The Species-Specific 282 Residue in the PB2 Subunit of the Polymerase Regulates RNA Synthesis and Replication of Influenza A Viruses Infecting Bat and Nonbat Hosts

ABSTRACT

Bat influenza viruses are genetically distant from classical influenza A viruses (IAVs) and show distinct functional differences in their surface antigens. Nevertheless, any comparative analyses between bat and classical IAV RNA polymerases or their specific subunits are yet to be performed. In this work, we have identified signature residues present in the bat influenza virus polymerase which are responsible for its altered fitness in comparison to the classical IAVs. Through comparative sequence and structural analysis, we have identified specific positions in the PB2 subunit of the polymerase, with differential amino acid preferences among bat and nonbat IAVs. Functional screening helped us to focus upon the previously uncharacterized PB2-282 residue, which is serine in bat virus but harbors highly conserved glutamic acid in classical IAVs. Introduction of E282S mutation in the human-adapted PB2 (influenza A/H1N1/WSN/1933) drastically reduces polymerase activity and replication efficiency of the virus in human, bat, and canine cells. Interestingly, this newly identified PB2-282 residue within an evolutionary conserved “S-E-S” motif, present across different genera of influenza viruses and serving as a key regulator of RNA synthesis activity of the polymerase. In contrast, bat influenza viruses harbor an atypical “S-S-T” motif at the same position of PB2, alteration of which with the human-like “S-E-T” motif significantly enhances its (H17N10/Guatemala/164/2009) polymerase activity in human cells. Together, our data indicate that the PB2-S282 residue may serve as an inherent restriction element of the bat virus polymerase, limiting its activity in other host species.

IMPORTANCE Influenza A viruses are known for their ability to perform cross-species transmission, facilitated by amino acid alterations either in the surface antigen hemagglutinin (HA) or in the polymerase subunit PB2. Recent isolation of influenza A-like viruses from bats raised concern about their epizootic and zoonotic potential. Here, we identify a novel species-specific signature present within the influenza virus polymerase that may serve as a key factor in adaptation of influenza viruses from bat to nonbat host species. The PB2-282 residue, which harbors a highly conserved glutamic acid for influenza viruses across all genera (A, B, C, and D), encompasses an atypical serine in the case of bat influenza viruses. Our data show that the human-adapted polymerase, harboring a bat-specific signature (PB2-S282,) performs poorly, while bat PB2 protein, harboring a human-specific signature (PB2-E282), shows increased fitness in human cells.

INTRODUCTION

Influenza A viruses (IAVs) are segmented negative-sense RNA viruses that cause infections in humans and a wide range of other animals, including, but not limited to, birds, pigs, dogs, cats, horses, and seals. Wild aquatic birds are the natural reservoirs for these viruses, which, through continuous adaptation through antigenic shift and antigenic drift, acquire the ability to infect new host species. Such adaptation oftentimes results in a new event of zoonotic infection leading to the expansion in the repertoire of human-infecting influenza A viruses (1). For successful zoonoses, influenza viruses have to conquer two barriers imposed by the host; the first is successful recognition of the cell surface receptors by viral hemagglutinin (HA) protein to mediate entry, and the second is species-specific adaptation of the viral RNA-dependent RNA polymerase (RdRp) (2, 3), majorly governed by its PB2 subunit. The PB2 protein of avian viruses contains a glutamic acid residue at the 627th position, while the human-adapted viruses harbor a lysine residue at the same position. The polymerase-containing avian-type signature residue in PB2 is fully functional in avian cells but shows severe attenuation in human cells (4, 5). Other mutations within the 627 domain of PB2 have also been reported to act as species-specific adaptation sites that help the avian viruses to establish successful infection in humans/mammals (6).

During 2012 and 2013, two new influenza A virus genomes were isolated from the little yellow-shouldered bat (Sturnira lilium) in Guatemala and the flat-faced fruit-eating bat (Artibeus planirostris) in Peru, respectively (7, 8). According to the phylogenetic analyses, these two viruses show significant divergence in their genetic architecture from classical influenza A viruses, classified as two new subtypes, H17N10 and H18N11. The H17 and H18 proteins do not recognize the canonical sialic acid receptors (8, 9), while the N10 and N11 proteins lack sialidase activity (10), suggesting that bat influenza viruses utilize a distinct mechanism for entry into the host cells compared to the classical influenza A viruses. Recent advancements with bat influenza viruses further substantiated this notion through identification of MHC class II (MHCII) as bona fide receptors for bat influenza viruses (11, 12). This also has raised the concerns of epizootic and zoonotic transmission of these viruses given the constitutive expression of MHCII molecules in a wide variety of tissues in human, pigs, chicken, and a wide range of other animals (12).

Polymerases from human or avian origin influenza viruses have been shown to quickly adapt in bat cell lines (13). However, adaptability of the bat influenza virus polymerase in avian, human, or other mammalian hosts is yet to be investigated. Chimeric bat viruses that express surface antigens (HA and neuraminidase [NA]) of classical influenza viruses (influenza A/H1N1/PR8 or A/H7N7/SC35M) but internal proteins from the bat influenza virus (H17N10/Guatemala/164/2009) replicate poorly in cell culture and infected animals compared to the H1N1 (PR8) or H7N7 (SC35M) viruses (14, 15). Additionally, recombinant bat virus polymerase reconstituted in human lung epithelial cells shows limited activity in a reporter-based polymerase activity assay (7). Together, this evidence suggests that bat influenza virus polymerases are severely restricted in nonbat hosts, possibly due to the higher genetic diversity of the internal genes that accommodate unique signature residues in the individual subunit proteins PB2 (polymerase basic protein 2), PB1 (polymerase basic protein 1) and PA (polymerase acidic protein) (16). In this regard, it should be noted that the bat virus PB2 protein harbors an atypical serine residue at the 627th position instead of highly conserved glutamic acid as in the case of avian viruses or lysine in the case of most of the human-adapted viruses.

In this work, we characterized unique molecular signatures present in the bat influenza virus PB2 protein that could be responsible for the compromised fitness of the polymerase in nonbat hosts. Considering the PB2-S627 as signature for the bat virus polymerase, we tried to identify more such positions in the PB2 protein, which harbors either serine or threonine in the bat virus but contains highly conserved glutamic acid or aspartic acid in the nonbat IAVs. We have identified the serine at the 282nd position as a key signature residue, responsible for restricted activity of the bat virus polymerase, in comparison to the human- or avian-adapted polymerases harboring a glutamic acid residue at the same position. Replacing the glutamic acid either with alanine, lysine, or bat-specific serine residue severely affects the RNA synthesis activity of the human virus polymerase when reconstituted through transient transfection. Furthermore, recombinant human viruses with alanine or bat-specific serine at the 282nd position are severely restricted in cell lines derived from human, bat, and canine, establishing the importance of the glutamic acid at this position of influenza virus polymerases. This is further supported by the fact that partially humanized H17N10 PB2, harboring glutamic acid at the 282nd position, boosts polymerase activity several-fold compared to its wild-type version harboring serine at the same position. Interestingly, the E282 resides within a “S279-E282-S286” motif, which is conserved across different genera of influenza viruses and is indispensable for RNA synthesis activity of the polymerase. In contrast, bat influenza viruses harbor an atypical S-S-T motif at the same position of PB2, which is restrictive toward the activity of the bat virus polymerase. Together, our work identified a key molecular signature in bat influenza virus PB2 protein, which is responsible for its restricted activity in nonbat host species and may serve as a potential site of adaptation in order to acquire improved fitness in other hosts, including human.

RESULTS

Identification of the bat-specific signature residues that regulate the activity of viral RNA polymerase in nonbat hosts.

In order to identify the species-specific signature residues in the bat virus polymerase, we focused upon the PB2 subunit due to its critical role in determining the species specificity of influenza A viruses (4, 5). Throughout this study, the influenza A/H1N1/WSN/1933 strain was used as a model for the human-adapted virus, while the influenza A/H17N10/Guatemala/60 strain was used as a representative for the bat virus. First, we replaced the human-specific lysine at the 627th position with the bat-specific serine residue in order to ascertain its impact upon the activity of the human virus polymerase in human (HEK293T) and avian (DF1) cells. Luciferase-based reporter ribonucleoprotein complexes (RNPs) were reconstituted through transient transfection of a genome sense reporter RNA template, polymerase subunits (PB1, PB2, and PA), and nucleoprotein (NP)-expressing plasmids derived from the A/H1N1/WSN/1933 strain as described previously (6, 17, 18). Different variants of PB2 harboring the human signature lysine (K), avian signature glutamic acid (E), or bat signature serine (S) residues at the 627th position were used to reconstitute the polymerase (Fig. 1A). The polymerase with PB2-627K shows high reporter activity in both cell lines, while the polymerase with PB2-627E was selectively attenuated in human cells as expected (4, 5, 19), hence validating the efficiency of our luciferase-based assay system in determining the species-specific fitness of the polymerase (Fig. 1B). Interestingly, the polymerase with PB2-627S still remained attenuated, although to a lesser extent than PB2-627E, in human cells but was fully functional in chicken cells (Fig. 1B). Different PB2 variants show comparable expression as evidenced from the Western blot analysis (Fig. 1B). These data suggest that the PB2-627S may serve as one of the key determinants for species-specific restriction of the bat virus polymerase in nonbat host species, specifically in human.

FIG 1.

Bat-specific serine at the PB2-627 position restricts the polymerase in human but not in avian cells. (A) Influenza A virus RNPs are the major determinant of its host-specific replication fitness. The 627th amino acid of the PB2 subunit of polymerase harbors a lysine (K) or a glutamic acid (E) residue for human- or avian-adapted viruses, respectively. Bat influenza viruses harbor a serine (S) at the same position. (B) Luciferase-based reporter assay was performed to assess the polymerase function in human and chicken cells. Viral RNP, with the genetic background of A/WSN/1933, was reconstituted in HEK293T or DF1 cells either with wild-type PB2 containing a lysine or mutant PB2 containing glutamic acid or serine residue at the 627th position. n = 3 ± standard deviation. *, P < 0.05; one-way analysis of variance (ANOVA) compared to PB2-627K.

Encouraged by this finding, we focused our study on identification of additional signature residues present in the bat influenza virus PB2 protein which may regulate the activity of viral RNA polymerase in a species-specific manner. Bioinformatics analysis revealed that bat influenza virus PB2 protein harbors a higher number of solvent-accessible serine residues than the classical influenza A virus PB2 protein, hence suggesting that the 627th serine may not be an isolated one in controlling its activity. Using the PB2-627S as a signature for bat influenza virus polymerase, we tried to identify specific positions throughout the primary sequence of PB2 protein, which harbors either serine or threonine in the bat virus polymerase but represents highly conserved glutamic acid or aspartic acid in the conventional influenza viruses of either human or avian origin. Through primary sequence alignment of the bat and not bat influenza A virus PB2 sequences, we have been able to identify a total of seven such amino acid residues (60, 282, 390, 472, 671, 678, and 681), which were then further screened based upon their solvent accessibility in the atomic structure of the polymerase (20, 21). Finally, we focused upon five amino acid residues, 282, 390, 472, 678, and 681, that are completely surface exposed (Fig. 2A and B; Table 1) in the heterotrimeric polymerase (in analogy with the PB2-627 position) and harbor serine or threonine in bat virus PB2 protein but are occupied by highly conserved aspartic acid (390, 678) or glutamic acid (282, 472, 681) in the nonbat IAVs (Table 1 and Fig. 1A). Residue numbers 60 and 671 were not included in the study, as they are not completely surface exposed in the heterotrimeric polymerase but, rather, reside within the PB2-PB1 and PB2-PB1-PA interaction interfaces, respectively. We speculated that some of these shortlisted residues may serve as a determining factor in regulating the bat virus polymerase activity in bat and (or) nonbat host species.

FIG 2.

Identification of species-specific signature residues in the PB2 subunit of bat influenza virus polymerase. (A) Residue-wise solvent-accessible surface area (SASA in Ų) of H17N10 monomeric PB2 and trimeric polymerase was calculated. Serine and threonine residues are plotted as blue dots on the basis of their exposed surface area in monomeric and trimeric forms. Serine or threonine residues in bat virus PB2 protein, which are a glutamic acid or aspartic acid residue in conventional influenza viruses, are marked as red and yellow dots, respectively. Logo plots of the shortlisted amino acid residues (marked with asterisks), which represent highly conserved glutamic/aspartic acid in the PB2 protein of influenza A viruses of human (sequences analyzed, 36,243) or avian (sequences analyzed, 19,762) origin but harbor serine in the PB2 of bat viruses (sequences analyzed, 7). (B) Spatial organization of the surface-exposed glutamic/aspartic acid residues in the heterotrimeric polymerase structure of a human-adapted influenza A virus (A/NT/60/1968; PDB ID 6RR7) that are shortlisted for functional screening. (C) Luciferase-based reporter assay was performed to assess the polymerase function of the recombinant RNA-dependent RNA polymerase with wild-type or alanine substitution mutants of PB2 proteins using the genetic background of A/WSN/1933 strain. n = 3 ± standard deviation. *, P < 0.05; one-way ANOVA compared to wild-type PB2.

TABLE 1.

Analysis of surface-exposed serine and threonine residues in H17N10 and H3N2 PB2 from the crystal structure of RNA-dependent RNA polymerase (PDB IDs 4WSB and 6RR7)

Criteria	Amino acid (no. of residues)	Positions
Total no. of surface-exposed serine and threonine residues in H17N10 PB2 (PDB ID 4WSB)	Serine (34)	14, 61, 92, 156, 178, 179, 225, 226, 279, 282, 288, 290, 291, 300, 322, 324, 366, 371, 442, 443, 481, 514, 533, 534, 544, 593, 622, 674, 678, 683, 684, 685, 688, 741
	Threonine (34)	6, 79, 129, 155, 184, 186, 224, 227, 235, 245, 286, 287, 303, 340, 346, 351, 353, 355, 390, 400, 444, 451, 468, 472, 549, 562, 566, 569, 582, 596, 598, 609, 631, 681
Total no. of surface-exposed serine and threonine residues in H3N2-PB2 (PDB ID 6RR7)	Serine (20)	14, 92, 179, 225, 226, 279, 286, 322, 324, 481, 514, 533, 534, 544, 593, 684, 688, 79, 155, 286, 582
	Threonine (20)	178, 371, 674, 683, 76, 129, 184, 186, 224, 235, 245, 287, 303, 346, 351, 468, 549, 569, 598, 609
Surface-exposed serine and threonine residue in H17N10-PB2 changed to other residues in classical influenza virus PB2	Serine (17)	61, 156, 178, 282, 288, 290, 291, 300, 366, 371, 442, 443, 622, 674, 678, 683, 685
	Threonine (18)	76, 155, 227, 286, 340, 353, 355, 390, 400, 444, 451, 472, 562, 566, 582, 596, 631, 681
Surface-exposed serine and threonine residue of H17N10-PB2 changed to aspartic and glutamic residues in classical influenza virus PB2	Aspartic acid (2)	390, 678
	Glutamic acid (3)	282, 472, 681

In order to test this, we first investigated the functional significance of the identified amino acid residues in the context of the human-adapted polymerase. For this purpose, we have substituted each of the identified aspartic acid or glutamic acid residues with alanine in order to generate a panel of the mutant PB2 proteins. Luciferase-based reporter RNPs were reconstituted either with wild-type or with mutant PB2 proteins along with other RNP components as described earlier. Polymerase reconstituted with the wild-type PB2 protein showed a high level of reporter activity in HEK293T cells, while mutant PB2 proteins supported polymerase activity to various extents (Fig. 2C). The PB2 D472A and E678A mutants supported polymerase activity comparable to the wild-type protein, suggesting that these residues are not critical for viral RNA synthesis. The D390A and E681A mutants showed only modest defects reducing polymerase activity only by 40%. Interestingly, the E282A mutation severely attenuated the polymerase, supporting only 20% reporter activity compared to the wild-type PB2. These data suggest that the glutamic acid at the 282nd position of PB2 is critical in supporting activity of the human-adapted polymerase in human cells. Western blot analysis of the E282A and other mutant PB2 proteins showed expression levels comparable to the wild-type protein, hence confirming that the defect in RNA synthesis results from the suboptimal activity of these proteins and not from their altered abundance in cells.

Recombinant influenza A/H1N1/WSN viruses harboring bat-specific signature residues at the PB2-627 and PB2-282 positions show reduced fitness in bat and nonbat host species.

Highly conserved PB2-K627 and PB2-E282 residues are functionally important in the context of human virus polymerase, suggesting that the bat-specific serine residues at corresponding positions may alter the polymerase activity and hence interfere with virus propagation in a species-specific manner. Hence, we investigated the impact of the bat-specific serine residues at the 627th and 282nd positions of PB2 in the context of the propagation of human-adapted viruses. The genetic background of the influenza A/H1N1/WSN/1933 strains was used, and K627S, E282A, and E282S mutations were introduced in the PB2 open reading frame (ORF) in order to generate recombinant viruses using the plasmid-based reverse genetics as described earlier (19). Initially rescued viruses were amplified in Madin-Darby canine kidney (MDCK) cells for three consecutive passages and then titrated using plaque assay and finally confirmed the presence of the mutations in the PB2 ORF through Sanger sequencing. Both PB2-282A and PB2-282S viruses showed plaques that are significantly smaller than the wild-type virus. The PB2-627S virus shows plaque sizes relatively larger than two other mutant viruses but smaller than the wild-type virus, indicating that the mutant viruses might have an inherent defect in their replication ability (Fig. 3A and B). To test this, multicycle replication kinetics of wild-type and mutant viruses were performed in MDCK cells. The wild-type virus replicated to high levels, reaching titers more than 10⁷ PFU/mL by 48 h postinfection, while the PB2-K627S virus was severely attenuated, showing up to a 1,000-fold reduction in viral titer throughout the time course (Fig. 3D). This clearly reflects that introduction of the bat-specific residue in the human polymerase severely restricts its activity in cells of mammalian origin, which is consistent with the results demonstrated with other strains of influenza A/H1N1 virus (22). Introduction of either alanine or the bat-specific serine residue at the 282 position showed similar attenuation with a 100-fold decrease in the titer with respect to wild-type virus (Fig. 3D). Interestingly, the PB2-E282S virus showed higher replication fitness than the PB2-K627S virus in MDCK cells.

FIG 3.

Conventional influenza A viruses harboring bat-specific signature serine residues are defective in bat and nonbat host species. Recombinant influenza A/H1N1/WSN/1933 was generated containing a serine or alanine residue at the 282nd position and serine at the 627th position of PB2. (A) Plaque morphology of recombinant viruses containing the mutation. (B) Mean plaque diameter of recombinant viruses with standard error. (C) Agarose gel electropherogram of multisegment reverse transcriptase PCR (RT-PCR) performed with different mutant viral RNA. (D to F) Madin-Darby canine kidney (MDCK) (D), human lung epithelial carcinoma (A549) (E), and Pteropus alecto kidney (PaKiT03) (F) cells were infected with recombinant viruses at an MOI of 0.01 in three biological replicates. Supernatants were collected at 8, 16, 24, 48, and 72 h postinfection. Viral titer was calculated for each time point by performing plaque assay.

Next, to have a better insight into the species-specific fitness of human influenza virus that could be modulated by the bat-specific serine residues at the PB2-627 and PB2-282 positions, replicability of the mutant viruses was evaluated in cell lines of human and bat origin. We have used human lung epithelial carcinoma cells (A549) and Pteropus alecto kidney (PaKiT03) cells as model cell lines of the respective host species. As described earlier, the PaKiT03 cells support influenza virus infection and reassortment (23), which makes them suitable for monitoring replication of the human viruses harboring bat-specific signatures in their replication machinery. In fact, both wild-type and mutant viruses replicated to higher extents in PaKiT03 cells than A549 or MDCK cells, showing high susceptibility and permissiveness of these cells toward H1N1 virus infection (Fig. 3F). In A549 cells, replication efficiency of the wild-type and mutant viruses showed trends similar to what was observed in MDCK cells. The PB2-K627S showed up to 100-fold attenuation, followed by the PB2-E282S showing around 10-fold reduction in viral titers compared to the wild-type virus (Fig. 3E). Interestingly, both mutant viruses showed similar replication fitness in PaKiT03 cells, with more than 100-fold reduction in virus titer compared to the wild-type one (Fig. 3F).

We performed whole-genome sequencing to identify any additional mutations in the viral genome that may have been generated during the serial passaging of the PB2 mutant viruses. Viral RNA was isolated from PB2-E282S and PB2-E282A viruses, and a multisegment reverse transcription and PCR amplification were performed to amplify all eight different segments of the genomic RNA as originally described by Zhou et al. (24) (Fig. 3C). Subsequently, these amplicons were sequenced using the Illumina NovaSeq 6000 platform, which resulted in an average coverage of 96.88% for each of the three viruses (wild type and two mutants) (Tables 1 and 2). While both the E282S and the E282A mutations were accurately identified by major genome sequence alignment of the wild-type, mutants, and reference influenza A virus sequences available in the database, no additional mutation in any of the viral genome segments was detected, suggesting that the attenuation of the mutant viruses are solely due to the mutations introduced into their reverse genetics plasmids.

TABLE 2.

Results of next-generation sequencing

Type	Position	Amino acid change	Gene	Codon no.	Change	Codon change	Isolate(s)
Polymorphism	1469-1471	E282S	PB2	282	CTC→GCT	GAG→AGC	E282S
Polymorphism	1469-1470	E282A	PB2	282	CT→GG	GAG→GCC	E282A
Polymorphism	2338		PB2		C→T		Wild-type, E282S, E282A

Together, our data suggest introduction of the bat-specific serine residues at the critical 627th and 282nd positions of PB2 can significantly restrict the replication of the prototypic WSN strain of human influenza virus in both bat and nonbat host species. Based upon the sequence and structural conservation of the above-mentioned residues, it is likely that similar phenotypes might be observed for other conventional influenza viruses as well.

A highly conserved S279-E282-S286 motif within the mid-link regions of the PB2 protein is crucial for optimum fitness of the polymerase in human and avian cells.

The E282 residue resides within the “mid-” domain of PB2, which connects the N-terminal “N2” subdomain and middle “cap binding domain.” As revealed by different atomic structures of the polymerase, the mid-domain, along with the “627 linker,” together form the stalk upon which the cap binding domain and the 627 domain can undergo structural reorganization in order to transition between the transcriptionally active and replicative forms (20, 21). The mid-domain consists of four α-helices (α14 to α17), two of which face toward the core of the polymerase, while the other two are solvent exposed. The E282 resides on the second turn of the α15 helix flanked by two other serine residues in the adjacent turns, all three facing toward the solvent (Fig. 4A). Due to this structural feature, the three residues S279, E282, and S286 constitute a unique motif which is completely conserved across the PB2 sequences derived from the conventional influenza A viruses (Fig. 2A). Furthermore, structural and sequence alignment studies revealed that this S₂₇₉-E₂₈₂-S₂₈₆ motif is also conserved in influenza C viruses and partially conserved in influenza B and D viruses (Fig. 4A). Interestingly, the bat influenza A virus PB2 harbors a serine at the 282nd position and a threonine at 286, which results in an altered S₂₇₉-S₂₈₂-T₂₈₆ motif with the same structural feature (Fig. 4A).

FIG 4.

Surface-exposed, highly conserved S-E-S motif is crucial for the activity of viral RNA polymerase. (A) The PB2 282nd residue is present in surface-exposed α15 helix (highlighted in red color) of the mid-domain (PDB ID 4WSB). Multiple-sequence alignment of influenza A/H17N10 (A/H17N10/Guatemala/060), influenza A/H1N1 (A/H1N1/WSN/1933), influenza C (C/Aichi/1/81), influenza D (D/Quebec/1 M-H/2019), and influenza B (B/Brisbane/163/2008) showing the variable degrees of conservation of the S-E-S motif. Structural alignment of α15 helix of different influenza virus PB2s. (B) Luciferase-based reporter assay was performed to assess the RNA-dependent RNA polymerase function (using the genetic background of influenza A/H1N1/WSN/1933 strain) in human cells either containing conserved S-E-S, S-S-S, or S-S-T motifs. (C) Luciferase-based reporter assay was performed to assess the RNA-dependent RNA polymerase function (using the genetic background of influenza A/H1N1/WSN/1933 strain) in human cells and chicken cells reconstituted by wild-type or mutant PB2 protein. n = 3 ± standard deviation. *, P < 0.05; one-way ANOVA compared to wild-type PB2. (D) Luciferase-based reporter assay was performed to assess the polymerase function of influenza B/Brisbane/2008 RNA-dependent RNA polymerase containing wild-type or mutant PB2 protein. n = 3 ± standard deviation.*, P < 0.05; one-way ANOVA compared to wild-type PB2.

Based upon the importance of the E282 position in supporting viral RNA synthesis and virus replication and the high conservation of the S-E-S motif across different influenza virus PB2 proteins, we next evaluated the importance of this motif in supporting polymerase activity in HEK293T and DF1 cells. For this purpose, we generated a series of mutant PB2 proteins where individual amino acids of the S-E-S motif were substituted either with alanine, serine, or lysine and tested their ability to support viral RNA synthesis using the reporter-based polymerase activity assay (Fig. 4C). Substitution of both of the peripheral serine residues with alanine resulted in a 40% decrease in reporter activity, while substitution of the central glutamic acid showed around 70 to 80% decrease (Fig. 4C), as also evidenced earlier (Fig. 2C). A triple alanine mutant substituting the entire S-E-S motif resulted in complete abrogation of RNA synthesis, pointing toward the importance of the entire motif in supporting the activity of viral RNA polymerase. It is interesting to note that either introduction of bat-specific serine residue only at the 282nd position, thereby reconstitution of the bat-like “S-S-S” motif, or complete replacement of the human-specific S-E-S with bat-specific “S-S-T” motif resulted in around 60 to 70% decrease in polymerase activity (Fig. 4B), hence emphasizing the importance of the negatively charged glutamic acid residue at the center of the motif (Fig. 4C). This is also substantiated by the fact that substitution of the negatively charged glutamic acid with positively charged lysine results in complete abrogation of the polymerase activity. All of the mutant proteins show expression and stability comparable with the wild-type PB2 protein as evidenced from the Western blot analysis (Fig. 4C). Together, our data revealed the existence of a highly conserved S-E-S motif present in the mid domain of influenza A virus PB2 protein, which is crucial in supporting the polymerase activity both in avian and human cell lines.

Structural and sequence comparison between influenza A and B virus PB2 proteins suggests that the B-PB2 protein lacks the canonical S279-E282-S286 motif in spite of having significant sequence similarity within the mid-domain (Fig. 4B). Instead, it contains two glutamic acid residues in the two subsequent turns of the alpha-helix, hence constituting the “E280-E284” motif where the E-284 of B-PB2 perfectly aligns with the E282 of A-PB2 protein (Fig. 4A). To investigate whether the conserved glutamic acid at 284 is important for influenza B polymerase activity, we employed our recently developed firefly luciferase-based influenza B virus (strain B/Brisbane/60/2008) polymerase activity assay (25). As shown in Fig. 4D, influenza B polymerase with wild-type PB2 supported high levels of reporter activity, while introduction of the E284A mutation in B-PB2 reduced the reporter activity by 84%, suggesting an indispensable role of this glutamic acid residue in supporting polymerase activity. These data further substantiate the importance of this highly conserved glutamic acid residue in the PB2 protein of influenza viruses across different genera. This also points toward the fact that the absence of such an important molecular signature may significantly impact efficiency of the bat influenza virus PB2 protein, thereby restricting the overall activity of the bat virus polymerase in comparison to conventional influenza A or influenza B viruses.

Alteration of the S-E-S motif specifically restricts the RNA synthesis ability of the influenza virus polymerase.

The importance of the PB2 S-E-S motif, specifically the E282 residue, in supporting influenza A virus polymerase activity motivated a detailed investigation of its precise role in viral RNA synthesis. Influenza viruses assemble their RNA synthesis machinery in the form of viral ribonucleoprotein complexes (RNPs), where the heterotrimeric RNA polymerase (PB1, PB2, and PA) recruits multiple copies of NP on the nascent RNA strand during its synthesis, resulting in the formation of a macromolecular RNA-protein complex of megadalton range. Hence, fruitful assembly of viral RNPs is a prerequisite for subsequent rounds of RNA synthesis. To investigate the molecular mechanism by which mutations introduced in the S-E-S motif altered viral RNA synthesis, we assessed their ability to support viral RNP assembly using an RNP reconstitution assay. Authentic viral RNPs are reconstituted in HEK293T cells by expressing the negative-sense genomic RNA template, NP, PB1, PA, and wild-type or mutant variants of Flag-tagged PB2 proteins. The efficiency of RNP formation was determined by immunoprecipitating the viral polymerase via PB2-FLAG and detecting coprecipitated NP as a part of the RNP complex by Western blotting. A strong signal of NP reflects coprecipitation of a large number of NP molecules as part of the RNP complex, while a faint signal represents limited coprecipitation as a result of direct PB2-NP interaction. Densitometric analysis of NP (coimmunoprecipitation [co-IP]) and PB2 (IP) band intensities from three independent experiments were used to quantitatively estimate the extent of RNP formation. As evidenced in Fig. 5A and B, wild-type PB2 efficiently copurified large portions of NP, indicating efficient RNP formation. The A-A-A and the S-K-S mutants completely abolished the RNP formation, while the S-A-S mutant also showed ∼70% reduction compared to the wild-type PB2. A-E-A and the S-S-S mutants showed an intermediate phenotype, with ∼45% RNP reconstitution efficiency compared to the wild-type protein. Together, the ability of the PB2 mutants to support RNP assembly perfectly corroborates with their ability to support viral RNA synthesis as evidenced in the reporter activity assay. In this regard, it should be remembered that any defect in RNP assembly could be rooted either in an inherent defect in polymerase activity or in the defective binary interactions between PB2 with PB1and NP.

FIG 5.

Alteration of the S-E-S motif affects RNP reconstitution without impacting interaction with other viral proteins NP and PB1. (A) RNP reconstitution was performed by transfecting NP, PB1, PA, and PB2 proteins and vRNA expression plasmid in HEK293T cells. Forty-eight hours posttransfection, cells were lysed, and PB2 protein was immunoprecipitated and blotted for coprecipitated NP. (B) Densitometric analysis of coprecipitated NP signal normalized with immunoprecipitated PB2 signal from three independent RNP reconstitution experiments. n = 3 ± SD. (C) PB2 and NP interaction was checked by transfecting NP expression plasmid along with wild-type or mutant PB2 expression plasmids in HEK293T cells. Forty-eight hours posttransfection, cells were lysed, and PB2 was immunoprecipitated and blotted for coprecipitated NP. (D) PB2 and PB1 interaction was checked by transfecting PB1 expression plasmid along with wild-type or mutant PB2 expression plasmids in HEK293T cells. Forty-eight hours posttransfection, cells were lysed, and PB2 was immunoprecipitated and blotted for coprecipitated PB1.

Next, we tested the ability of the mutant PB2 proteins to participate in these protein-protein interactions. Flag-tagged wild-type or mutant PB2 proteins were coexpressed either with NP or with PB1 in HEK293T cells, and the lysates containing the binding partners were subjected to PB2-FLAG immunoprecipitation followed by Western blotting to detect the coprecipitation. To eliminate any nonspecific complex formation with cellular RNA, lysates were treated with RNase A before immunoprecipitation. All of the PB2 mutants showed NP and PB1 coprecipitation comparable to the wild-type protein, suggesting mutations in the S-E-S motif do not impact the ability of PB2 to interact with other RNP-associated proteins (Fig. 5C and D). We have also evaluated the nuclear localization ability of PB2 mutants by overexpressing them in A549 cells followed by indirect immunofluorescence assay. PB2 contains a nuclear localization signal which directs its importin alpha-mediated nuclear import in order to participate in RNP formation (26, 27). All of the mutant proteins showed preferential nuclear accumulation comparable to the wild-type PB2 when expressed through transient transfection in A549 cells (Fig. 6). Based upon these data, it can be inferred that neither impaired protein-protein interaction nor alteration of the subcellular localization was the key for the altered RNP formation ability as evidenced in case of the PB2 mutants.

FIG 6.

Motif mutants show similar subcellular localization to the wild-type PB2 protein. Plasmids expressing wild type or mutant PB2 proteins were transfected in A549 cells. Twenty-four hours posttransfection, cells were fixed and permeabilized. Cells were first stained with anti-PB2 mouse antibody followed by Alexa Fluor 488 anti-mouse rabbit antibody. Nuclei were stained with DAPI. Images were taken with a Leica fluorescence microscope.

Synthesis of viral genomic/antigenomic RNA and their assembly into progeny viral RNPs take place in a concomitant fashion. Hence, any defect in viral RNA synthesis would get manifested in the form of defective RNP assembly and vice versa. To test whether PB2 mutants are actually defective in their ability to support viral RNA synthesis, we have performed a mini-viral RNA (vRNA) template-based primer extension assay where the RNA synthesis activity of the polymerase could be monitored independently of the RNP assembly process. The mini-vRNA templates are short RNAs (76 nucleotides long) containing viral untranslated regions (UTRs) that could be replicated and transcribed by the polymerase in the absence of NP (28). Hence, any defect in RNA synthesis could solely be attributed to the defect in polymerase activity and not to the defective RNP assembly process. Mini-vRNA template (NP-77) was expressed in HEK293T cells along with the polymerase constituted either with wild-type or mutant PB2 proteins, and a primer extension assay was performed in order to monitor vRNA (replication) and mRNA (transcription) synthesis in the absence of NP. As evidenced clearly, PB2 mutants show various degrees of defects in vRNA synthesis (Fig. 7A and B), which could perfectly be correlated with their ability to support polymerase activity (evidenced in the reporter assay) (Fig. 4C) and the RNP assembly process. Synthesis of mRNA was almost completely abrogated for all of the mutants (only the AEA mutant showed minimal activity) (Fig. 7A and B), which could be a consequence of the low abundance of the vRNA template due to a defect in the replication process. These data clearly elucidate an inherent defect in RNA synthesis activity of the polymerase harboring the PB2 mutant proteins with alteration in their S-E-S motif.

FIG 7.

The S-E-S motif is critical for supporting RNA synthesis activity of the polymerase. (A) Primer extension was performed with a 77-nucleotide-long mini-viral RNA template containing 3′ and 5′ UTRs of the NP gene. Trimeric polymerase was reconstituted in HEK293T cells with wild-type or mutant PB2. Total RNA was extracted by TRIzol method and primer extension performed with fluorescently labeled specific primer for vRNA, mRNA, and 5S rRNA. Reaction products were analyzed by urea-PAGE and imaged in Bio-Rad ChemiDoc. (B) Densitometric analysis of primer extension experiment from three independent experiments.

The host’s acidic nuclear protein (ANP32A) may not be associated with the 282nd residue-mediated restriction.

ANP32A protein is a major determinant of PB2-mediated adaptation of avian influenza viruses in mammalian hosts (29). It has been shown that ectopic expression of chicken ANP32A, but not its human ortholog, can rescue the restriction associated with avian virus polymerase harboring a glutamic acid at the 627th position, hence supporting the replication of both avian- and human-adapted influenza viruses in human cells (19, 29). In contrast, human ANP32A supported bat influenza virus (both H17N10 and N18N11) polymerase activity to higher extents than the chicken ANP32A (30). Hence, to investigate any possible correlation between ANP32A and the restriction associated with the PB2-E282S mutation, we have tested the activity of either wild type (WT) or PB2-K627S or PB2-E282S polymerases in the context of overexpression of either chicken or human ANP32A. As expected, overexpression of ChANP32A rescued the defect of the avian-type polymerase in HEK293T cells, while HuANP32 failed to do so (Fig. 8A). Interestingly, ChANP32A, but not HuANP32, also boosted the activity of the PB2-K627S polymerase to the wild-type levels, further substantiating the strong association of ANP32A protein with the PB2-627 phenotype. In contrast, neither human nor chicken ANP32A proteins could rescue the defect associated with the PB2-S282 polymerase, which still remained significantly attenuated in the context of overexpression of either of these proteins (Fig. 8A). These data suggest that the PB2-282 residue possibly regulates viral RNA synthesis in an ANP32A-independent manner. It is important to note that human ANP32A showed preferential enhancement of the wild-type polymerase with respect to the polymerase harboring bat-specific signature residues either at the 627th or 282nd positions (Fig. 8A), which may reflect the generic preference of the human host factor toward the polymerase of a human-adapted virus. On the other hand, human ANP32A and bat ANP32A and B proteins supported bat virus polymerase activities (H17N10 and H18N10) to similar extents (30), suggesting that they may not show differential effects toward the PB2-627- or -282-associated phenotypes.

FIG 8.

Overexpression of chicken ANP32A does not rescue the 282 phenotype, unlike 627 in human cells. (A) Luciferase-based reporter assay was performed by reconstituting the influenza A/H1N1/WSN/1933 RNP in HEK293T, either with overexpression of chicken ANP32A or human ANP32A. n = 3 ± standard deviation. *, P < 0.05; one-way ANOVA compared to wild-type control. (B, Top) Influenza C virus polymerase dimer in complex with human ANP32A (light yellow). (B, Bottom) Influenza C virus polymerase dimer in complex with chicken ANP32A (dark yellow). Conserved S-E-S motif is marked in red. The 649K (homologous to 627K of H1N1) residue is marked in blue. (C) Polymerase-polymerase dimer formation was checked by reconstitution of influenza virus polymerase with differential HA or FLAG epitope-tagged PB2. Immunoprecipitation was performed with anti-FLAG antibody, and Western blot analysis was performed for coprecipitated HA-tagged PB2.

During viral genome replication, heterotrimeric RdRp forms an asymmetric dimer through a series of contacts between the C-terminal domain of PA, N1 subdomain of PB2, thumb subdomain of PB1, and also through the intermediacy of the host ANP32A protein (Fig. 8B) (20, 31). Crystallographic studies and nuclear magnetic resonance (NMR) spectroscopy analysis have shown that the 627K resides in close proximity to the dimerization interface and also participates in interaction with the flexible C-terminal tail of both chicken or human ANP32A proteins (31, 32). Given that mutation of the S-E-S motif harboring the E282 residue severely impacted the genomic RNA replication but does not affect PB2-PB1 or PB2-NP interactions, we decided to test its effect upon homodimer formation ability of the polymerase by performing an in-cell polymerase-polymerase dimer formation experiment as described previously (33, 34). Briefly, differentially epitope-tagged heterotrimeric polymerases were reconstituted by transient transfection of plasmids expressing PB1, PA, and FLAG-tagged or HA-tagged PB2 proteins in HEK293T cells. This is followed by immunoprecipitation of FLAG-tagged polymerase and evaluating the coprecipitation efficiency of the HA-tagged polymerase at 48 h of posttransfection. As evidenced, polymerase harboring either wild-type or mutant FLAG-tagged PB2 proteins dimerize successfully with the polymerase harboring wild-type HA-tagged PB2 protein, thereby leading to coimmunoprecipitation of HA-tagged and FLAG-tagged PB2 together (Fig. 8C). Clearly, the RNA synthesis defect associated with the mutant PB2 proteins does not originate from defective dimerization of polymerase, nor can it be rescued through ectopic expression of ANP32A proteins, pointing toward a novel regulatory mechanism behind the PB2-282 residue-mediated host adaptation process. In this regard, it should be noted that the E282, unlike the 627K, resides away from the homodimerization interface, which excludes any possible involvement of this residue in dimer formation or participation in interaction with ANP32A proteins, as also suggested by our results.

Introduction of human virus-specific residues in the bat virus PB2 protein boosts the bat influenza virus RNP activity.

Investigation of the molecular mechanism revealed that the highly conserved S-E-S motif, specifically the E282 residue, in the mid-domain of PB2 is critical in supporting optimum activity of classical influenza A virus polymerase. Alteration of the S-E-S motif via introduction of the bat-specific serine residue at the 282nd position significantly restricts polymerase activity and hence inhibits virus propagation, not only in human but also in bat cells (Fig. 3F). Similar attenuation was also observed in the case of viruses harboring the bat-specific serine residue at the 627th position (Fig. 3F) (22). Based upon these observations, we hypothesize that the serine residues at critical positions of PB2 may serve as the key restriction elements present within the bat virus polymerase that are responsible for the lower replication fitness of bat influenza virus in comparison to the classical influenza A viruses.

To test our hypothesis, we reconstituted a chimeric polymerase composed of the PB1 and PA subunits of the A/H1N1/WSN/1933 strain and the PB2 subunit of the A/H17N10/Guatemala/060 strain. Influenza A/H1N1/WSN/1933 virus RNPs were reconstituted either with wild-type or chimeric polymerases in HEK293T cells as described earlier (13). As evidenced, introduction of the bat virus PB2 in the human virus RNP severely restricts the polymerase, resulting in around 2-log decrease in reporter activity (Fig. 9A). Although an increasing concentration of the H17N10 PB2 results in a dose-dependent increase in reporter activity, the chimeric polymerase still remained attenuated with respect to the wild type, suggesting the severe restriction imposed by the bat PB2 upon the other subunits of the polymerase. This is in corroboration with the results presented by other groups, where alteration of the individual RNP component of human-adapted influenza virus with that of the bat influenza virus restricted viral RNA synthesis (13–15). Subsequently, we have replaced the bat-specific serine residues at 627 and 282 positions with human signature lysine and glutamic acid residues in the H17N10 PB2 protein in order to generate either single mutants (S627K and S282E) or a double mutant (S672K/S282E). Chimeric polymerases and RNPs were reconstituted either with wild-type or mutant H17N10 PB2 proteins to measure their ability to support viral RNA synthesis. Interestingly, substituting the bat virus-specific serine residues with human virus-specific glutamic acid or lysine boosted the polymerase activity by 100% and 50%, respectively, thereby providing evidence that serine at the 282nd and 627th positions of the H17N10 PB2 protein may restrict the activity of bat influenza polymerases in human cells (Fig. 9B). The double mutant S282E/S627K, harboring both of the human virus-specific signature residues, showed minor enhancement compared to the single mutants, hence suggesting that these restriction factors in the polymerase do not act in a cumulative fashion. This also suggests that species tropism governed by PB2-627 residue is an independent event and is not mechanistically connected with the PB2-282 residue-mediated restriction of bat virus polymerase.

FIG 9.

Introduction of S627K and S282E mutations in H17N10 PB2 protein boosts activity of the polymerase. (A) Luciferase-based reporter assay was performed by reconstituting the influenza A/H1N1/WSN/1933 RNP in HEK293T cells, either with H1N1 PB2 or with different amounts of H17N10 PB2 expression plasmid. n = 3 ± standard deviation. *, P < 0.05; one-way ANOVA compared to the previous set. (B) Luciferase-based reporter assay was performed by reconstituting the influenza A/H1N1/WSN/1933 RNP with either wild-type H17N10 PB2 or mutant H17N10 PB2 protein. n = 3 ± standard deviation. *, P < 0.05; one-way ANOVA compared to PB2 wild-type or S282E or S627K. (C) H17N10 RNP was reconstituted in HEK293T cells with wild-type or mutant PB2 protein. n = 3 ± standard deviation. *, P < 0.05; one-way ANOVA compared to PB2 wild type. (D) H17N10 RNP was reconstituted in HEK293T cells with wild-type or mutant PB2 protein. Either human or chicken ANP32A was overexpressed with the RNPs in HEK293T cells. n = 3 ± standard deviation. *, P < 0.05 one-way ANOVA compared to PB2 wild type.

To investigate the importance of the PB2-282 position in the context of authentic bat virus polymerase, we have reconstituted authentic H17N10 reporter RNPs in HEK293T cells. For this purpose, we have constructed the plasmid expressing reporter genome where the firefly luciferase gene was inserted in between the 3′ and 5′ UTRs from segment 6 of the H17N10 genome as described in Materials and Methods. Subsequently, the reporter construct, along with the plasmids expressing PB1, PA, and PB2 proteins, was transiently transfected in HEK293T cells to reconstitute the bat influenza virus RNPs. While wild-type PB2 leads to successful RNP reconstitution and hence enhanced reporter activity compared to the no-PB2 control, RNPs reconstituted with PB2 proteins harboring mutations within the highly conserved S-S-T motif showed variable degrees of activity (Fig. 9C). A triple-alanine mutation (A-A-A) of the conserved S-S-T severely compromised polymerase activity with 72% reduction in reporter activity, thereby validating the importance of this motif for optimum activity of influenza virus polymerase. Interestingly, incorporation of a glutamic acid residue at the center of the triple alanine, hence reconstituting an “A-E-A” motif, resulted in reporter signal comparable to the wild-type protein (Fig. 9C). Clearly, a single glutamic acid can compensate for the deficiency of the S-S-T in the H17N10 PB2 protein in terms of supporting viral RNA synthesis. Finally, the replacement of the bat signature S-S-T motif with the human signature “S-E-T” motif led to an ∼75% increase in the activity of the H17N10 polymerase in human cells, hence establishing the importance of the E282 residue in influenza virus polymerase function and also in potential adaptation of bat virus polymerase in nonbat host species. Furthermore, complete H17N10 RNPs showed a similar trend with the chimeric H1N1 RNPs harboring H17N10 PB2, hence suggesting that other components of RNPs do not have any epistatic effect upon the PB2-282-associated phenotype.

Finally, we tested the effects ANP32A proteins upon bat influenza virus polymerases harboring either bat or human signature residues at position 282. Reporter H17N10 RNPs were reconstituted either with wild-type (S282) or mutant (E282) PB2 proteins in the context of overexpression of human or chicken ANP32A in HEK293T cells. As observed, H17N10 polymerase harboring the PB2-S282E mutation showed higher fitness than the wild-type polymerase, even in the context of overexpressed human or chicken ANP32A (Fig. 9D). These data, along with the results shown earlier (Fig. 8A), suggest that ANP32A proteins neither can rescue the defect of the H1N1 polymerase harboring bat-specific serine residue at PB2-282 nor can they subside the enhanced fitness of the bat virus polymerase harboring a human-specific glutamic acid at the same position.

Together, our data not only establish the importance of the PB2-282 residue in the RNA synthesis activity of the polymerase but also establish it as a novel species-specific signature that can dictate the fitness of the polymerase in different host species.

DISCUSSION

Higher genetic plasticity of the bat influenza viruses has been shown to facilitate their rapid adaptation in mice (16, 35), hence revealing their potential of cross-species transmission in nonbat host species. Interestingly, these adapted bat viruses replicate poorly in ferrets and in human-derived cell lines, which could be attributed to the suboptimal activity of the viral polymerase in nonbat host species (7, 16, 35). In this work, we have identified a unique molecular signature in the PB2 subunit of bat influenza virus polymerase, which is responsible for its restricted activity in human cell lines. Combined sequence and structural comparison between different influenza virus polymerases led us to the identification of a previously uncharacterized S-E-S motif in the PB2 subunit, which is completely conserved in classical influenza A viruses but is replaced with an S-S-T motif in bat influenza virus. Alteration of this canonical S-E-S motif with bat-like S-S-T significantly reduced the polymerase activity and hence the fitness of influenza A/H1N1/WSN virus. In corroboration with these results, replacing the bat-specific S-S-T with the semicanonical S-E-T motif results in significant enhancement in the activity of the A/H17N10/Guatemala virus polymerase and hence viral RNA synthesis in a reporter-based RNP activity assay. Clearly, this S-E-S motif, specifically the 282 E residues, serves as a key molecular determinant of the fitness of classical influenza A virus polymerase in a host-independent manner, and the serine at the 282nd position of the bat influenza virus PB2 may serve as an inherent restriction element, impeding its activity in human cells. It is interesting to note that the bat-specific serine residue at the PB2-627 position also imparts similar restriction but specifically in human cells and not in avian cells, data corroborating well with the previously published reports (22).

Our results presented strong evidence that the complete integrity of the S-E-S motif, specifically the negatively charged glutamic acid at the center of the motif, is critical for supporting RNA synthesis activity of the polymerase, although the precise mechanism is yet to be characterized. The PB2-282 resides within the mid-domain and remains completely surface exposed in the atomic structure of the polymerase, hence suggesting its possible involvement in interaction with host factors. However, our results showed that neither human nor chicken ANP32A proteins can alter the phenotype associated with the alteration of the PB2-282 residues in the context of H1N1 or H17N10 polymerases, thereby making the involvement of ANP32A less likely in PB2-282-mediated regulation of viral RNA synthesis. This is further supported by the atomic structure of the polymerase in complex with ANP32A protein, which shows that the PB2-282 residue positions far away from ANP32A interacting interface, hence possibly excluding its involvement in ANP32A-mediated regulation of viral RNA synthesis. In this regard, it is noteworthy that the 282 resides within close proximity to the cap binding domain (CBD) of PB2 and hence may participate either in structural reorganization of the CBD (36, 37) or even in direct interaction with the capped RNA primer as reported previously (38). Additionally, the M283 residue has been recognized as virulence determinant and human adaptation site for avian influenza viruses (39), further substantiating the importance of this region in influenza virus replication, host adaptation, and pathogenesis. A more detailed investigation would be likely to elucidate the precise molecular mechanism by which PB2-282 and other residues within the mid-link region play a role in influenza viral RNA synthesis and virus life cycle.

The evolutionary history of the bat influenza viruses and their ancestral relationship with other flu variants are not clear (16). However, the New World bat influenza viruses (H17N10 and H18N11 subtypes) have been designated distinctly divergent entities of IAVs (7, 8). As per the phylogenetic backdating, the New World bat IAVs possibly have separated from all of the classical lineages around 650 years ago (16). This is supported by the fact that the amino acid sequences of the internal segments of the New World bat viruses tolerate higher diversity in contrast to the classical IAVs and hence form a distinct outgroup (8). It is possible that the bat influenza viruses served as the ancestor for all other IAVs, which, during the course of adaptation into other host species, acquired specific mutations in different internal gene segments. The change of serine to glutamic acid at the 627th (avian) and 282nd positions of PB2 may thus have served as crucial adaptation sites of the virus from bat to nonbat hosts, which have conferred additional fitness to the viral RNA polymerase in terms of replication in corresponding host species. In another possible scenario, bat and nonbat IAVs may have segregated from a common ancestor (40), where the bat-specific serine residues may have been selected over glutamic acid in order to generate a slower-replication variant of the influenza virus, which might be suitable in terms of establishing host-parasite equilibrium in the bat species. It is interesting that the Old-World bat IAVs (H9N2), which show relatively lesser divergence from the classical variants (16), harbor glutamic acid at the 627th and 282nd positions of PB2 (41), further supporting the importance of these residues in adaptation of the viruses from bat to nonbat hosts or vice versa.

It is interesting to note that bat influenza virus polymerase, specifically the PB2 subunit, harbors a higher number of serine residues in the solvent-exposed surface in comparison to the conventional IAVs (Table 1). For example, all of the aspartic acid or glutamic acid residues identified in this study are serine or threonine in bat virus polymerase. It is tempting to speculate that phosphorylation of some of these serine residues, specifically at the 627th and 282nd positions, is important for proper functionality of the polymerase, and that is why these residues may get adapted to negatively charged glutamic acid or aspartic acid residues in classical IAVs so that the dependence over the host kinases could be bypassed. In the opposite sense, a glutamic acid-to-serine adaptation could also impart functional regulation to the bat virus polymerase, which could be controlled via phosphorylation by host kinases. This notion is indirectly supported by our data obtained from the H17N10 polymerase, where a single phosphomimetic glutamic acid residue at the 282nd position can compensate for the defect caused due to complete replacement of the S-S-T motif with triple alanine. It is likely that phosphorylation of one or more residues in the S-S-T motif is a prerequisite for supporting RNA synthesis activity of the bat influenza virus polymerase. A comparative phospho-proteome analysis of the bat and nonbat influenza virus polymerase in various host species may shed light upon this possibility of such naturally occurring phosphomimetic adaptation sites in the conventional influenza A viruses. From a different perspective, mutation of the 282nd serine residue to glutamic acid within the bat influenza virus polymerase significantly increases its replication fitness in human cells, which, together with the adaptive mutations in the HA and NA (35, 40), could lead to the spillover of the bat viruses in nonbat host species.

Together, this work leads to the identification of a species-specific signature amino acid residing within a highly conserved motif within the PB2 protein, which plays a crucial role in maintaining proper replication fitness of the influenza virus polymerase in a wide variety of host species, including human, chicken, canine, and bats. The extent of conservation of this E282 residue across different genera of influenza viruses and unusual occurrence of serine at the same of position of bat influenza virus PB2 protein presented a strong indication toward the potential role that this residue may play in adaptation of bat influenza virus polymerase toward nonpermissive or poorly permissive nonbat host species. Identification of more such adaptation sites would be helpful in assessing the zoonotic potential of bat influenza viruses and also help us understand the molecular mechanism of spillover of bat-borne viruses that remain a major concern for global human health and economy.

MATERIALS AND METHODS

Cell culture.

Human embryonic kidney (HEK293T), Madin-Darby canine kidney (MDCK), and Pteropus alecto kidney (PaKiT03) cells were maintained in Dulbecco’s modified Eagle medium (DMEM; Gibco) cell culture medium supplemented with 10% fetal bovine serum (FBS), 1× penicillin-streptomycin (PenStrep) solution (Gibco), and 1× GlutaMax (Gibco) at 37°C and 98% relative humidity in the presence of 5% CO₂. Human lung epithelial carcinoma(A549) cells were maintained in DMEM-F12 (Gibco) medium supplemented with 10% FBS, 1× penicillin-streptomycin solution, and 1× GlutaMax at 37°C and 98% relative humidity in the presence of 5% CO₂. DF-1 (chicken fibroblast) cells were maintained at 39°C and 98% relative humidity in the presence of 5% CO₂ in the same DMEM culture medium as HEK293T.

Plasmids and viruses.

All genes were derived from the influenza A (A/WSN/33) or influenza B (B/Brisbane/60/2008) viruses. Polymerase proteins and NP were expressed in cells from the plasmids, pCDNA3-PB2-3XFLAG (encoding a C-terminal FLAG tag), pCDNA3-PA, pCDNA3-PB1, pCDNA6.2-NP-V5 (encoding a C-terminal V5 tag for influenza A), and pcDNA3.1-NP for influenza B. The PB2 ORF of influenza A/H17N10/Guatemala/060 was cloned also under the cytomegalovirus (CMV) promoter in pCDNA3.1 vector with C-terminal 3× FLAG epitope tag. pHH21-vNA-luc reporter plasmid encodes firefly luciferase in the negative sense, flanked by UTRs from the NA gene derived from A/H1N1/WSN/1933, is a kind gift from Andrew Mehle. Influenza B reporter plasmid was constructed using UTRs derived from NA gene of Influenza B/Brisbane/60/2008 virus (25). For expression of luciferase reporter RNA in DF-1 cells, pgHH21-vNA-Luc plasmid was prepared by inserting the reporter cassette of pHH21-vNA-Luc under chicken polymerase I promoter. Viruses were generated using the bidirectional reverse genetics plasmid system. For expression of polymerase subunits, NP proteins and corresponding genomic RNA pBD plasmids were used. The pTMΔRNP plasmid (kind gift from Andrew Mehle) was transfected for providing the other genomic RNA segments. ANP32A expression plasmids were kindly provided by Andrew Mehle and Steven F. Baker.

H17N10 reporter plasmid construction.

The H17N10 reporter plasmid has been created by modifying the pHH21-vNA-Luc plasmid. The 3′ and 5′ UTRs of segment 6 of H1N1/WSN have been replaced by the same of H17N10/Guatemala/060. H17N10 5′-UTR forward (5′-ACTACTCGTCTCCATTCATTGTCCATTGGTCATTACACGGCGATCTTTCCGCCCTTC-3′) and H17N10 3′-UTR reverse (5′-ACTACTCGTCTCTTAATAATGGAAGACGCCAAAAACATAAAGAAAGGC-3′) primers were used to amplify the luciferase cassette flanked by the 5′ and 3′ UTRs. pHH21 vector was linearized with H17N10 vector forward (5′-ACTACTCGTCTCTATTAAAAACTCCTGCTTCTGCTCCCCCCC-3′) and H17N10 vector reverse (5′-ACTACTCGTCTCTGAATGAAAAAACTCCTTGTTTCTACTAATAACCCGGCGGC-3′) primers. Amplified PCR products were purified and digested with BsmBI-v2 (NEB) restriction enzymes to make compatible sticky ends. The digested luciferase cassette was dephosphorylated by calf intestinal alkaline phosphatase. Finally, the luciferase cassette and the digested vector were ligated with T4 DNA ligase.

Site-directed mutagenesis.

Site-directed mutagenesis primers were designed by QuikChange Primer Design software (Agilent Technologies). Mutation was performed by PCR amplification of the plasmids using PfuTurbo DNA polymerase enzyme (Agilent Technologies) followed by DpnI digestion and transformation in E. coli DH5α cells. Mutations were confirmed by Sanger sequencing.

Polymerase activity assays.

HEK293T or DF-1 cells were reverse transfected in triplicates using Lipofectamine 3000 with plasmids expressing PA, PB1, PB2-FLAG, NP-v5/NP, and vNA-Luc. Cells were harvested 36 h posttransfection, and polymerase activity was measured using the luciferase assay system (Promega) on GloMax20/20 luminometer (Promega). Equivalent PB2 and NP protein expression was confirmed by Western blotting. The H17N10 reporter assay was performed by transfecting the reporter construct (mentioned above) and PB2, PB1, PA, and NP protein expression plasmids (kindly provided by Andrew Mehle) into HEK293T cells. Cells were harvested at 18 h posttransfection.

Rescue of recombinant viruses and plaque assay.

Cocultures of HEK293T and MDBK cells were reverse transfected using Lipofectamine 3000 reagent (Invitrogen) with virus rescue plasmids pTMΔRNP, pBD_PB2, pBD*_PB1, pBD_PA, and pBD_NP. Twenty-four hours posttransfection (hpt), the culture medium was replaced with virus growth medium (1× DMEM, 1× PenStrep, 4 mM GlutaMax, 0.2% bovine serum albumin [BSA], 25 mM HEPES buffer, and 0.5 μg/mL L-1-tosylamido-2-phenylethyl chloromethyl ketone[TPCK] treated-trypsin). Supernatants were harvested 48 to 72 hpt, and the viruses were subsequently amplified in MDBK cells.

Plaque assay.

Plaque assay was performed using MDCK cells. We seeded 0.3 million cells in each well of a 12-well cell culture plate the day before the assay. Virus stocks were diluted in 10-fold series in virus growth media. Cells were infected by the dilutions for 50 min at 37°C with intermittent shaking every 10 min. After 50 min, the infected cells were overlaid with a 1:1 mixture of 2.4% Avicel and 2× DMEM (2× DMEM, 1× PenStrep, 4 mM GlutaMax, 0.4% BSA, 50 mM HEPES buffer, and 1.0 μg/mL TPCK-trypsin). Plates were harvested at 72 h postinfection, fixed with 70% ethanol, and stained with 1% crystal violet. Plaques were counted from the stained plates, and titer was calculated.

Multicycle replication kinetics.

A549, MDCK, and PakiT03 cell lines were infected with wild-type or mutant influenza A/H1N1/WSN/1933 viruses at a multiplicity of infection (MOI) of 0.01 in three biological replicates. Cell culture supernatants were harvested at 8, 16, 24, 48, and 72 h postinfection from the infected dish. Viral titer in the supernatant was quantitated by performing plaque assay in MDCK cells.

RNA isolation, reverse transcription, and PCR.

Supernatants from virus rescue experiments were harvested and clarified by centrifugation at 12,000 × g for 10 min, and viral RNA was extracted using TRIzol reagent (Invitrogen). Reverse transcription was performed by M-MLV RT PB2-specific primers. PCR amplification of cDNA was done by Phusion DNA polymerase with the PB2 gene-specific primer pair. PCR-amplified fragments were used for Sanger sequencing. Multisegment reverse transcription (RT) was performed using MBTuni12 primers (5′-ACGCGTGATCAGCAAAAGCAGG-3′ and 5′-ACGCGTGATCAGCGAAAGCAGG-3′ in 1:1 molar ratio) and SuperScript III reverse transcriptase following previous methods. PCR amplification was performed with MBTuni12 and MBTuni13 primers (5′-ACGCGTGATCAGTAGAAACAAGG-3′) using Q5 DNA polymerase along with GC enhancer. PCR products were purified and subjected to next-generation sequencing.

Next-generation sequencing and data analysis.

Equimolar quantities of gel-purified PCR products were pooled and subjected to paired-end whole-genome sequencing (2 × 250 bp) on an Illumina NovaSeq 6000 platform (Illumina, San Diego, USA). Raw sequence data generated were trimmed and filtered using BBDuk software from BBtools package (https://sourceforge.net/projects/bbmap/), and reference-guided assembly was performed using Bowtie2 (Table 3) (42). H1N1 reference sequences were obtained from GenBank (accession numbers LC333182.1, LC333183.1, LC333184.1, LC333185.1, LC333186.1, LC333187.1, LC333188.1, and LC333189.1). Major genome or consensus sequences were generated based on variants having a frequency of >0.5 compared to the reference sequences. Major genome sequence alignments were executed using MUSCLE version 3.8.425 with 8 iterations (43), and variants were identified.

TABLE 3.

Alignment details

Isolates	No. of paired-end reads	No. of filtered reads	No. of aligned reads (%)
WT-1	4,876,506	3,885,154	3,784,313 (97.41)
E282A	5,118,704	3,982,264	3,882,163 (97.49)
E282S	4,576,378	3,611,552	3,457,629 (95.74)

Primer extension.

PB1, PA, and wild-type or mutant PB2-encoding plasmids with a small 77-nucleotide-long viral RNA with 3′ and 5′ UTRs of NP (NP77) expression plasmid were transfected in HEK293T cells. Forty-eight hours posttransfection, the cells were harvested. Total RNA was isolated by TRIzol reagent (Invitrogen). Primer extension was performed by SuperScript III reverse transcriptase enzyme (Thermo) and fluorescence-labeled primers as described earlier (28). The reactions were separated in urea-PAGE and the gel imaged in Bio-Rad ChemiDoc.

Coimmunoprecipitations.

HEK293T cells were transfected for RNP reconstitution with expression plasmids encoding NP-V5, PB1, PA, and FLAG-tagged WT or mutant PB2, along with vRNA (NA segment). Cells were lysed 48 hpt using co-IP buffer (50 mM Tris-HCl [pH 7.4], 150 mM NaCl, 1 mM EDTA, 1% NP-40, 1% Na-deoxycholate, and 0.1% SDS) supplemented with 1× protease inhibitor cocktail (10 μL of 50× propidium iodide [PI] in 500 μL of co-IP buffer) and Halt phosphatase inhibitor (10 μL for 500 μL of lysis buffer) and incubated at 4°C for 20 min on a rocker. Lysates were clarified by centrifuging at 20,000 × g for 20 min at 4°C. Total protein samples were separated from the lysate. Lysates were supplemented with 0.5 mg/mL BSA and precleared with 20 μL of preequilibrated protein A agarose beads. After preclearing, the lysates were incubated overnight with the antibody. The next day, preequilibrated protein A magnetic beads were added to the lysate and incubated for 1 h. Protein A beads were then recovered using magnetic racks and washed thrice with co-IP buffer. Finally, sample beads were treated with 30 μL of 5× Laemmli buffer, heated at 98°C for 5 min, and centrifuged at 10,000 × g for 10 min. Samples were separated by SDS-PAGE and identified by Western blotting.

Western blotting.

Cell lysates were separated by SDS-PAGE and transferred to methanol-activated polyvinylidene difluoride (PVDF) membrane (Bio-Rad) using Trans-Blot Turbo transfer system. The membrane was blocked in 5% skimmed milk solution in 1× Tris-buffered saline with Tween 20 (TBST) at room temperature for 2 h. Primary antibody incubation was done at 4°C in rocking conditions overnight. The next day, after washing three times with 1× TBST, the membrane was incubated with horseradish peroxidase (HRP)-conjugated secondary antibody at room temperature for 1 h. Before developing, the blots were washed thrice with 1× TBST. Chemiluminescent substrate was applied over the blot and incubated for 3 to 5 min and developed in Bio-Rad ChemiDoc.

Immunofluorescence assay.

A549 cells grown on coverslips were transfected with plasmids encoding wild-type or mutant PB2 proteins. Twenty-four hours posttransfection, cells were washed in phosphate-buffered saline (PBS), fixed with 3% formaldehyde (20 min at room temperature), quenched with 0.1 M glycine, and permeabilized with 0.1% Triton X-100 in PBS (10 min at room temperature). After blocking with 3% BSA (20 min at room temperature), cells were incubated with primary antibody (anti-FLAG) for 1 h at room temperature and washed thrice. Secondary antibody (Alexa Fluor 488-conjugated donkey anti-mouse IgG) incubation was done for 40 min at room temperature. DAPI (4′,6-diamidino-2-phenylindole) staining was done along with the secondary antibody incubation. After secondary antibody incubation, cells were washed thrice with PBS and once with Nanopure water and mounted on slides with Fluoroshield (Sigma-Aldrich). Images were taken by a fluorescence microscope (Leica).

Bioinformatics and structural analysis.

Specific PB2 protein sequences were obtained from the Influenza Research Database (www.fludb.org) by selecting data type-protein, Virus type-A, and proteins-PB2 and separately selecting host-human (n = 36,243), avian (n = 19,762), and bat (n = 7). The aligned FASTA files were viewed and analyzed in BioEdit sequence alignment editor (44). Multiple-sequence alignment was performed in BioEdit software by using ClustalW followed by manual trimming. Logo plots were generated by WebLogo server (https://weblogo.berkeley.edu/logo.cgi) using the aligned FASTA files (http://www.ebi.ac.uk/pdbe/prot_int/pistart.html) (45). PDBePISA tool (46) was used to analyze the solvent-accessible surface area of PB2 proteins from the following PBD IDs: 4WSB, 4WSA, 6RR7, and 5D98 (20, 21, 47). Structural alignment of the PB2 protein from the trimeric polymerase crystal structure was performed in UCSF Chimera software.

Statistical analysis and data analysis.

Graph preparations and statistical analysis were done in Microsoft Excel software. Densitometric analysis was performed in Bio-Rad Image Lab software.

Data availability.

Sequences obtained from the analysis have been submitted to NCBI GenBank (accession numbers OL831132, OL831133, OL831134, OL831135, OL831136, OL831137, OL831138, OL831139, OL831157, OL831158, OL831159, OL831160, OL831161, OL831162, OL831163, OL831164, OL831179, OL831180, OL831181, OL831182, OL831183, OL831184, OL831185, and OL831186).