Key Amino Acids of M1-41 and M2-27 Determine Growth and Pathogenicity of Chimeric H17 Bat Influenza Virus in Cells and in Mice

ABSTRACT

Based on our previous studies, we show that the M gene is critical for the replication and pathogenicity of the chimeric H17 bat influenza virus (Bat09:mH1mN1) by replacing the bat M gene with those from human and swine influenza A viruses. However, the key amino acids of the M1 and/or M2 proteins that are responsible for virus replication and pathogenicity remain unknown. In this study, replacement of the PR8 M gene with the Eurasian avian-like M gene from the A/California/04/2009 pandemic H1N1 virus significantly decreased viral replication in both mammalian and avian cells in the background of the chimeric H17 bat influenza virus. Further studies revealed that M1 was more crucial for viral growth and pathogenicity than M2 and that the amino acid residues M1-41V and M2-27A were responsible for these characteristics in cells and in mice. These key residues of the M1 and M2 proteins identified in this study might be important for influenza virus surveillance and could be used to produce live attenuated vaccines in the future.

IMPORTANCE The M1 and M2 proteins influence the morphology, replication, virulence, and transmissibility of influenza viruses. Although a few key residues in the M1 and M2 proteins have been identified, whether other residues of the M1 and M2 proteins are involved in viral replication and pathogenicity remains to be discovered. In the background of the chimeric H17 bat influenza virus, the Eurasian avian-like M gene from the A/California/04/2009 virus significantly decreased viral growth in mammalian and avian cells. Further study showed that M1 was implicated more than M2 in viral growth and pathogenicity in vitro and in vivo and that the key amino acid residues M1-41V and M2-27A were responsible for these characteristics in cells and in mice. These key residues of the M1 and M2 proteins could be used for influenza virus surveillance and live attenuated vaccine applications in the future. These findings provide important contributions to knowledge of the genetic basis of the virulence of influenza viruses.

INTRODUCTION

The M gene of segmented influenza viruses contributes frequently to recombination between different subtypes of influenza A viruses (IAVs). For example, the H9N2 avian influenza virus M gene can reassort with other subtypes to generate new influenza viruses—H5N1 (1, 2), H7N9 (3, 4), and H10N8 (5)—and cause serious zoonotic diseases in humans and animals. The M gene packaging signals of bat and conventional IAVs are compatible; none of the other five internal gene segments (NP, NS, PB2, PB1, and PA) can be replaced by the respective segment from conventional IAVs in the background of the chimeric bat influenza virus (6, 7). Different M gene reassortant viruses (from the classical human M gene, the North American classical M gene, and the Eurasian swine M gene) can be generated through reverse genetics in the background of the chimeric H17 bat influenza virus (Bat09:mH1mN1) (6, 8), demonstrating the high reassortment ability of the M gene in the background of bat IAV.

In addition to the higher compatibility of the M gene than of other internal genes for IAV reassortment, the M1 and M2 proteins can also influence virion morphology (9–11), replication and virulence (8, 12–16), and transmissibility (10, 17, 18). The M gene is also associated with host restrictions, probably due to frequent adaptation in different hosts during IAV evolution (19, 20). There are five main lineages of M genes from different hosts (human, swine, and avian): the classical human M gene, the North American classical swine M gene, the Eurasian swine M gene, the Eurasian avian M gene, and the Eurasian avian-like M gene (20). In our previous study, the reassortant virus with the PR8 M gene (classical human M gene) in the background of the Bat09:mH1mN1 virus displayed higher pathogenicity in mice than the reassortant virus with the TX98 M gene (North American classical swine M gene) or with the KS-107824 M gene (Eurasian swine M gene) (8). The 2009 pandemic H1N1 virus (pH1N1) was a reassortant between the North American triple reassortant (which provided PB1, PB2, PA, HA, NP, and NS) and Eurasian avian-like swine influenza viruses (which provided NA and M), had enhanced pathogenicity and transmission, and caused the latest global influenza pandemic, which started in Mexico in April 2009 (10, 11, 21). Whether the pH1N1 M protein can enhance the growth and pathogenicity of the chimeric bat virus, and which key amino acids of the M protein determine pathogenicity, remain unknown. Although a few key residues in the M1 and M2 proteins have been identified in previous studies (12, 14, 15, 17), more residues of the M1 and M2 proteins involved in viral replication and pathogenicity need to be discovered.

In the current study, in the background of the chimeric H17 bat influenza virus (Bat09:mH1mN1), a reassortant virus with the M gene from A/California/04/2009 (the CA09/H1N1 virus with the Eurasian avian-like swine M gene), the growth kinetics in the MDCK, A549, and LMH cell lines were compared with those of chimeric H17 bat viruses with the M gene from A/Puerto Rico/8/1934 (PR8/H1N1 virus with the classical human M gene, a mouse-adapted virus) or A/Swine/Texas/4199-2/98 (the TX98/H3N2 virus with the North American classical swine M gene) (8). Through site-directed mutation, reverse genetics, viral growth curves, and an animal infection study, the key amino acids of the M1 and M2 proteins determining viral growth in cells and pathogenicity in mice were further characterized in the background of the chimeric H17 bat influenza virus and the CA09/H1N1 virus.

RESULTS

The Eurasian avian-like M gene significantly decreased viral replication in cells.

The chimeric H17 bat influenza virus with an M gene from either the CA09/H1N1 virus, the PR8/H1N1 virus, or the TX98/H3N2 virus was generated as described previously (8) (Fig. 1A) and was named Bat09:mH1mN1-CA09(M), Bat09:mH1mN1-PR8(M), or Bat09:mH1mN1-TX98(M), respectively. The replication of the three chimeric bat viruses with different M genes was compared in different cell lines (MDCK, A549, and LMH) (Fig. 1B to D). The growth curves showed that the PR8 M gene reassortant virus underwent significantly higher levels of replication in all cell lines tested than the TX98 M and CA09 M gene reassortant viruses at 24, 48, and 72 h postinoculation (hpi), as seen in Fig. 1 (P, <0.05 or <0.01). The CA09 M (Eurasian avian-like M) gene virus significantly decreased viral replication in MDCK, A549, and LMH cells at 24, 48, and 72 hpi (P, <0.05 or <0.01), after replacing the PR8 M gene in the background of Bat09:mH1mN1-PR8(M), as seen in Fig. 1.

FIG 1.

Schematic diagram of chimeric H17 bat influenza viruses with different M genes and graphs showing viral growth kinetics in mammalian and avian cells. (A) Schematic diagram of viruses with different M genes in the background of Bat09:mH1mN1. Dark blue bars represent the five internal segments or HA/NA packaging signal sequences from A/little yellow shouldered bat/Guatemala/164/2009 (H17N10); green bars represent HA/NA coding regions or the M segment from A/Puerto Rico/8/1934 (H1N1); the light blue bar represents the M segment from A/Swine/Texas/4199-2/98 (H3N2); the brown bar represents the M segment from A/California/04/2009 (H1N1). (B to D) Growth dynamics of chimeric H17 bat influenza viruses in MDCK (B), A549 (C), and LMH (D) cells. Cell monolayers were infected with each virus at an MOI of 0.01, and samples were collected at the indicated time points. The virus titers were determined in MDCK cells. Each data point indicates the mean ± standard error of the mean from three independent experiments, *, P < 0.05; **, P < 0.01.

The M1-mut virus showed significantly lower levels of replication than the M2-mut virus in cells.

To determine whether M1 or M2 protein differences influenced viral replication or growth in cells (Fig. 1), 11 common amino acid substitutions for PR8 M compared with both TX98 M and CA09 M were characterized (Fig. 2A and B), including 4 positions in the M1 protein and 7 in the M2 protein. The M1-mut (V41A I115V A137T N231D) and M2-mut (G21D A27V G61R K70E K78Q A86V S93N) viruses were generated in the background of Bat09:mH1mN1-PR8(M) with four substitutions in the M1 protein and seven in the M2 protein (Fig. 2B). From the growth curves in Fig. 2, when the M1-mut and M2-mut viruses were compared with Bat09:mH1mN1-PR8(M), the M1-mut virus showed significantly reduced replication in MDCK, A549, and LMH cells at 24, 48, and 72 hpi (P < 0.01), but the M2-mut virus showed significantly reduced replication only in A549 cells at 24 and 48 hpi (P < 0.05). The results indicated that both the M1-mut and M2-mut viruses could reduce viral replication in different cells, with a greater reduction with the M1-mut virus.

FIG 2.

Schematic diagram of IAV segment 7 and its encoded proteins labeled with amino acid differences of TX98 M and CA09 M from PR8 M and graphs of the growth dynamics of the M1-mut and M2-mut viruses in mammalian and avian cells. (A) Schematic diagram of gene sequences encoding the M1 and M2 proteins. Protein positions where amino acids in TX98M and CA09 M commonly differ from those in PR8 M are shown. (B) Primary structural maps for the M1 and M2 proteins, labeled with the common amino acid differences of TX98 M and CA09 M from PR8 M. For example, V41A represents V (Val) in position 41 of PR8 M1 and A (Ala) in position 41 of TX98 M1 and CA09 M1. (C to E) Monolayers of MDCK (C), A549 (D), and LMH (E) cells were infected with each virus at an MOI of 0.01, and samples were collected at the indicated time points. The virus titers were determined in MDCK cells. Each data point indicates the mean ± standard error of the mean from three independent experiments, *, P < 0.05; **, P < 0.01.

The key amino acids determining viral replication in cells.

To further explore the key amino acids of the M1 and M2 proteins determining viral replication in cells, six single-site mutation viruses (V41I, I115V, A137T, and N231D in M1; G21D and A27V in M2) were generated in the background of Bat09:mH1mN1-PR8(M), and a multisite mutation virus with five substitutions in the M2 cytoplasmic domain was also generated and was named the M2-C domain virus (G61R K70E K78Q A86V S93N). The replication of these M mutant viruses was compared in MDCK and LMH cells by the growth curves seen in Fig. 3. Since no chimeric bat viruses could replicate well in the A549 cell line, especially the chimeric bat virus with CA09 M or TX98 M (Fig. 1) and the M1-mut and M2-mut viruses (Fig. 2), which replicated with very low titers in the A549 cell line, this study did not use A549 cells for the next growth curve study. In MDCK cells (Fig. 3A), the M1-V41A and M2-A27V viruses showed significantly lower levels of replication than Bat09:mH1mN1-PR8(M) (P < 0.05); the M1-I115V and M2-G21D viruses also showed significantly reduced replication at 72 hpi (P < 0.05), but not at 12, 24, and 48 hpi (P > 0.05), while the other mutant viruses (M1-A137T, M1-N231D, and M2-C domain) showed almost the same growth curve as the Bat09:mH1mN1-PR8(M) virus at 24, 48, and 72 hpi. In LMH cells (Fig. 3B), the M1-V41A, M1-I115V, M2-G21D, and M2-A27V viruses showed significantly lower levels of replication than the Bat09:mH1mN1-PR8(M) virus (P < 0.05), while other mutant viruses (M1-A137T, M1-N231D, and M2-C domain) showed a growth curve similar to that of Bat09:mH1mN1-PR8(M), with only small reductions at 24 and 48 hpi (P > 0.05), as seen in Fig. 3B. Our results indicated that the M1-V41A, M1-I115V, M2-G21D, and M2-A27V viruses showed reduced replication in both mammalian and avian cells; these were selected for further investigation in mice.

FIG 3.

Growth dynamics of different M mutant viruses in mammalian and avian cells. The Bat09:mH1mN1-PR8(M) virus was set as the parental virus control. Monolayers of MDCK (A) and LMH (B) cells were infected at an MOI of 0.01 with each mutant virus, including the M1-V41A, M1-I115V, M1-A137T, M1-N231D, M2-G21D, M2-A27V, and M2-C domain (G61R K70E K78Q A86V S93N) viruses, and samples were collected at the indicated time points. The virus titers were determined in MDCK cells. Each data point indicates the mean ± standard error of the mean from three independent experiments, *, P < 0.05; **, P < 0.01.

The key amino acids determining viral replication and pathogenicity in mice.

Mice are not natural hosts for IAVs, but they are useful models for studying antiviral immune responses and the pathogenesis of influenza viruses. Based on the results from growth curves in cells, the key amino acids determining viral replication and pathogenicity in mice were further characterized. Six mutant viruses—the M1-mut, M2-mut, M1-V41A, M1-I115V, M2-G21D, and M2-A27V viruses—were selected to infect mice, and the Bat09:mH1mN1-PR8(M) virus and phosphate-buffered saline (PBS) groups were used as controls (Fig. 4).

FIG 4.

Weight loss, survival rates, and viral replication in mice infected with each indicated virus at a dose of 2.5 × 10⁴ TCID₅₀s/mouse. In each group, eight mice were infected intranasally. PBS was given to negative-control mice. (A) Body weights were monitored daily until 14 dpi, and animals with a >25% loss of original body weight were humanely euthanized. (B) Survival rates were calculated for mice until 14 days after viral infection. (C and D) Three mice from each group were randomly selected to be euthanized at 4 dpi. Viral replication was determined in MDCK cells for mouse nasal turbinate (C) and lung (D) samples collected at 4 dpi.

To evaluate viral replication in nasal turbinate and lung samples, three mice from each group were randomly selected to be euthanized at 4 days postinoculation (dpi). Prior to necropsy, the Bat09:mH1mN1-PR8(M), M2-mut, M1-I115V, and M2-G21D viruses caused severe clinical signs of ruffled fur, dyspnea, or lethargy, which became worse after 4 dpi. Severe weight loss up to 25% was observed starting from 4 dpi, and 100% mortality was observed at 6 dpi for Bat09:mH1mN1-PR8(M), at 7 dpi for M2-mut and M1-I115V, and at 8 dpi for M2-G21D (Fig. 4A and B). These three mutant viruses (M2-mut, M1-I115V, and M2-G21D) maintained high pathogenicity for mice, like the parental Bat09:mH1mN1-PR8(M) virus, while the other three mutant viruses (M1-mut, M1-V41A, and M2-A27V) showed reduced pathogenicity in mice, with zero mortality and only transient weight loss up to 14 dpi. The M2-A27V virus caused weight loss of no more than 15% for 6 to 10 dpi, showing higher virulence than the M1-mut and M1-V41A viruses. Body weight and survival rates indicated that the M1-mut, M1-V41A, and M2-A27V viruses could significantly reduce viral pathogenicity in mice when the indicated substitutions occurred.

At 4 dpi, viral replication in nasal turbinate and lung samples was determined by titration in MDCK cells. The results showed efficient replication in mouse lungs of the parental Bat09:mH1mN1-PR8(M), M2-mut, M1-I115V, and M2-G21D viruses (>6 log₁₀ median tissue culture infectious doses [TCID₅₀s]/ml), while lower virus titers were found for the M1-mut, M1-V41A, and M2-A27V viruses (P < 0.05), as seen in Fig. 4D. No viral titers in nasal turbinate samples were detected at 4 dpi for the M1-mut, M1-V41A, and M2-A27V viruses (Fig. 4C). Viral replication in mouse tissues consistently showed that the M1-mut, M1-V41A, or M2-A27V substitution significantly attenuated the chimeric bat virus.

Histopathological analysis showed that the Bat09:mH1mN1-PR8(M), M2-mut, M1-I115V, and M2-G21D viruses induced typical influenza-induced pneumonia in mice, manifesting as mild to severe bronchoalveolar epithelial cell degeneration and necrosis, interstitial pneumonia, and even pulmonary congestion at 4 dpi (Fig. 5). The M1-mut, M1-V41A, and M2-A27V viruses caused little or no histopathological lung damage at 4 dpi (P < 0.01), as seen in Fig. 5. All the data indicated that the M1 protein and the key amino acids M1-41V and M2-27A played important roles in the replication and pathogenicity of the chimeric bat influenza virus in vitro and in vivo.

FIG 5.

H&E staining and histopathological scores for lung sections collected at 4 dpi from mice infected with each indicated virus at a dose of 2.5 × 10⁴ TCID₅₀s/mouse. In the H&E-stained sections, there are numerous sloughed cells, necrotic cells within the bronchiolar lumen, pyknotic cells and erythrocytes in adjacent alveoli, multifocal necrosis of luminal epithelium and interstitial pneumonia, severe disorganization of the bronchiolar epithelium, and macrophage and neutrophil infiltration in alveoli for the Bat09:mH1mN1-PR8(M), M2-mut, and M1-I115V viruses. For the M2-G21D and M2-A27V viruses, there are medium levels of bronchoalveolar disorganization, congestion, embolism, and neutrophil infiltration. There are no lesions or only slight lesions for the M1-mut, M1-V41A, and PBS groups. In the bar graph showing histopathological scores, microscopic lung scores are presented as means ± standard errors of the means for three mice in each group at 4 dpi. Significant differences between different groups are indicated (*, P < 0.05; **, P < 0.01).

To confirm the roles of M1-41V and M2-27A in viral pathogenicity in mice, the M1-A41V and M2-V27A viruses were further generated based on the viral backbones of Bat09:mH1mN1-CA09(M) and CA09/H1N1, and the effect on viral pathogenicity in mice was investigated. The results showed that these two substitutions in CA09 M could enhance virulence in mice (Fig. 6). The M1-A41V and M2-V27A mutant viruses could cause 50% and 25% mortality in mice, respectively, while the backbone virus Bat09:mH1mN1-CA09(M) could not kill the mice, resulting only in a weight loss of <13% over 1 to 10 dpi (Fig. 6A and B). In the CA09/H1N1 virus backbone, these two substitutions caused 100% mortality in mice, while the backbone virus CA09/H1N1 could cause only 25% mortality in mice (Fig. 6C and D). Both M1-41V and M2-27A were therefore characterized as responsible for increased growth and virulence in mice based on the viral backbone of Bat09:mH1mN1 or CA09/H1N1in this study.

FIG 6.

Weight loss and survival rates in mice infected with each indicated CA09 M-related virus at a dose of 2.5 × 10⁴ TCID₅₀s/mouse. In each group, eight mice were infected intranasally. PBS was given to negative-control mice. (A and C) Body weights were monitored daily until 14 dpi, and animals with a >25% loss of original body weight were humanely euthanized. (B and D) After viral infection, survival rates were calculated for mice until 14 dpi.

DISCUSSION

The M1 and M2 proteins are produced by differential splicing from segment 7 of IAVs (Fig. 2A). The M1 protein is multifunctional, playing many essential roles throughout the viral replication cycle. M1 is a 252-amino-acid (aa) protein that forms the major structural component of influenza virus particles and lines the inner surface of the viral membrane, so it is a major determinant of influenza virion morphology (11, 22). M1 also plays an essential role in viral assembly and budding, recruits the viral ribonucleoprotein (vRNP) complex to the cell surface, functions in transcription inhibition, and controls RNP incorporation by interacting with RNP, RNA, hemagglutinin (HA), neuraminidase (NA), and NP (23–25). M2 is an ion channel protein required for viral entry and replication (26, 27). In conjunction with interactions with HA, NA, NP, and the plasma membrane, M2 also binds to M1. M2 may directly associate with HA or with both HA and M1 to be recruited to the viral budding sites and functions as a mediator of membrane scission, which is important for virion formation and for virus budding and release (27–30).

Little has been reported about the direct contributions of the M1 and M2 proteins to the pathogenicity of IAVs. M1-A41V has been implicated in adaptation to laboratory substrates, including mice (15, 31). Previous studies have also suggested that position 43 of the M1 protein is responsible for the difference in pathogenicity between two highly pathogenic H5N1 avian influenza viruses; however, substitution at this position might not affect the fundamental functions of the M1 protein (12). The role of the M2 protein cytoplasmic tail in viral replication was examined by deletion of 28 amino acids, which resulted in significantly lower production of infectious viral particles than that of the wild-type virus (29). Since the functional domains of M1 and M2 are not completely understood, the mechanism for changes in M-related replication and pathogenicity remains unclear (32).

Our previous study showed that the bat M gene could be replaced by M genes from human and swine IAVs to generate live chimeric bat viruses and could influence the pathogenicity of reassortant viruses (8). The current study further evaluated the Eurasian avian-like M gene (from the CA09/H1N1 pandemic virus) and its roles in viral replication in different cells. Neither CA09/H1N1 M nor TX98 M (North American classical swine M) could increase viral growth in mammalian and avian cells (Fig. 1); they showed significantly lower levels of growth in cells than the chimeric H17 bat virus with PR8 M. This study performed M mutagenesis studies by mutating to lower-virulence viruses, theoretically in the background of a higher-virulence virus, i.e., Bat09:mH1mN1-PR8(M) (Fig. 2 to 4) (8). M1-V41A and M2-A27V were characterized as significantly attenuated substitutions for the chimeric H17 bat virus with PR8 M in vitro and in vivo. In this study, the PR8 M used is from a mouse-adapted virus (A/Puerto Rico/8/1934) that has the reported mouse adaptation mutation of M1-41V (15). Based on human and swine IAV sequences obtained from the Influenza Research Database (http://www.fludb.org) from 2000 to 2020, the amino acids at position 41 of the M1 protein and position 27 of the M2 protein were analyzed (Table 1). A total of 17,256 M1 sequences and 17,261 M2 sequences of viruses isolated from humans, and 2,828 M1 sequences and 2,821 M2 sequences of viruses isolated from swine, were examined. These analyses showed that the amino acid at M1-41 was highly conserved as Ala (A) for human (17,231/17,256 [99.86%]) and swine (2,808/2,828 [99.29%]) IAVs, suggesting that Ala at position 41 of the M1 protein is important for the viral life cycle, and viral growth and virulence were significantly increased in cells and in mice after replacement of Ala with Val at this position, which is in the RNP binding domain for M1 (Fig. 2). Structurally, amino acid 41 is located in the N-terminal domain and folds into the third helix of the M1 protein (Fig. 7); mutation at this position is likely to affect pH-dependent association/dissociation of M1 with the vRNP and to control virulence as well as growth by binding RNP less strongly and allowing replication to proceed to higher levels (33). The WSN M1 protein with the A41V mutation has been shown to dissociate at a higher pH, which was implicated in increased growth rates in MDCK cells (15, 31, 32). In contrast, substitution of M1-P41A in the strains of the Eurasian avian-like swine lineage has been reported to be associated with increased viral replication and transmission in guinea pigs and has also been implicated in determining the virion morphology as filamentous or spherical (17).

TABLE 1.

Natural amino acid residues at M1-41 and M2-27 for virus isolates obtained from the Influenza Research Database from 2000 to 2020^a

Amino acid position	Source of virus isolateb	Mutations (no. of isolates bearing each residue)	Total no. of virus isolationsc
M1-41	Human	A (17,231), V (23), S (1), T (1)	17,256
	Swine	A (2,808), V (16), P (1), R (1), S (1), E (1)	2,828
M2-27	Human	V (17,146), A (58), I (34), F (16), T (3), G (2), X (2)	17,261
	Swine	A (1,293), V (943), I (331), T (248), S (4), F (1), L (1)	2,821

^aAs of 3 September 2020.

^bFrom subtype searches for human H1N1 and swine H3N2.

^cFrom a search of all full-length protein sequences.

FIG 7.

M1 and M2 protein structure modeling by Phyre2 (39). (Top) M1-41-Val and M2-27-Ala are labeled in orange and red, respectively (protein structure imaged by PyMOL Viewer). (Bottom left) Three helix ribbons of M1 are shown in the image. The other nine helix ribbons of the M1 protein are not shown. M1-41 is located at the second amino acid position of the third helix of the N terminus of the M1 protein. (Bottom right) All three helix ribbons of M2 are shown in the image. M2-27 is located at the second amino acid position of the first helix of the N terminus of the M2 protein. (The helix secondary structure was drawn by Swiss-PdbViewer, version 4.1.0; helix ribbons are colored in rainbow order from the N to the C terminus.)

The natural isolation rate of M1-41V viruses is higher for swine viruses (16/2,828 [0.57%]) than for human viruses (23/17,256 [0.13%]) (Table 1). In 2009, the influenza pandemic was caused by a novel reassortant virus that originated from swine influenza viruses providing a Eurasian avian-like swine M gene, which increased viral pathogenicity and transmission between humans (10, 11, 21). The current study provided evidence that the M1-A41V substitution of CA09/H1N1 M could possibly increase viral growth and virulence, and this change in influenza surveillance should be a warning. This finding shows that M1-41 is a crucial position associated with the influenza virus life cycle, and that changes in this position, such as P or V, will significantly affect viral replication and pathogenicity. Deletion of the M2 cytoplasmic tail could decrease the production of infectious viruses by coordinating the efficient packaging of genome segments into influenza virus particles with a defect in vRNP packaging, and expression of the full-length M2 protein restored the replication of the M2-truncated virus (29). In the current study, the M2-C domain virus (with five substitutions in the M2 cytoplasmic tail) displayed the same growth curve as the parental Bat09:mH1mN1-PR8(M) virus in mammalian and avian cells, indicating that the substitutions (G61R K70E K78Q A86V S93N) in the M2 cytoplasmic tail did not influence viral replication as deletions did (Fig. 3). In cells in vitro, the M2-C domain virus (G61R K70E K78Q A86V S93N) had almost no change in viral replication, but each of these five mutations probably has different effects, since only M2-A86S was ever characterized as different between the virulent virus and the cold adaptation attenuated virus (34), while other mutations probably have virulence-increasing functions. The transmembrane (TM) domain (amino acids 24 to 42 in M2) is responsible for M2 ion channel activity (Fig. 2B). Watanabe and colleagues reported that an M2-del (with aa 29 to 31 deleted) mutant A/WSN/33 (H1N1) influenza virus with defective M2 ion channel activity replicated as efficiently as the wild-type virus in cells (35), although viral growth was attenuated in mice (35) and in tissue culture (36).

The M2 mutations in the TM domain—L26F, V27A, A30T/V, S31N, and G34E—are known to confer resistance to amantadine (36–38). The results of this study showed that the M2-A27V substitution significantly decreased viral growth in cells and pathogenicity in mice, proving that the M2–27 position is responsible for replication and virulence as well as for amantadine resistance. This position is in the N-terminal functional domain of the ion channel for M2 (Fig. 2B), and position 27 is in the domain folding into the first helix of M2 (Fig. 7). In drawing the image for the helix secondary structure by Swiss-PdbViewer, version 4.1.0 (39), this study found that the M2-27 position is like the M1-41 position in that it is located at the second amino acid position of a helix of the N terminus, but M2-27 is in the first helix of the M2 protein within ion channel areas and is suspected to have a greater effect on viral replication and virulence. Natural amino acid mutation analyses showed that the amino acid at M2-27 is comparatively variable, with seven residues for human IAVs and seven residues for swine IAVs (Table 1). The M2-27A residue characterized as responsible for increased growth and virulence in our study has already presented in 45.83% (1,293/2,821) of natural swine isolates and in 0.34% (58/17,261) of natural human isolates. The variation in M2-27 is possibly selected in pigs by host or immune factors; besides M2-27V (943/2,821 [33.42%]), the residue at M2-27 for swine IAVs is also naturally selected as M2-27I (331/2,821 [11.73%]) or M2-27T (248/2,821 [8.79%]), but the roles of these residues in viral growth and pathogenicity remain unresolved.

In summary, our results indicated that the Eurasian avian-like M gene (from the CA09/H1N1 pandemic virus) significantly decreased viral growth in mammalian and avian cells. The M1 protein is implicated more than the M2 protein in viral growth and pathogenicity in vitro and in vivo. The key amino acid residues M1-41V and M2-27A significantly enhanced viral growth in cells and pathogenicity in mice. These key residues of the M1 and M2 proteins could be used for influenza virus surveillance and live attenuated vaccine applications in the future.

MATERIALS AND METHODS

Ethics statement on animal use.

The present study was conducted in strict accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals of the Ministry of Science and Technology of the People’s Republic of China (40). All animal studies were conducted in a biosecurity level 2 laboratory approved by the Shanghai Veterinary Research Institute (SHVRI) of the Chinese Academy of Agricultural Sciences (CAAS) (SHVRI-SZ-20200816-02) and were approved by the Animal Association of Science and Technology Commission of Shanghai Municipality, China.

Cell culture and reagents.

Human embryonic kidney 293 T cells were maintained in a reduced-serum minimal Eagle’s medium (Opti-MEM; HyClone, Los Angeles, CA, USA) supplemented with 10% fetal bovine serum (FBS; PAN-Biotech, Aidenbach, Germany) and a 1% penicillin–streptomycin (PS) antibiotic solution (Thermo Fisher Scientific, Waltham, MA, USA). A549 human lung carcinoma epithelial cells and Madin-Darby canine kidney (MDCK) cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM; HyClone) containing 5% FBS and 1% PS. Chicken hepatocellular carcinoma LMH cells were maintained in DMEM–F-12 medium containing 5% FBS and 1% PS. Cells were inoculated with virus in MEM-infecting medium containing 0.3% bovine serum albumin (Sigma, St. Louis, MO, USA), 1 μg/ml tosylsulfonyl phenylalanyl chloromethyl ketone-treated trypsin (Sigma), and 1% PS. All cells were incubated at 37°C under 5% CO₂.

Chimeric bat influenza viruses with M genes of different origins.

The M segment viral RNA from CA09/H1N1 was subcloned into a pHW2000 vector. Based on the reverse genetic system of the modified H1N1 chimeric H17 bat influenza virus (Bat09:mH1mN1), which contained six internal genes from the H17N10 bat influenza virus (A/little yellow-shouldered bat/Guatemala/164/2009) and the chimeric hemagglutinin (HA) and neuraminidase (NA) from PR8/H1N1 modified with H17 and N10 packaging signals, a Bat09:mH1mN1-CA09(M) virus was generated by replacing bat M with CA09 M, as described previously (6, 8) (Fig. 1A). The other two chimeric bat viruses with PR8 M and TX98 M, named Bat09:mH1mN1-PR8(M) and Bat09:mH1mN1-TX98(M), respectively, were generated by the same strategy. The genomes of the rescued viruses were sequenced to confirm the absence of unexpected mutations. All three reassortant viruses were characterized and propagated in 9-day-old specific-pathogen-free chicken eggs (Merial-Vital, Beijing, China). All confirmed virus stocks were titrated on MDCK cells.

Viral growth kinetics of chimeric bat viruses with different M origins in mammalian and avian cells.

To evaluate the growth kinetics of reassortant viruses in mammalian and avian cells, monolayers of mammalian MDCK and A549 cells and avian LMH cells were infected with each virus at a multiplicity of infection (MOI) of 0.01. The supernatants from infected cells were collected at 12, 24, 48, and 72 h postinoculation (hpi) and were titrated on MDCK cells in 96-well plates. Three replicates were set for each sample. The virus titers were determined as the median tissue culture infectious dose (TCID₅₀) per milliliter using the method of Reed and Muench (41).

M1 and M2 mutant viruses.

The M1 and M2 protein sequences from PR8 M, TX98 M, and CA09 M were compared by MegAlign (Lasergene 7.0; DNAStar, Madison, WI, USA). After constructing different PR8 M mutagenesis plasmids with the commercial Mut Express II Fast Mutagenesis kit, v2 (Vazyme, Nanjing, China), we generated an M1 mutant virus (M1-mut) and an M2 mutant virus (M2-mut) in the background of the Bat09:mH1mN1-PR8(M) virus, with four substitutions in the M1 protein and seven in the M2 protein (Fig. 2 and 4).

To further explore the key amino acids in the M1 or M2 protein, the six site mutant viruses (V41A, I115V, A137T, and N231D in M1; G21D and A27V in M2) and an M2-C domain mutant virus (with five mutations in the M2 cytoplasmic domain [G61R K70E K78Q A86V S93N]) were generated with the same strategy in the background of the Bat09:mH1mN1-PR8(M) virus (Fig. 2B and 3), and the CA09-M1-A41V and CA09-M2-V27A mutant viruses were generated in the background of the Bat09:mH1mN1-CA09(M) virus and the CA09/H1N1 virus (Fig. 2B and 6). The primers used for site-directed mutagenesis experiments are available upon request. All mutant viruses were characterized, propagated, and titrated, and their replication in cells was compared as described above (Fig. 3 and 4).

Mouse experiments.

To determine the replication and pathogenicity of the chimeric bat viruses in vivo, eight 5-week-old female BALB/c mice (Jie Si Jie Laboratory Animals Co., Shanghai, China) were infected intranasally. These mice were anesthetized with dry ice prior to intranasal inoculation; each mouse received 2.5 × 10⁴ TCID₅₀s per 50 μl intranasally, and 50 μl PBS was given to each negative-control mouse (eight mice per group). Body weights were monitored daily until 14 days postinoculation (dpi), and animals with a >25% loss of original body weight were humanely euthanized. All remaining mice were kept for daily monitoring of body weight; clinical signs were observed twice daily after disease onset; and mice were euthanized at 14 dpi. Three mice from each group were randomly selected to be euthanized at 4 dpi. To determine virus replication in mice, nasal turbinate and lung samples were collected at 4 dpi and homogenized in PBS buffer (1 ml PBS/g) containing PS (100 U/ml penicillin and 0.05 mg/ml streptomycin). The tissues with PBS containing PS were homogenized using a TissueLyser instrument at 70 times/s for 2 min and were centrifuged at 12,000 × g for 10 min, and aliquots of the supernatants were collected and stored at −80°C for virus titration. The lung samples were collected, fixed in 10% formalin, and subsequently embedded in paraffin for hematoxylin-and-eosin (H&E) staining (Servicebio, Wuhan, China). The viruses were titrated on MDCK cells in order to determine virus replication in mouse tissues collected at 4 dpi (6, 8). Lung lesions were examined by a veterinary pathologist in a blinded fashion and were given scores from 0 to 10 to reflect the severity of bronchial epithelial injury as described previously (42).

Statistical analysis.

Statistical analysis was performed using Prism software (version 7; GraphPad, La Jolla, CA, USA). Pairwise comparisons between groups were conducted by Tukey’s multiple-comparison test (one-way or two-way analysis of variance). Differences were considered significant at P values of <0.05 or <0.01.