|Home | About | Journals | Submit | Contact Us | Français|
The direct infection of humans with highly pathogenic avian H5N1 influenza viruses has suggested viral mutation as one mechanism for the emergence of novel human influenza A viruses. Although the polymerase complex is known to be a key component in host adaptation, mutations that enhance the polymerase activity of avian viruses in mammalian hosts are not fully characterized. The genomic comparison of influenza A virus isolates has identified highly conserved residues in influenza proteins that are specific to either human or avian viruses, including 10 residues in PB2. We characterized the activity of avian polymerase complexes containing avian-to-human mutations at these conserved PB2 residues and found that, in addition to the E627K mutation, the PB2 mutation T271A enhances polymerase activity in human cells. We confirmed the effects of the T271A mutation using recombinant WSN viruses containing avian NP and polymerase genes with wild-type (WT) or mutant PB2. The 271A virus showed enhanced growth compared to that of the WT in mammalian cells in vitro. The 271A mutant did not increase viral pathogenicity significantly in mice compared to that of the 627K mutant, but it did enhance the lung virus titer. Also, cell infiltration was more evident in lungs of 271A-infected mice than in those of the WT. Interestingly, the avian-derived PB2 of the 2009 pandemic H1N1 influenza virus has 271A. The characterization of the polymerase activity of A/California/04/2009 (H1N1) and corresponding PB2 mutants indicates that the high polymerase activity of the pandemic strain in mammalian cells is, in part, dependent on 271A. Our results clearly indicate the contribution of PB2 amino acid 271 to enhanced polymerase activity and viral growth in mammalian hosts.
Influenza A viruses are negative-sense, single-stranded, segmented RNA viruses that infect a wide range of hosts, including humans and many avian species. In humans, influenza A viruses of limited subtypes cause seasonal epidemics as well as occasional pandemics affecting a disproportionately larger percentage of the population with increased morbidity and mortality (35, 49). The subtypes of currently circulating seasonal influenza A strains in humans are H1N1 and H3N2 (48). In contrast, all 16 HA and 9 NA subtypes are thought to circulate in wild avian species, with most causing little or no disease (49). One potential mechanism for the instigation of a pandemic is the direct infection of humans with an avian virus that contains mutations allowing it to easily infect and spread among humans. Before 1997, the direct infection of humans with avian influenza viruses was not considered a threat to human health, and there was little evidence in the literature for direct avian-to-human transmission (49). However, several cases of human infection with highly pathogenic avian H5N1 influenza viruses occurred in 1997 in Hong Kong with a 33% fatality rate (4, 44). In 2003, H5N1 viruses reemerged, and as of December 2009, there have been 445 cases of human infection, with a mortality rate of almost 60% (47). These cases of H5N1 infection in humans show that the direct avian-to-human transmission of influenza can occur, and furthermore, it can cause severe disease and mortality in the human host. While these infections have been limited mainly to persons in close contact with infected poultry and human-to-human transmission has been extremely rare (2, 46), the acquisition of additional mutations allowing the efficient human-to-human transmission of highly pathogenic avian viruses could lead to a severe pandemic.
Many viral components play a role in the host adaptation of influenza A viruses. Receptor binding on host cells is one factor. Studies of differentiated primary human lung epithelial cell cultures have shown that the hemagglutinin (HA) of human viruses preferentially binds to α-2,6 receptors on nonciliated, secretory cells, whereas avian HA preferentially binds to α-2,3-linked sialic acid receptors on ciliated cells (30). Furthermore, the human lung distribution of infection varies: human viruses preferentially bind α-2,6 receptors, which are found throughout the airway, while avian viruses bind α-2,3 receptors primarily localized to the alveoli. It has been suggested that this restriction of the replication of avian viruses to the deep lung limits human-to-human transmission (38). The presence of a multibasic cleavage site in the HA of highly pathogenic avian H5N1 viruses also increases virulence (15, 18).
The influenza virus polymerase complex also is known to play a role in host adaptation. The complex is comprised of three proteins, PA, PB1, and PB2. In virions, a copy of the complex binds each viral RNA (vRNA) segment, which also is encapsidated by the nucleocapsid protein NP, forming the viral ribonucleoprotein (vRNP). During infection, vRNPs are released from the endocytosed virion after low-pH exposure in the late endosome (29) and then transported to the nucleus, where the viral genome is transcribed and replicated by the polymerase complex (35). PB1 is the catalytic component of the complex. PB2 binds the cap on host pre-mRNA molecules as part of the mechanism known as cap snatching (8, 14, 22). PA recently has been shown to possess the endonuclease activity required for cap snatching (7), and mutations in PA can affect transcription and replication.
PB2 is an important factor in the determination of host range (1), and influenza virus replication efficiency in mammalian cells is significantly affected by the amino acid at position 627 in PB2 (43). Furthermore, studies of highly pathogenic H5N1 avian influenza viruses have identified that PB2, especially residue 627, plays a major role in enhanced pathogenicity in mammalian hosts. While a glutamic acid (E) is located at this position in avian viruses, a lysine (K) is found at this position in human viruses (15, 39). H5N1 viruses with 627E show low pathogenicity and lung-restricted viral replication in mice, while viruses with 627K are highly pathogenic with systemic viral replication in mice (16). However, the presence of 627K in PB2 is not essential for high pathogenicity in the mammalian host, since some H5N1 avian viruses isolated from patients, such as A/HK/481/97, A/HK/532/97, and A/Vietnam/1204/2004, are highly pathogenic in mice but have the avian residue PB2 627E (19, 26). Additional studies, including those with the avian-adapted H7N7 influenza virus SC35 and its mouse-adapted derivative SC35M, have found asparagine (N) at 701 to be important for mammalian host adaptation, possibly due to enhanced interaction with importin α1 (10-12). A recent study has shown that 701N can compensate functionally for the absence of lysine at 627 in viral transmission in guinea pigs (42). The absence of 627K and 701N from some highly pathogenic strains suggests that other residues in PB2 contribute to mammalian host adaptation.
To identify mutations in polymerase genes that enhance the polymerase activity of avian viruses in mammalian hosts, we compared the polymerase activity of a mildly pathogenic avian influenza virus, A/chicken/Nanchang/3-120/01 (H3N2) (Nan), and a well-characterized human strain, A/Aichi/68 (H3N2) (Aichi). Our results using a reporter gene assay confirmed previous findings that PB2 is a key component for high polymerase activity in mammalian cells (23, 32). We then used mutagenesis to introduce highly conserved human residues into Nan PB2. The analysis of the polymerase activity of these mutants revealed that amino acid 271 is a key residue for enhanced activity in mammalian cells. We confirmed that this residue enhances mammalian adaptation using recombinant viruses both in vitro and in vivo. Interestingly, the 2009 pandemic swine-origin H1N1 virus (S-OIV) also contains the human conserved residue at 271 in its avian-like PB2, and mutation to the avian virus-conserved residue reduced polymerase activity. Overall, our studies indicate that PB2 residue 271 plays a major role in virus growth in mammalian hosts.
MDCK, 293T, DF-1, and A549 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 8% fetal calf serum (FCS). A/Aichi/2/68 (H3N2) (Aichi) was obtained from A. Klimov (Centers for Disease Control, Atlanta, and Prevention, GA). A/chicken/Nanchang/3-120/01 (H3N2) (Nan) (25) and A/California/04/2009 (H1N1) (Cal) were provided by R. Webster and R. Webby (St. Jude Children's Research Hospital, Memphis, TN). A/WSN/33 (H1N1) (WSN) was rescued from cDNAs provided by Y. Kawaoka (University of Wisconsin, Madison). Aichi, Cal, and WSN were propagated in MDCK cells, and Nan was propagated in 10-day-old embryonated eggs.
Aichi, Nan, and Cal NP, PA, PB1, and PB2 genes were cloned by reverse transcription-PCR (RT-PCR) using total RNA extracted from infected cells and inserted into pCAGGS/MCS for expression (20, 34). Mutations in PB2 genes were created using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). Nan NP, PA, PB1, PB2, and mutant PB2 genes were subcloned to pPolI to rescue recombinant virus. pCAGGS/MCS containing WSN NP, PA, PB1, and PB2 genes, as well as pPolI-WSN vectors, were obtained from Y. Kawaoka. The pPolI-NP-Luc construct was obtained from T. Wolff (Robert-Koch Institute, Berlin, Germany). The pRL-SV40 vector (Promega, Madison, WI) expresses Renilla under the control of a simian virus 40 promoter. pPRC425-FluA-Luc, containing the firefly luciferase gene flanked by the noncoding 3′ and 5′ regions of A/WSN/33 NP under a chicken RNA polymerase I promoter, was developed from pPRC425-FluA CAT(−), which was obtained from N. Naffakh (Institut Pasteur, Paris, France) (27). The luciferase gene was amplified from pPolI-NP-Luc using appropriate primers and then subcloned into the pPRC425 vector, replacing the CAT gene. All constructs were sequenced for confirmation.
For polymerase activity assays, 293T cells in a 12-well plate were transfected with 0.4 μg pCAGGS-NP, pCAGGS-PA, pCAGGS-PB1, pCAGGS-PB2, 0.1 μg pPolI-NP-Luc, and 0.1 μg pRL-SV40 using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) in Optimem for 24 h. Luciferase production was assayed using the dual-luciferase reporter assay system (Promega) according to the manufacturer's instructions. Polymerase activity was normalized to Renilla expression. To measure polymerase activity in avian cells, DF-1 cells in a 12-well plate were transfected with 0.6 μg pCAGGS-NP, pCAGGS-PA, pCAGGS-PB1, and pCAGGS-PB2 and 0.1 μg pPRC425-FluA-Luc using Fugene 6 (Roche, Palo Alto, CA) in Optimem for 6 h, and then the medium was changed to DMEM plus 8% FCS for 18 h. At 24 h posttransfection, polymerase activity was assayed using the luciferase assay system (Promega). All results shown are the averages with standard deviations from three independent experiments. Data were analyzed using GraphPad Prism 5; statistical tests include one-way repeated-measure analysis of variance (ANOVA) followed by Tukey's multiple comparison test.
Viruses were rescued using the 12-plasmid rescue system developed by Neumann et al. (33). Briefly, a 293T/MDCK coculture in a 6-well plate was transfected with 0.1 μg of each pPolI plasmid and 0.4 μg each of pCAGGS-NP, pCAGGS-PA, pCAGGS-PB1, and pCAGGS-PB2 using Lipofectamine 2000 (Invitrogen) in Optimem. Twenty-four hours posttransfection, medium was changed to DMEM plus 1% FCS. Rescued viruses were plaque purified, and stock viruses were prepared in MDCK cells. Viruses rescued have WSN HA, NA, M, and NS, with Nan NP, PA, PB1, and wild-type or mutant Nan PB2. The PB2 sequence of all viruses was confirmed. The titers of virus on MDCK cells were determined using the method of Reed and Muench (37).
A549 or MDCK cells were infected at a multiplicity of infection (MOI) of 0.01 for 1 h, washed once with Dulbecco's phosphate-buffered saline (PBS) with magnesium and calcium (Invitrogen), and then cultured at 34 or 37°C in DMEM containing 0.15% bovine serum albumin (BSA) and tosylsulfonyl phenylalanyl chloromethyl ketone (TPCK)-trypsin at 1 μg/ml. At the indicated time points, 10% of the culture supernatant was harvested, and viral titers were determined as described above.
To determine morbidity and mortality, 4-week-old BALB/c female mice (Taconic Farms, Inc., Germantown, NY) were anesthetized with Avertin and infected intranasally with the indicated viruses at 4 × 105 50% tissue culture infectious doses (TCID50) in 30 μl. Body weight was measured daily for 10 days, and mice were observed for morbidity. Thirty percent weight loss was considered fatal, and mice reaching this limit were humanely sacrificed. To determine virus replication in the lungs of infected mice, a minimum of three mice were sacrificed on days 2, 4, and 6. Lungs were removed and homogenized in 1 ml of Dulbecco's PBS. The clarified supernatant was stored at −80°C until analysis. Virus titers in lung homogenates were determined as described above. Two independent experiments were performed for a minimum of six mice per virus per time point. For the histological analysis of infected lungs, two mice from each group were sacrificed on day 6. Lungs were perfused with 10% formalin, fixed, and embedded in paraffin, and then 50 μm hematoxylin and eosin sections were prepared. Images were taken using a Leica DMIRB inverted microscope. All animal experiments were approved by the University of Rochester University Committee on Animal Resources.
Previous studies have found that the activity of avian influenza virus polymerase complexes is low in mammalian cells compared to the activity of human polymerase complexes (23, 32). To confirm these results, we determined the activity of a chicken isolate, A/chicken/Nanchang/3-120/01 (H3N2) (Nan), a human isolate, A/Aichi/2/68 (H3N2) (Aichi), and a mouse-adapted human isolate, A/WSN/33 (H1N1) (WSN), polymerase complex in human embryonic kidney 293T and immortalized chicken embryo fibroblast DF-1 cells using a reporter gene assay. First, we analyzed the polymerase activity of Nan, Aichi, and WSN complexes in avian DF-1 cells at 34, 37, and 39°C (Fig. (Fig.1A).1A). The temperatures chosen were used to approximate the conditions which the influenza virus polymerase would face in different hosts. All complexes were active in DF-1 cells, with Nan and Aichi activity being higher at 37 and 39°C than at 34°C (Fig. (Fig.1A).1A). The differences in activity between Nan and Aichi at 34, 37, and 39°C are 8-, 7-, and 4-fold, respectively. We next compared the polymerase activity in human 293T cells (Fig. (Fig.1B).1B). The Aichi and WSN polymerase complexes were highly active, especially at lower temperatures, but both complexes showed slightly lower activity at 39°C. Compared to peak activity, the reduction in Aichi and WSN at 39°C was 4- and 5-fold, respectively. In sharp contrast to Aichi and WSN, the Nan complex was much less active in 293T cells at all temperatures. The difference in activity between Nan and Aichi was 181-, 69-, and 35-fold at 34, 37, and 39°C, respectively. These results indicate that the polymerase activity of the avian Nan complex is much lower in human 293T cells than that of the human Aichi strain, which is active in both human and avian cells.
To determine the gene responsible for the low polymerase activity of the Nan complex in 293T cells, we analyzed the activity of mixed polymerase complexes composed of Nan and Aichi components (Fig. (Fig.2).2). The individual replacement of Nan NP, PA, or PB1 with that of Aichi did not significantly increase the polymerase activity of the avian complex. However, the replacement of Nan PB2 with that of Aichi significantly increased the activity. Compared to that of the all-Nan complex, the activity of Nan with Aichi PB2 increased 12-, 9-, and 7-fold at 34, 37, and 39°C, respectively. Similarly, the replacement of Aichi PB2 with Nan PB2 in an otherwise-Aichi complex reduced activity 115-, 255-, and 236-fold at 34, 37, and 39°C, respectively. The replacement of Aichi NP or PB1 with that of Nan did not significantly affect activity; however, Nan PA reduced activity 2.7- to 17.6-fold. These results clearly indicate that the origin of PB2 is the major factor that determines influenza virus polymerase activity in 293T cells.
The amino acid at position 627 in PB2 has been found previously to affect polymerase activity and viral pathogenicity in mammalian hosts, although it does not fully explain all pathogenic cases. We performed genome-wide sequence comparisons of avian and human isolates and identified amino acid residues in PB2 that are highly conserved host-specific residues. Studies with similar sequence analysis also have reported the presence of host-specific residues in PB2 and other genes (3, 9). As shown in Table Table1,1, there are 10 highly conserved residues in PB2. The presence of these highly conserved residues likely indicates selective pressure to allow viral function and growth in specific hosts. To determine if these highly conserved host-specific residues in PB2 affect polymerase activity in mammalian cells, we first created each avian-to-human mutation in pCAGGS Nan PB2 and determined the polymerase activity of Nan complexes containing PB2 mutants at 34 (Fig. (Fig.3A),3A), 37 (Fig. (Fig.3B),3B), and 39°C (Fig. (Fig.3C)3C) in 293T cells and at 34 (Fig. (Fig.3D)3D) and 39°C (Fig. (Fig.3E)3E) in DF-1 cells. As expected from previous reports, the mutation E627K significantly increased the activity of an otherwise-avian Nan complex at all three temperatures compared to that of the complex containing WT PB2 (39- to 62-fold). We also found that mutation at 271 from the avian virus conserved residue threonine (T) to the human virus conserved alanine (A) enhanced polymerase activity 9.1- and 4.9-fold at 37 and 39°C, respectively (Fig. 3B and C). Both increases are statistically significant (P < 0.05). At 34°C, the activity of the PB2 T271A mutant did not significantly increase above that of the WT (Fig. (Fig.3A).3A). The Nan PB2 A588I mutation slightly increased polymerase activity at 37 and 39°C (Fig. 3B and C) 3.2- and 2.5-fold, respectively. We also examined the polymerase activity of these mutations in DF-1 cells and found that they do not significantly affect polymerase activity at low (34°C; Fig. Fig.3D)3D) or high (39°C; Fig. Fig.3E)3E) temperature, except the Nan T271A mutation increased polymerase activity 2.5-fold at 34°C compared to that of the WT.
We next determined if additional mutations at the conserved residues further enhanced the activity of a 627K-containing polymerase. We made a series of double mutants in pCAGGS-NanPB2 containing the E627K mutation with one of the other nine host-specific residues and examined the polymerase activity in 293T and DF-1 cells (Fig. (Fig.4).4). The activity of all combinations in 293T cells at 34 (Fig. (Fig.4A),4A), 37 (Fig. (Fig.4B),4B), and 39°C (Fig. (Fig.4C)4C) was similar to the activity of Nan complexes containing 627K alone. At 34°C, the polymerase activity of the 271A/627K double mutant was increased 2.4-fold above the activity of the 627K mutant. We also examined the polymerase activity of these double mutants in DF-1 cells and found that generally they do not significantly affect polymerase activity (Fig. 4D and E).
Our in vitro reporter gene assay results showed that the T-to-A mutation at 271 increased polymerase activity in 293T cells. We next wanted to confirm this enhanced activity using rescued viruses. To determine virus growth in tissue culture as well as in a mouse model, we used a WSN viral background. We rescued four viruses containing WSN HA, NA, M, and NS with Nan NP, PA, PB1, and either Nan WT or Nan mutant (271A, 627K, or both 271A/627K) PB2. The viruses were rescued successfully, and stock viruses were prepared after plaque cloning in MDCK cells.
We first compared the multicycle growth of these viruses in A549 and MDCK cells at 34 (Fig. 5A and B) and 37°C (Fig. 5C and D). Cells were inoculated at a multiplicity of infection of 0.01, and supernatant was sampled at various time points after infection. At both temperatures in A549 cells (Fig. 5A and C), the PB2 271A mutant grew more rapidly and to a higher titer than that of the WT. This effect was more pronounced at 34°C, although the 271A mutation did not enhance the polymerase activity significantly at 34°C (Fig. (Fig.3).3). The 271A/627K and 627K mutants also grew more rapidly than the WT. Similar results were obtained in MDCK cells (Fig. 5B and D). At both 34 and 37°C, the 271A mutant grew more rapidly in MDCK cells than the WT.
To test whether the phenotypes of the mutant viruses seen in vitro manifested in vivo, we evaluated the growth and pathogenicity of these viruses in a mouse model. Four-week-old BALB/c mice were infected intranasally with 4 × 105 TCID50, and body weight was monitored daily for 10 days (Fig. (Fig.6A).6A). As expected from previous reports, mice infected with the 627K mutant rapidly lost weight, and four of four infected mice died by day 8. Mice infected with the 271A/627K mutant lost weight similarly to those infected with the 627K mutant, with most mice succumbing by day 8. Interestingly, although the 271A mutant grew to a higher titer than the other mutants in A549 or MDCK cells, mice infected with the 271A mutant did not lose a significant amount of weight, and no mortality was observed. The weight change pattern of WT virus-infected mice was similar to that of 271A mutant-infected mice, although WT virus-infected mice gained weight slightly more rapidly than 271A mutant-infected mice.
We next determined lung virus titers on days 2, 4, and 6 postinfection (Fig. (Fig.6B).6B). In WT virus-infected mice, the virus titer was at its highest, 2.4 × 106 TCID50/ml, at day 2, then the titer decreased rather quickly to 3.6 × 105 by day 4 and 7.7 × 104 on day 6. In 271A mutant-infected mice, the virus titer in lungs was 4.7 × 106 TCID50/ml, approximately 2-fold higher than that of the WT on day 2. However, in sharp contrast to WT virus-infected mice, the virus titer of 271A-infected mice remained high at day 4 (5.0 × 106 TCID50/ml), 14-fold higher than that of WT virus-infected mice. On day 6, however, the virus titer decreased to 3.1 × 105 TCID50/ml. The titer of the 627K mutant was higher than that of the WT or the 271A mutant at 1 × 107 TCID50/ml at day 2, which gradually decreased about 3-fold by day 4 and then approximately 6-fold further by day 6. The 271A/627K mutant virus was at its highest titer on day 2, approximately 2.7 × 107 TCID50/ml. The titer decreased about 5-fold at day 4 but maintained a similar level of virus (4.7 × 106 TCID50/ml) at day 6.
To further compare the pathogenicity of these viruses in vivo, we performed histological analysis of lungs from mice infected with WT, 271A, 627K, and 271A/627K mutants on day 6 postinfection (Fig. (Fig.7).7). Uninfected mice displayed thin alveolar walls and large air spaces, which are indicative of normal lung function (data not shown). Mice infected with the WT virus exhibited mild inflammation with immune infiltrates in alveolar interstitial spaces and the area surrounding bronchioles. In contrast to the WT, infection with the 271A mutant led to a significant increase of cellular infiltration in the lung parenchyma and the thickening of the basement membrane of bronchioles. Similarly, infection with the 627K mutant caused excessive infiltration. In addition, unlike 271A, 627K mutant infection was associated with severe hemorrhage. The lungs of mice infected with the 271A/627K double mutant showed the most severe inflammatory cell infiltration around blood vessels and in the lung parenchyma. This infection also led to severe hemorrhage and alveolar destruction (Fig. (Fig.7D7D).
Our data described above indicate that a T-to-A mutation at position 271 in PB2 enhances polymerase activity and virus growth in mammalian cultured cells and mice, suggesting that the T271A mutation contributes to avian polymerase adaptation to mammalian hosts. During the course of our experiments, the pandemic swine-origin H1N1 virus (S-OIV) began to circulate in Central and North America. Sequence analysis revealed that S-OIV isolates contain an avian-type PB2 (5). The PB2 of the S-OIV contains most of the avian isolate conserved residues, including 627E. However, two avian virus conserved residues were changed: at 271 to A and at 588 to T. As shown above, the T271A mutation increases polymerase activity in mammalian cells. The T at 588 is not a conserved residue for either avian or human isolates. To determine if the residues at 271 and 588 affect the polymerase activity of S-OIV, we first compared the activity of the S-OIV polymerase complex using the NP, PA, PB1, and PB2 genes cloned from A/California/04/2009 (Cal). The polymerase activity of Cal was compared to that of WSN, Aichi, and Nan by reporter gene assay in 293T cells. The Cal polymerase complex was highly active in mammalian cells (Fig. (Fig.8).8). At 34°C, the Cal complex was 35.5-fold more active than the avian Nan complex, but it was slightly less active than the human complexes: 5.1-fold less active than Aichi and 1.9-fold less active than WSN (Fig. (Fig.8A).8A). At 37 (Fig. (Fig.8B)8B) and 39°C (Fig. (Fig.8C),8C), the Cal polymerase was as active as Aichi and WSN. To determine if 271A or 588T contributes to the enhanced activity of the avian-like Cal polymerase in mammalian cells, we used site-directed mutagenesis to introduce the avian residue T at position 271 and the human and avian residues, isoleucine (I) and A, respectively, at position 588. We also made 271T/588A and 271T/588I double mutants. The 271T mutation in Cal PB2 decreases polymerase activity, especially at low temperature: 6-fold lower at 34°C (Fig. (Fig.8A)8A) and 2-fold lower at 37°C (Fig. (Fig.8B).8B). The reduction at 34°C is statistically significant (P < 0.05). The effects of the individual mutations in residue 588 were less than those of the 271T mutation; however, the 588 mutation to the human virus-conserved I was more active than that containing the avian virus-conserved A at all three temperatures determined. Also, the activity of PB2 containing 271T was enhanced by the 588I mutation, although the effect of the mutation at 588 was less significant than that at 271. These results suggest that, as observed with Nan PB2, mutation at 271 in an avian-origin PB2 contributes to the polymerase activity of Cal in mammalian hosts.
To further evaluate the potential role of residue 271 in swine influenza infection, we next collected 220 full-length PB2 sequences of swine influenza viruses from the Influenza Research Database (41) and examined the residues at positions 271, 627, and 701 (Table (Table2).2). Among the swine viruses that contain 627E, 51% have 271A and 28% contain 701N, suggesting that these two residues can compensate for the lack of 627K in PB2. We next analyzed the PB2 sequences from swine viruses isolated in North America (Table (Table3).3). Almost all (98.9%) of the viruses isolated before 1998 contained PB2 627K paired with the avian residue 271T. Interestingly, most of the viruses isolated after 1998, when triple-reassortant viruses were introduced into the North American swine population (13, 40), contained PB2 627E and 271A. Only 6.3% of swine isolates contained avian residues at both 271 and 627. Of the North American swine viruses isolated between 1998 and 2009, 17 are of the H1N1 subtype, and 10 of them have 271A and 627E. Furthermore, the analysis of all PB2 sequences available for the S-OIV as of December 2009 showed 271A/627E/701D. These results suggest that 627K is not essential for polymerase function in the swine host and that 271A is likely to contribute to virus growth and spread in swine and possibly other mammals.
The direct transmission of avian H5N1 influenza A viruses to humans suggests a new mechanism for influenza outbreaks. However, potential mutations in avian viruses that confer increased pathogenicity or transmissibility in humans and other mammalian hosts are not fully examined. In this study, we identified PB2 residue 271 as a key amino acid for polymerase function and virus growth in mammalian cells. Our data show that a single mutation at 271 from the avian virus-conserved T to the human virus-conserved A enhances polymerase activity in a reporter gene assay, as well as the growth of a recombinant virus containing an otherwise-avian polymerase both in vitro and in vivo. Our results indicate a role for PB2 residue 271 in the mammalian adaptation of avian influenza A viruses.
PB2 residues 627 and 701 have been shown previously to enhance viral pathogenicity and transmission in mammalian hosts. The role of PB2 627 is the most extensively studied. A single mutation from E to K dramatically improves the function of an avian polymerase complex and enhances viral growth in mammalian cells, especially at lower temperatures (28, 32, 39, 43). Some H5N1 viruses with E at this position show low pathogenicity in mice, while those with K are highly pathogenic (15). In vivo experiments have further revealed that 627K allows for the growth of H5N1 viruses in the upper respiratory tracts of mice (16), and viruses with PB2 627E show reduced transmission in a guinea pig model compared to that of viruses containing 627K (42). These studies suggest that enhanced replication at lower temperatures contributes to enhanced growth in the upper airway of infected animals or humans and thus improves transmission. Almost all human viruses have 627K, whereas avian viruses have 627E in PB2 (3, 9). The analysis of the polymerase activity of the avian Nan polymerase complex using a reporter gene assay indicates that the introduction of the E627K mutation enhanced activity at both low and high temperatures in 293T cells, while the T271A mutation enhanced activity only at higher temperatures (37 and 39°C) (Fig. (Fig.2).2). This result suggests that the mechanism of the enhancement of polymerase activity differs between the mutations at 271 and 627. It is highly likely that enhanced polymerase activity by the 271A mutation contributes to increased virus growth in the mouse lung (Fig. (Fig.6B).6B). Interestingly, the T271A mutation did not increase polymerase activity at low temperature as determined in vitro, although the growth of the T271A mutant virus in cultured cells was more efficient than that of WT virus at 34°C (Fig. (Fig.5).5). It is not clear how the T271A mutation enhances virus growth in cultured cells at 34°C. It is possible that the slight increase in polymerase activity detected in the reporter gene assay (Fig. (Fig.3A)3A) is sufficient for enhanced virus growth in tissue culture (Fig. (Fig.5).5). Another possibility is that, in addition to its role in polymerase activity, the highly conserved host-specific amino acid 271 also plays an important role in virus assembly and spread in mammalian cells.
The PB2 mutation D701N also has been implicated in the adaptation of H5N1 viruses to mammalian hosts (6, 24). Studies using an avian-adapted H7N7 strain and its mouse-adapted variant showed that the single mutation D701N in PB2 improved polymerase activity 3-fold in a reporter gene assay in 293T cells (10, 11). Other studies using recombinant A/Panama/2007/99 (H3N2) and A/Vietnam/1203/04 (H5N1) viruses showed that the D701N mutation enhances transmission between guinea pigs (42). A recent study also has shown that the D701N mutation enhances the binding of PB2 to importin α1 and correspondingly increases the level of PB2 in the nucleus in mammalian cells, suggesting that the adaptation of the viral polymerase to the nuclear import machinery plays an important role in the interspecies transmission of influenza virus (12).
Of the 10 highly conserved PB2 residues we examined in this study, only two, E627K and T271A, enhanced the polymerase activity of an otherwise-avian complex in mammalian cells. None of the other mutations significantly enhanced polymerase activity in mammalian 293T cells, except the A588I mutation, which increased activity about 3-fold (Fig. (Fig.3).3). It is not clear if other conserved residues play roles in adaptation to mammalian hosts. Although these residues did not enhance polymerase activity in vitro, these residues could be involved in interactions with mammalian host proteins during the viral life cycle and thus are required for the efficient replication and assembly of the virus in mammalian host cells. They also may affect the interaction of PB2 with other influenza virus proteins in mammalian cells. Since other structural components also contain host-specific residues, it is of interest to further determine the role of these conserved residues in virus growth.
The results of our in vitro reporter gene assay (Fig. (Fig.8)8) and sequence analysis (Tables (Tables22 and and3)3) support the idea that PB2 residue 271 contributes to the 2009 novel swine-origin influenza A (H1N1) virus human pandemic. The S-OIV possesses PB2 and PA genes of North American avian virus origin; a PB1 gene of human H3N2 virus origin; HA, NP, and NS genes of classical swine virus origin; and NA and M genes of Eurasian avian-like swine virus origin (5, 13). Previous studies as well as our data in Fig. Fig.22 clearly indicate that polymerase complexes containing avian virus PB2 show poor activity in mammalian host cells. However, the polymerase activity of the S-OIV is much higher than that of the Nan complex (Fig. (Fig.8).8). The avian-origin S-OIV PB2 contains both 627E and 701D, the same residues found in Nan and most other avian viruses. However, the S-OIV contains PB2 271A, which we showed enhances polymerase activity and virus growth in mammalian hosts (Fig. (Fig.3,3, ,5,5, and and6).6). In fact, the activity of the Cal polymerase complex at 34°C was reduced 6-fold by mutation at 271 from A to the avian virus-conserved residue T (Fig. (Fig.8A).8A). In addition, the 271A mutation significantly enhanced virus spread in cultured mammalian cells, especially at lower temperature (Fig. (Fig.5).5). Therefore, 271A in S-OIV is likely to contribute to its efficient transmission through both enhanced polymerase activity and virus growth in mammalian hosts.
PB2 271A has been maintained in the majority of swine isolates since the introduction of triple-reassortant viruses around 1998 (Table (Table3),3), and all pandemic S-OIV isolates have 271A/627E/701D. However, the polymerase activity of the Cal complex with 271T still is higher than that of the Nan polymerase complex. In addition, the A271T mutation in the Cal polymerase complex did not significantly affect polymerase activity at higher temperatures, suggesting that additional unidentified residues in PB2 or other polymerase components also contribute to the high polymerase activity of the Cal complex in mammalian cells.
Unlike the E627K mutation, the T271A mutation by itself was not sufficient to cause lethal infection by WSN-based viruses containing Nan NP and polymerase proteins. However, the virus titer in lungs of mice infected with the 271A mutant was much higher than that of WT virus-infected mice (Fig. (Fig.6).6). Additionally, the double mutant 271A/627K showed a higher lung virus titer than the E627K mutant, especially at 6 days postinfection, when the virus titer in lungs infected with the 271A/627K mutant was almost 10-fold higher than that of 627K mutant-infected mouse lungs. Pathological analysis of lungs showed the most severe lesions in 271A/627K mutant-infected mice with massive inflammatory cell infiltration, severe hemorrhage, and alveolar destruction (Fig. (Fig.7).7). These results suggest the contribution of the T271A mutation to the pathogenicity of influenza virus in mice. However, the fact that the 271A mutation by itself cannot cause lethal infection, despite the high viral titer in lungs, suggests that virus growth in the mouse lung does not necessarily correlate with virus pathogenicity. Pathological analysis of infected lungs indicates that both the T271A and E627K mutants induce massive cell infiltration. However, unlike 271A, infection with the 627K mutant was associated with severe hemorrhage, suggesting that the 627K virus causes more severe damage in lung tissue (Fig. (Fig.7).7). We also noticed an earlier and more severe cytopathic effect in 627K virus-infected A549 and MDCK cells compared to that of cells infected with the T271A mutant (data not shown). These observations suggest that, in addition to enhanced virus growth, lung tissue damage strongly influences the pathogenicity of influenza A viruses.
While our studies have shown that residue 271 is important for the mammalian adaptation of influenza virus, the mechanism of action of 271 still needs to be determined. Studies of 627 mutants suggest that residue 627 affects the interaction of PB2 with NP (36). Another study suggests that the E-to-K mutation at 627 facilitates escape from an inhibitory factor that restricts the function of avian-derived polymerase in human cells (31). Additional studies by Kuzuhara et al. (21) have found that a charged “basic groove” in 627K-containing PB2 binds RNA, especially the 5′ vRNA promoter, with high affinity compared to that for 627E, suggesting that the RNA-binding activity of PB2 contributes to enhanced polymerase activity (21). PB2 627E and PB2 627K possess distinct electrostatic properties (45) and may interact differentially with host inhibitory factors (31) and/or viral RNA (21) as a result. It is not known whether PB2 residue 271 also is involved in the interactions of PB2 with NP or RNA. Early cross-linking studies suggested that two separate sequences of PB2 at residues 242 to 282 and residues 538 to 577 are involved in cap binding (17). However, the crystal structure of a PB2 fragment containing residues 318 to 483 includes a PB2 cap-binding domain within the solved structure (14). It is possible that residue 271 locates close to the cap binding domain in the three-dimensional structure and thus affects the interaction of PB2 with the cap structure. Future structural studies of the PB2 domain containing residue 271 will help to elucidate the function and mechanism of the action of residue 271 of PB2 and its effect on host adaptation.
This work was supported by the New York Influenza Center of Excellence (NYICE), a member of the NIAID CEIRS network, under NIH contract HHSN266200700008C.
We thank L. Martinez-Sobrido (University of Rochester Medical Center), A. Klimov, R. Webster, R. Webby, Y. Kawaoka, T. Wolff, and N. Naffakh for reagents.
Published ahead of print on 24 February 2010.