/ZMD-22-2/2022(H3N8) strain, isolated from a human[25], while the NS gene was most similar to A/Duck/Jiangxi/447/2022(H3N8) isolated from duck. The PB2, NP and M genes showed the highest homology to these AIVs isolated from chickens. These results indicate that the H3N3 virus likely originated from multiple host species, including humans, ducks, and chickens, highlighting the complex reassortment dynamics occurring within the avian influenza virus population.
Phylogenetic analysis of HA and NA genes
To explore the evolutionary relationships of the novel H3N3 AIV, phylogenetic trees were constructed by using the full-length nucleotide sequences of the surface genes (HA gene and NA gene). The best-fit models for the phylogenetic trees based on HA and NA genes were determined by using MEGA software. As listed in supplementary Table S1, the HA genetic tree was performed by using the GTR + G + I model, and the TN93 + G model was utilized for the phylogenetic tree of the NA gene. Phylogenetic analyses of the HA and NA genes revealed a clear division of the reference AIV isolates into Eurasian and North American lineages. The newly characterized A/chicken/China/16/2023(H3N3) strain was confidently assigned to the Eurasian lineage (Fig. 3).
H3 subtype influenza viruses are typically classified based on their host range, including human, avian, swine and equine. Phylogenetic analysis of the HA gene further supports this classification. As shown in Fig. 3, the maximum likelihood (ML) tree of the HA gene definitively places the novel H3N3 strain within the Eurasian lineage of avian-origin H3 AIVs and confirms its derivation from H3N8 AIVs.
Furthermore, the novel H3N3 strain had the closest genetic evolutionary relationship to A/chicken/China/NXFWB5/2022(H3N8) isolated from chicken in 2022 (Fig. 3A). Strikingly, within the Eurasian lineage of the HA gene-based evolutionary tree, both human-derived H3N8 strains, A/China/ZMD-22-2/2022(H3N8) and A/China/CSKFQ-22-5/2022(H3N8), clustered closely with the novel H3N3 strain. These strains shared high maximum inter-lineage sequence identities of 98.8% and 97.6%, respectively (Fig. 3A). Interestingly, the novel identified H3N3 AIV occupied a basal position within a subclade containing A/chicken/China/NXFWB5/2022(H3N8) and A/China/ZMD-22-2/2022(H3N8), suggesting a potential evolutionary origin of the HA genes in these H3N8 isolates from the novel H3N3 strain.
Phylogenetic analysis of the NA gene, as depicted in Fig. 3B, revealed a clear division into American and Eurasian lineages. The novel H3N3 isolate clustered within the Eurasian lineage of H3 subtype AIVs. Within the subclade containing the novel H3N3 AIV, four viruses were closely related. A/Environment/China/04940NA/2021(H10N3) exhibited the closest relationship to the H3N3 isolate (Fig. 3B). Additionally, a human-derived H10N3 strain, A/China/0428NA/2021, shared a high 98.2% sequence identity with the H3N3 virus. Two other chicken-derived H10N3 isolates, A/chicken/China/189-6/2021 and A/chicken/China/146-6/2021, also demonstrated high sequence similarity (98.5%) to the H3N3 isolate. These findings strongly support a common ancestral origin for the H3N3 and these four H10N3 viruses, further solidifying the phylogenetic placement of the H3N3 isolate based on the neuraminidase (NA) gene.
Phylogenetic analysis of internal genes
To further elucidate the evolutionary history of this novel H3N3 AIV, phylogenetic analyses were conducted on the full-length nucleotide sequences of the remaining six internal genes. The optimal substitution model and its corresponding parameters for tree construction are detailed in Supplementary Table S1. The GTR + G + I model was used for the phylogenetic analysis of PB2, PB1, PA and NP genes, and the phylogenetic trees based on M and NS genes were performed using the HKY + G model. Overall, among the internal genes of the novel H3N3 isolate, four genes (PB2, PB1, PA and NP) were classified in the Eurasian lineage of H9N2 AIVs, as well as two genes (M and NS) were categorized into the Eurasian lineage of H3N8 AIVs.
In the ML tree of the PB2 gene (Supplementary Fig. S1), the H3N3 AIV isolated in this study were closely related to prevalent H9N2 AIVs, specifically A/chicken/China/71/2019 and A/chicken/China/1102/2019, circulating in Chinese poultry. These findings suggest that the PB2 gene of the novel H3N3 strain may have originated from chicken-derived H9N2 AIVs in 2019. In contrast, analysis of the PB1 gene (Supplementary Fig. S2) indicated a closer genetic relationship between the H3N3 strain and A/environment-air/Kunshan/NIOSH-BL20/2018 (H9N2) and A/swine/Shandong/TA009/2019 (H9N2). Strong bootstrap support for these relationships suggests a common ancestor for these viruses. Importantly, a human-derived H7N9 strain (A/Changsha/41/2017(H7N9)) identified in 2017 shares 97.7% genetic similarity with the novel H3N3 strain within the same subclade. This close relationship suggests that the PB1 gene of human-origin H7N9 influenza viruses may have originated from the prevalent H9N2 AIVs. Furthermore, it highlights the potential for the PB1 gene of the novel H3N3 strain to recombine with human-origin H7N9 influenza viruses.
In the Eurasian lineage of the PA gene-based phylogenetic tree (Supplementary Fig. S3), the novel H3N3 strain was rooted as a subclade that included some chicken-derived H9N2 or H3N8 isolates, as well as a human-derived H3N8 isolate. This suggests a potential evolutionary relationship where the PA genes of these H9N2 and H3N8 isolates may have originated from the novel H3N3 strain. Conversely, all other strains within this subclade were isolated in 2022, excluding the novel H3N3 strain. This suggests that the PA gene of the newly isolated H3N3 strain likely originated from circulating H9N2 AIVs present before 2022. Evolutionary analysis further indicates that the most probable source is A/chicken/Shanghai/11/2018 (H9N2), identified in 2018.
The ML tree of the NP gene reveals that the novel H3N3 strain had the closest evolutionary affiliation with A/chicken/China/HP-1/2018(H9N2), followed by A/chicken/Anhui/12.25_YHZGS004-O/2018(H9N2) (Supplementary Fig. S4), proved that the NP gene of the novel H3N3 strain might also be provided by H9N2 AIVs prevalent in China. The M gene of the newly isolated H3N3 strain clustered within the Eurasian lineage of AIVs, specifically among H3N8 AIVs isolated from chickens (Supplementary Fig. S5). This suggests that the M gene of the novel H3N3 strain originated from chicken-derived H3N8 AIVs. For the phylogenetic analysis of NS gene, the novel H3N3 strain demonstrates its closest genetic connection with A/silkie chicken/Huizhou/411/2022(H3N8), followed by A/Duck/Jiangxi/447/2022(H3N8) and A/chicken/Huizhou/2542/2021(H3N8) (Supplementary Fig. S6). Meanwhile, the novel H3N3 strain and the three isolates belonged to a distinct subclade, the H3N3 strain being the root of the subclade, suggesting that the NS gene of the other three H3N8 AIVs may have originated from it.
Overall, Comprehensive phylogenetic analysis of all the novel H3N3 AIV isolate’s genes revealed a complex reassortment history. This virus emerged from the reassortment of H3N8 (from 2022), H9N2 (from 2018–2022), and H10N3 (from 2021) influenza virus subtypes. These parental viruses were isolated from chickens, humans, and the environment and were widely distributed in China, highlighting the intricate reassortment events leading to the novel H3N3 AIV (Fig. 4).
Molecular characterization
To assess the potential risk of the novel H3N3 strain spilling over to mammals, we conducted a molecular characterization of its amino acid sequences. The H3N3 AIV cleavage site between HA1 and HA2 contains the amino acid motif PEKQTR/GIF, typical of avian H3 subtype influenza viruses. This monobasic cleavage site characteristic classifies the novel isolate as a low pathogenic avian influenza virus (LPAIV). It is well known that the receptor binding sites (RBSs) motif of HA protein are critically crucial for cellular receptor specificity and determine the host range of influenza viruses[26, 27]. We analyzed the amino acid sequences at the RBSs of the HA protein in the novel H3N3 strain and other referenced H3N3 and H3N8 influenza viruses. According to Table 3, the amino acids at the RBSs of the novel H3N3 strain were well conserved and did not show variation compared to the related H3N8 viruses isolated from humans and chickens. Consistent with other H3N3 AIVs, the conserved amino acids Q226 and T228 (H3 numbering) were identified in the RBS of this novel H3N3 isolate (Table 3), suggesting that this H3N3 isolate also binds to the α-2,3-linked sialic acid receptor, which is generally considered to be the primary receptor in avian species.
Table 3
Amino acid comparison of cleavage and receptor binding sites in the HA protein of the novel H3N3 isolate and reference influenza viruses.
Strains
|
Cleavage sites
|
Receptor-binding sites (H3 numbering)
|
98
|
138
|
153
|
155
|
183
|
190
|
194
|
195
|
226
|
228
|
A/chicken/China/16/2023(H3N3)
|
PEKQTR/GIF
|
E
|
T
|
S
|
C
|
T
|
F
|
Y
|
I
|
Q
|
T
|
A/China/ZMD-22-2/2022(H3N8)a
|
PEKQTR/GLF
|
E
|
T
|
S
|
C
|
T
|
F
|
Y
|
I
|
Q
|
T
|
A/China/CSKFQ-22-5/2022(H3N8)a
|
PEKQTR/GLF
|
E
|
T
|
S
|
C
|
T
|
F
|
Y
|
I
|
Q
|
T
|
A/Iran/Clinical Sample/2019(H3N3)a
|
PEKQTR/GIF
|
K
|
N
|
S
|
C
|
T
|
F
|
Y
|
I
|
Q
|
A
|
A/bantam/Nanchang/9-366/2000(H3N3)
|
PEKQTR/GLF
|
E
|
T
|
G
|
C
|
T
|
F
|
Y
|
I
|
Q
|
T
|
A/duck/Zhejiang/D16/2013(H3N3)
|
PEKQTR/GLF
|
E
|
T
|
N
|
C
|
T
|
F
|
Y
|
I
|
Q
|
T
|
A/chicken/China/NXFWB5/2022(H3N8)
|
PEKQTR/GLF
|
E
|
T
|
S
|
C
|
T
|
F
|
Y
|
I
|
Q
|
T
|
A/chicken/Shantou/481/2022(H3N8)
|
PEKQTR/GLF
|
E
|
T
|
S
|
C
|
T
|
F
|
Y
|
I
|
Q
|
T
|
A/duck/Hunan/199/2014(H3N8)
|
PEKQTR/GLF
|
E
|
T
|
G
|
C
|
T
|
F
|
Y
|
I
|
Q
|
T
|
aThe strain was isolated from human. |
The specific polypeptide sequence for N-linked potential glycosylation was defined by the amino acid configuration of Asn-X-Ser/Thr, where X can represent any amino acid apart from proline[28]. The N-linked potential glycosylation of HA protein is associated with viral pathogenicity and affinity for the influenza virus receptor[29, 30]. For the novel H3N3 isolated in this study, the HA protein shared six N-linked potential glycosylation sites (PGSs) at positions 38, 54, 61, 181, 301 and 499, which were highly conserved with these referenced H3N8 viruses isolated from chickens and humans (Supplementary Table S2). Notably, eight PGSs of NA protein were detected at the positions of 14, 57, 66, 72, 146, 308, 387, and 401 (Supplementary Table S3). Comparative to the most similar H10N3 isolate at the NA gene level, the novel H3N3 strain isolated in this study has two additional PGSs at positions 387 and 401, respectively; whether this will alter the viral characteristic needs to be further investigated.
Furthermore, a detailed amino acid analysis of the gene segment of the novel H3N3 virus isolated in this study was carried out. As shown in Table 4, several mutations were observed including in the NA protein (M26I), PB2 protein (L89V, K251R, G309D, T339K, Q368R, R389K, H447Q, R477G and I495V), PB1 protein (D/A3V, L13P, R207K, K328N, S375N/T, H436Y, L473V and D622G), PA protein (H266R, K356R, N383D, S409N and S/A515T), NP protein (I353V), M1 protein (V15I, N30D and T215A), and NS1 protein (P/A42S, D97E and V149A). These mutations were reported to be related to an increase in the viral replication ability and virulence of AIVs in mice or mammalian cells (Table 4). Besides, mutations of PB1 protein (I368V) and PA protein (F277S, C278Q and L653P), which may facilitate the adaptation of AIVs to mammalian hosts, were also found (Table 4).
Table 4
Molecular characterization of the novel H3N3 isolate presented in this study.
Viral protein
|
Amino acid
|
A/chicken/China
/16/2023(H3N3)
|
Comments
|
References
|
HA
|
T160A
|
A
|
Increased binding to human-type influenza receptor
|
[1]
|
Q226L
|
Q
|
Increased binding to human-type influenza receptor
|
[2]
|
NA
|
M26I
|
I
|
Increased virulence in mice
|
[3]
|
V116A
|
V
|
Resistance to neuraminidase inhibitors
|
[4]
|
I117T
|
T
|
Reduces susceptibility to oseltamivir and zanamivir
|
[5]
|
E119V
|
E
|
Resistance to neuraminidase inhibitors
|
[6, 7]
|
Q136L
|
Q
|
Resistance to neuraminidase inhibitors
|
[8]
|
R152W
|
R
|
Resistance to neuraminidase inhibitors
|
[6, 9]
|
PB2
|
L89V
|
V
|
Enhanced polymerase activity and virulence in mammals
|
[10]
|
E158G
|
E
|
Enhanced polymerase activity and virulence in mice
|
[11]
|
K251R
|
R
|
Increased virulence in mice
|
[12]
|
G309D
|
D
|
Enhanced polymerase activity and virulence in mice
|
[10]
|
T339K
|
K
|
Enhanced polymerase activity and virulence in mice
|
[10]
|
Q368R
|
R
|
Increased polymerase activity and virulence in mice
|
[13, 14]
|
R389K
|
K
|
Enhanced polymerase activity and virulence in mice/Mammalian host adaptation
|
[15]
|
H447Q
|
Q
|
Increased polymerase activity and virulence in mammals
|
[13, 14]
|
R477G
|
G
|
Enhanced polymerase activity and virulence in mice
|
[10]
|
I495V
|
V
|
Enhanced polymerase activity and virulence in mice
|
[10]
|
Q591K
|
Q
|
Enhanced replication efficiency and virulence in mice
|
[11]
|
E627K
|
E
|
Mammalian host adaptation
|
[10, 16, 17]
|
D701N
|
D
|
Increased polymerase activity and viral replication in mammalian
|
[18]
|
PB1
|
D/A3V
|
V
|
Increased polymerase activity and virulence in mice
|
[14]
|
L13P
|
P
|
Increased polymerase activity and virulence in mammals, Mammalian host marker
|
[19]
|
H99Y
|
H
|
Increased transmissibility in ferrets
|
[11, 15]
|
R207K
|
K
|
Increased polymerase activity in mammalian cells
|
[20]
|
K328N
|
N
|
Increased polymerase activity and virulence in mammals
|
[14]
|
I368V
|
V
|
Mammalian host adaptation
|
[11]
|
S375N/T
|
N
|
Increased polymerase activity and virulence in mammals, human host marker
|
[14]
|
H436Y
|
Y
|
Increased polymerase activity and virulence in mallards, ferrets and mice
|
[20]
|
L473V
|
V
|
Increased polymerase activity and replication efficiency
|
[21]
|
D622G
|
G
|
Increased polymerase activity and virulence in mice
|
[22]
|
PB1-F2
|
N66S
|
N
|
Increased replication, virulence and antiviral response in mice/Increased virulence in mammals
|
[23]
|
T68I
|
T
|
Increased virulence in mice
|
[24]
|
PA
|
S37A
|
S
|
Increased polymerase activity and viral replication in mammalian cells
|
[25]
|
T97I
|
T
|
Enhances polymerase and virulence
|
[7, 11]
|
V100A
|
V
|
Contributed to the virulence and mammalian adaptation
|
[26, 27]
|
H266R
|
R
|
Increased polymerase activity and virulence in mammals and birds
|
[13]
|
F277S
|
S
|
Mammalian host adaptation
|
[13]
|
C278Q
|
Q
|
Mammalian host adaptation
|
[13]
|
K356R
|
R
|
Enhanced virulence and mammalian adaptation
|
[28]
|
N383D
|
D
|
Increased polymerase activity and mammalian adaptation
|
[29]
|
A404S
|
A
|
Human host marker
|
[30]
|
S409N
|
N
|
Increased polymerase activity, viral replication and virulence to mammalian; enhanced transmission
|
[31]
|
S/A515T
|
T
|
Increased polymerase activity and virulence in mammals and birds
|
[20]
|
L653P
|
P
|
Mammalian host adaptation
|
[13]
|
NP
|
F253I
|
I
|
Results in attenuated pathogenicity of the virus in mice
|
[25]
|
A286V
|
A
|
Affect the pathogenicity of the virus in mice
|
[32]
|
N319K
|
N
|
Increased virulence in mammalian cells
|
[7, 33]
|
I353V
|
V
|
Increased virulence in mice
|
[12]
|
T437M
|
T
|
Affect the pathogenicity of the virus in mice
|
[32]
|
M1
|
V15I
|
I
|
Increased virulence in mammals
|
[15, 21, 34]
|
N30D
|
D
|
Increased pathogenicity to mice
|
[35]
|
A166V
|
A
|
Increased polymerase activity and virulence in mammals
|
[21]
|
T215A
|
A
|
Increased virulence in mice
|
[35]
|
M2
|
L26P
|
L
|
Reduced susceptibility to amantadine and rimantadine
|
[36]
|
V27A/I
|
V
|
Reduced susceptibility to amantadine and rimantadine
|
[36]
|
A30T
|
A
|
Reduced susceptibility to amantadine and rimantadine
|
[36]
|
S31N
|
N
|
Reduced susceptibility to amantadine and rimantadine
|
[36, 37]
|
L55F
|
F
|
Mammalian host marker
|
[38]
|
NS1
|
P/A42S
|
S
|
Increased virulence in mice; Antagonism of
IFN induction
|
[39]
|
T/D92E
|
D
|
Increased virulence and antiviral response in mammals
|
[40]
|
D97E
|
E
|
Increased pathogenicity to mice
|
[15]
|
V149A
|
A
|
Pathogenicity in mice; Antagonism of IFN induction
|
[41]
|
NEP/NS2
|
M31I
|
M
|
Increased virulence in mice
|
[7, 42]
|
Moreover, the detection of a T160A substitution in the HA protein suggests enhanced binding affinity of the H3N3 isolate to human-type influenza receptors [37]. In contrast, the absence of Q591K, E627K, and D701N mutations in the PB2 protein indicates a low potential for cross-species transmission to mammals [38–42]. The NA protein's I117T mutation may reduce susceptibility to oseltamivir and zanamivir [43], while P/A42S and V149A substitutions in the NS1 protein could enhance interferon resistance [44, 45]. Additionally, the S31N mutation in the M2 protein confers resistance to amantadine and rimantadine [46, 47].
Overall, molecular characterization reveals that the novel H3N3 isolate predominantly binds to avian-type receptors despite possessing several mutations in internal genes that could potentially increase viral resistance, virulence, and transmission in mammalian hosts.