|Home | About | Journals | Submit | Contact Us | Français|
We have previously shown that infection with laboratory-passaged strains of influenza virus causes both specific degradation of the largest subunit of the RNA polymerase II complex (RNAP II) and inhibition of host cell transcription. When infection with natural human and avian isolates belonging to different antigenic subtypes was examined, we observed that all of these viruses efficiently induce the proteolytic process. To evaluate whether this process is a general feature of nonattenuated viruses, we studied the behavior of the influenza virus strains A/PR8/8/34 (PR8) and the cold-adapted A/Ann Arbor/6/60 (AA), which are currently used as the donor strains for vaccine seeds due to their attenuated phenotype. We have observed that upon infection with these strains, degradation of the RNAP II does not occur. Moreover, by runoff experiments we observe that PR8 has a reduced ability to inhibit cellular mRNA transcription. In addition, a hypervirulent PR8 (hvPR8) variant that multiplies much faster than standard PR8 (lvPR8) in infected cells and is more virulent in mice than the parental PR8 virus, efficiently induces RNAP II degradation. Studies with reassortant viruses containing defined genome segments of both hvPR8 and lvPR8 indicate that PA and PB2 subunits individually contribute to the ability of influenza virus to degrade the RNAP II. In addition, recently it has been reported that the inclusion of PA or PB2 from hvPR8 in lvPR8 recombinant viruses, highly increases their pathogenicity. Together, the data indicate that the capacity of the influenza virus to degrade RNAP II and inhibit the host cell transcription machinery is a feature of influenza A viruses that might contribute to their virulence.
The genome of the influenza A viruses consists of eight single-stranded RNA segments of negative polarity, encoding a total of 11 proteins. Upon entry into susceptible cells, infecting ribonucleoprotein complexes (RNPs) are transported to the nucleus, where transcription and replication take place. Replication of viral RNAs (vRNAs) involves the synthesis of positive-strand replicative intermediates (cRNAs) that are exact copies of the virion RNAs (for a review, see reference 15). A functional link between viral and cellular transcription has been proposed since influenza virus mRNA transcription is initiated using short capped RNA oligonucleotides as primers that are obtained by endonucleolytic cleavage of de novo-synthesized cellular pre- mRNAs (6, 56). This cap-snatching process is performed by the viral polymerase, a heterotrimeric complex comprised of the PB1, PB2, and PA subunits (15, 30, 40).
The carboxy-terminal domain (CTD) of the largest subunit of the RNA polymerase II (RNAP II) complex plays an essential role in cellular transcription. This domain is differentially phosphorylated during the transcription cycle, dynamically permitting or impeding its association with a large number of factors (27). Two major forms of RNAP II can be found in cells when the CTD of its largest subunit is hyperphosphorylated or hypophosphorylated. The main phosphorylation target is the YS2PTS5PS heptapeptide that is repeated 52 times in the mammalian CTD and that is mainly phosphorylated at Ser-2 and Ser-5 (52). Phosphorylation of serine 5 is detected in the promoter region of the transcribing genes, whereas serine 2 phosphorylation increases as the RNAP II leaves the promoter and proceeds along the gene during the transcription elongation process (55). In agreement with the proposed coupling between influenza virus and cellular transcription, it has been reported that the vRNA polymerase interacts with the CTD of the initiating RNAP II phosphorylated at Ser-5 (16). Indeed, the latter is the RNAP II form engaged in the capping process during cellular transcription. Therefore, it was suggested that this interaction might be required for the vRNA polymerase to gain access to capped RNA substrates for their endonucleolytic cleavage. Several cellular factors that interact with the viral polymerase or RNPs could be involved in the viral-cellular RNA polymerase interaction. Among other candidates, the interaction with splicing factors such as hnRNPA1, hnRNPM, or PSF/SFPQ (37), transcription-related factors such as Ebp-1 (29), PARP-1 (45), transcription intermediary factor 1β (45), and DDX5 (37) or with the positive mRNA transcription modulator hCLE/CGI-99 (32, 54) could be particularly important.
Some viruses use replication strategies that alter the host cell transcription machinery. For example, cytomegalovirus (1, 67), Bunyamwera virus (70), Epstein-Barr virus (3), human immunodeficiency virus type 1 (12, 34, 74), or adenovirus (51) induce changes in the phosphorylation state of the CTD of the RNAP II. Other viruses, such as herpes simplex virus type 1 (13) or influenza virus (58), provoke the degradation of the largest subunit of the RNAP II complex. Both the dephosphorylation of the CTD and the degradation of the RNAP II could be considered as a potential mechanism in viral pathogenesis. Indeed, degradation of the RNAP II upon influenza virus infection has been shown to correlate with a subsequent decrease in cellular transcriptional activity (58).
Influenza PR8 strain has been used for the past 30 years to produce inactivated influenza vaccine. This vaccine contains the six so-called internal gene protein segments of PR8 and the neuraminidase (NA) and hemagglutinin (HA) genes of the circulating viruses. The PR8 donor virus has been passaged in mice, ferrets and chicken embryos, leading to virulence attenuation for humans (4). Indeed, experimental evidence has demonstrated that PR8 is also attenuated in chickens (44). Although it has been shown that the PR8 virus is pathogenic in standard laboratory mouse strains, these mice lack a complete antiviral defense system because they carry defective alleles of the Mx1 gene (66), and consequently the PR8 strain displays low virulence in Mx+/+ mice (22). In contrast, a hypervirulent PR8 mutant strain (hvPR8) that grows very fast in cultured cells is highly virulent in Mx+/+ mice (22).
The aim of the present study was to study the ability of the attenuated PR8 virus to degrade the RNAP II in infected cells and to investigate the presence of this phenotypic trait in other influenza viruses. Our results indicate that degradation of RNAP II is a general feature of nonattenuated influenza A viruses, since every virus tested with the exception of the vaccine donor strains PR8 and A/Ann Arbor/6/60 (AA) induce the degradation process. In addition, the hvPR8 virus produces efficient RNAP II degradation and the use of reassortant viruses indicates that PA and PB2 individually contribute to both RNAP II degradation and virulence.
The cell lines used in the present study were HEK293T, HeLa, and A549 (human); COS-1 (monkey); NIH 3T3 (mouse); NLB2 (canine); and DF-1 (chicken). The influenza viruses A/Victoria/3/75 (H3N2) (VIC), A/WSN/33 (H1N1) (WSN), A/PR8/8/34 (H1N1) (PR8), A/New Caledonia/20/99 (H1N1) (NC), A/England/1/51 (H1N1) (E), A/Wyoming/3/2003 (H3N2) (Wy), A/Ann Arbor/6/60 ca (H2N2) (AA), A/Vietnam/1203/04 (H5N1), A/Turkey/Wisconsin/66 (H9N2), and mouse-adapted PR8 strains of low (lvPR8) and high (hvPR8) pathogenicity (22) were used. Different PR8 and WSN viruses coming from the Centro Nacional de Biotecnología, Centers for Disease Control, Instituto de Salud Carlos III, and Mount Sinai School of Medicine were also used. All of the PR8 viruses are closely related to the Mount Sinai strain of A/PR/8/34 (62), except the hvPR8 and the PR8 virus from Centers for Disease Control and Prevention that are related to the Cambridge strain (72). α-Amanitin was obtained from Sigma. The protease inhibitor “Complete” was from Roche. Benzonase was purchased from Novagen.
Cells were infected at a multiplicity of infection of 3 PFU/cell. At different times postinfection, cells were collected in phosphate-buffered saline with protease and phosphatase inhibitors (500 μM sodium orthovanadate, 500 μM β-glycerophosphate, and 500 μM sodium molybdate), and the cell pellet was resuspended in Laemmli sample buffer. Western blotting was carried out as described previously (58). The following primary antibodies were used: for RNAP II, monoclonal antibodies 8WG16 (1:500), H14 (1:500), and H5 (1:500) from BabCo; for PA, monoclonal antibody 14 (1:250) (2); for PB1, a rabbit polyclonal antibody (1:1,000) (20); for nucleoprotein (NP), a rat polyclonal antibody generated using as antigen a His-NP protein expressed and purified from bacteria (1:2,000); and for β-tubulin, a mouse monoclonal antibody (1:15,000) from Sigma.
The immunoprecipitation studies were performed as described previously (42). Briefly, 107 HEK293T cells were mock infected or infected with either the PR8 or VIC influenza virus strains (at 3 PFU/cell) and at 6 h postinfection (hpi) the cells were collected and lysed in buffer composed of 150 mM NaCl, 1.5 mM MgCl2, 10 mM Tris-HCl (pH 8.5), and 0.5% Igepal, with protease and phosphatase inhibitors (500 μM sodium orthovanadate, 500 μM β-glycerophosphate, 500 μM sodium molybdate) and 100 U of benzonase. The lysate was centrifuged at 10,000 × g, and the supernatant was used for immunoprecipitation studies with the 8WG16 antibody or an unrelated antibody (monoclonal antibody to HA epitope from Covance). The immunocomplexes were washed 10 times with a buffer containing 100 mM NaCl, 10 mM Tris-HCl (pH 8.0), 0.5% Triton X-100, 0.5 mM dithiothreitol, and 0.2 mM EDTA, and the coimmunoprecipitated proteins were analyzed by Western blot.
To analyze RNA synthesis, cultures of HEK293T cells were mock infected or infected with the VIC or PR8 strains, and their nuclei were isolated and frozen as described previously (54). These nuclei were used to detect in vitro RNA synthesis by incorporation of [32P]α-GTP (250 μCi/ml) during a 30-min pulse, in the presence or absence of α-amanitin (5 μg/ml), as described previously (58). After in vitro RNA synthesis, total RNA was isolated by phenol extraction and ethanol precipitation. The RNA was quantified by ethidium bromide staining in a Bio-Rad Chemi Doc equipment. The radioactivity incorporated was quantified in a phosphorimager after blotting to a nylon membrane.
Total RNA from HEK293T infected cells was isolated using Ultraspec RNA isolation reagent from Biotex. Northern blotting was performed using standard conditions and the membranes were hybridized with oligonucleotide probes radiolabeled with [γ-32P]ATP. The probes recognized NP positive- or negative-sense RNA, and their sequences were 5′-TCTTAGGATCTTTCCCCGC-3′ and 5′-GTCTTCGAGCTCTCGGAC-3′, respectively.
Rescue of lvPR8 and hvPR8 recombinant viruses was done as previously described (22).
To examine whether the PR8 virus induces RNAP II degradation, we infected HEK293T cells with either the PR8 or the VIC strains of influenza virus and evaluated the accumulation of RNAP II after infection. The VIC strain was used as a positive control since we have previously characterized its ability to degrade the RNAP II (58). The results obtained showed that, whereas the levels of the RNAP II forms detected with the 8WG16 antibody (mainly hypophosphorylated form and also partially phosphorylated at Ser5 [36, 46]) start to decrease at around 6 hpi and are almost undetectable at 15 hpi in cells infected with the VIC strain, this was not the case in cells infected by the PR8 strain (Fig. (Fig.1A),1A), even when analyzed at later times postinfection (Fig. (Fig.1B).1B). As we reported previously, the hyperphosphorylated forms of the RNAP II detected with the H5 (phosphorylated at Ser2, Ser5, and at Ser2+Ser5) and H14 (phosphorylated at Ser5 and Ser5+Ser2) antibodies and other cellular proteins such as β-tubulin remained unaltered. It should be noticed that a shift to the upper band recognized by the 8WG16 is observed in the PR8 infection. A similar shift has been found during herpes simplex virus type 1 infection and seems to be unlinked to virus-induced repression of host cell transcription (57, 65).
Next, we sought to determine whether the failure of PR8 to induce RNAP II proteolysis was host cell dependent. Studies using cells of human (HeLa and A549), monkey (COS-1), canine (NLB2), mouse (NIH 3T3), and chicken (DF-1) origin (Fig. (Fig.2)2) indicated that none of these PR8-infected cells had decreased RNAP II levels, in contrast to VIC-infected cells. It should be noticed that NIH 3T3 and DF-1 cells have lower levels of the RNAP II upper band recognized by the 8WG16 antibody. These results indicate that the PR8 strain of influenza virus is unable to induce degradation of the largest subunit of the RNAP II regardless of the origin of the host cell being infected.
As previously reported, the three subunits of the influenza virus polymerase associate with the Ser-5 hyperphosphorylated forms of the RNAP II as a heterotrimeric complex and not as individual subunits or heterodimers (16). In addition, we have previously observed that the in vivo reconstituted polymerase complex is sufficient to trigger the RNAP II degradation (58). Thus, viral-cellular polymerase association could be a requisite to induce the proteolytic pathway leading to RNAP II degradation. To test the importance of this interaction, HEK293T cells were mock infected or influenza virus infected with the VIC or PR8 strains and at 6 hpi, immunoprecipitation of RNAP II was performed, and the presence of viral polymerase in the immunocomplexes was evaluated. The results showed that the PB1 polymerase subunit, as a representative of the entire polymerase complex, of both the VIC and PR8 strains coprecipitated with RNAP II with similar efficiency (Fig. (Fig.3).3). Hence, the association of viral polymerase with cellular RNAP II may be a prerequisite, but it is not sufficient to trigger its degradation.
We previously described that global cellular transcription is inhibited upon influenza virus infection, as is the specific synthesis of mRNAs corresponding to several cellular genes (58). This inhibition could be the direct consequence of virus-induced RNAP II degradation associated with infection, although other mechanisms related to the vRNA transcription-replication strategy could modulate the activity of the host cell transcription machinery. To evaluate whether the absence of RNAP II proteolysis during PR8 infection was correlated with changes in cellular mRNA synthesis, cell transcription was analyzed following infection with the PR8 or VIC strains. Cultures of HEK293T cells were infected with either the PR8 or the VIC strains or they were mock infected. The nuclei of these cells were then isolated at various times postinfection and used for runoff transcription experiments with or without α-amanitin, to evaluate specifically the activity of the RNAP II, as previously reported (58). The results (Fig. (Fig.4)4) showed that there is less inhibition of cellular transcription after PR8 infection than that following infection with the VIC strain, even at a late time postinfection. These data support that the RNAP II degradation contributes to the inhibition of cellular transcription that takes place upon influenza virus infection. These results are in agreement with those showing decreased amounts of RNAP II associated with the internal gene regions in infected cells, suggesting that influenza virus infection inhibits transcription elongation and triggers premature RNAP II termination (9). On the other hand, the obtained data with the PR8 strain indicate that, in addition to the RNAP II degradation, other mechanisms should be involved in the transcription inhibition during influenza virus infection, such as the cap-snatching process (41), the inhibition of cleavage and the polyadenylation of cellular pre-mRNAs (10, 49) or the nuclear retention of poly(A)-containing cellular mRNAs (18) among others.
Since the failure of PR8 infection to produce RNAP II proteolysis could be explained by differences in vRNA transcription and/or replication pattern of this particular strain, we characterized vRNA transcription and replication in HEK293T cells infected with the PR8, VIC, or WSN strains. At different times postinfection, total RNA was isolated and used for Northern blot analysis, detecting NP positive- or negative-sense RNA with specific oligonucleotide probes, as described previously (58). The accumulation of vRNA starts at around 2 to 4 hpi, and it continues throughout the infection period (Fig. (Fig.5A,5A, vRNA). Higher levels of vRNA accumulated after infection by the VIC strain than by infection with PR8 or WSN. The lower RNA replication rate of these latter strains is in agreement with specific virus-host cell interactions modulating RNA replication, since these two viruses are chicken- and mouse-adapted strains, respectively. Conversely, VIC is a human strain and, accordingly, its replication rate is higher in the human cell line used in the present study. Given that WSN and PR8 accumulate comparable amounts of vRNA with similar kinetics and that the WSN strain produces extensive RNAP II degradation (Fig. (Fig.5C),5C), it is unlikely that vRNA replication has any effect on cellular RNAP II proteolysis. In addition, the accumulation of viral positive sense RNA (Fig. (Fig.5B),5B), was similar in the different strains studied (since the amount of cRNA is much lower than that of mRNA, the data presented would mainly represent the mRNA levels), suggesting that the utilization of cellular cap-oligonucleotides per se does not activate the degradation process.
Next, we tested laboratory passaged influenza viruses coming from different sources to discard that the stocks that we routinely use had acquired the degradation properties by a particular passage story. To this aim, we used WSN and PR8 strains from different sources (see Materials and Methods), as well as a recombinant virus lacking NS1 rescued in the PR8 genetic background (19). The WSN viruses successfully induced degradation, whereas the PR8 viruses were all unable to degrade RNAP II (Fig. (Fig.6A6A).
Previous studies demonstrated that the cold-adapted AA influenza virus is avirulent in humans, as well as in animal models (mice and ferrets) (38). Different viral genes (PA, PB2, PB1, and M) independently contribute to its attenuated phenotype (63). Therefore, this cold-adapted strain has been used to produce safe live influenza A virus vaccines derived by genetic reassortment of its six internal genes with the HA and NA of the circulating viruses (11, 38). Due to its attenuation we sought to determine whether this virus could degrade RNAP II. Thus, cells from different origins such as human (HEK293T and A549), canine (NLB2), mouse (NIH 3T3), and chicken (DF-1) were infected with the AA strain at 33°C. At 15 hpi the accumulation levels of the RNAP II were analyzed by Western blotting. The results (Fig. (Fig.6B)6B) showed that similarly to the other vaccine donor virus, the PR8 strain, the infection with AA virus does not induce degradation of the RNAP II, even at late times postinfection (Fig. (Fig.6C).6C). These results suggest that the absence of RNAP II degradation is linked to the influenza virus attenuation.
Having characterized the ability of laboratory-passaged influenza virus strains to trigger RNAP II degradation, we wanted to evaluate whether this degradation was also brought about by natural isolates of influenza virus. We infected HEK293T cells with different human strains of the H1N1 subtype, such as A/England/1/51 and A/New Caledonia/20/99, or with the H3N2 subtype A/Wyoming/3/2003. The A/England/1/51 strain was chosen to assess whether RNAP II degradation was a recently acquired phenomenon or whether it was already present in older naturally circulating viruses. Antigens of A/New Caledonia/20/99 and A/Wyoming/3/2003 viruses have been included in the vaccine strains, since these viruses have been circulating recently (7, 24). We also used avian strains of the H5N1 subtype (A/Vietnam/1203/04), which was isolated from humans (43) and is a very pathogenic isolate with high virus titers in many organs, including brain (21), and an avian H9N2 subtype strain (A/Turkey/Wisconsin/66) (28, 47) that was analyzed due to the potential of H9N2 viruses to transmit to mammalian species (8, 53). In all cases, the accumulation of RNAP II, the influenza virus PB1 and PA subunits, and of β-tubulin was analyzed after 15 hpi by Western blotting. On this occasion, PA protein from the avian viruses was almost undetectable, probably due to the absence of the corresponding epitope recognized by the monoclonal antibody in the avian proteins. The results (Fig. (Fig.7)7) indicated that all human and avian influenza virus isolates induced RNAP II degradation as the laboratory strains (VIC and WSN), supporting the notion that this process is a general feature of nonattenuated influenza A viruses.
Next, we examined whether the alterations in RNAP II were correlated with the different degrees of virulence in vivo. Recently, it has been reported the characterization of a particular highly pathogenic PR8 virus that was generated by serial lung passages in Mx+/+ mice (23). Its phenotype includes a much faster growth in cell culture and higher pathogenicity for Mx+/+ mice (22). We compared the behavior of this highly pathogenic PR8 virus (hvPR8) to that of a common PR8 strain described as low-pathogenic PR8 (lvPR8) (22) and that of the PR8, VIC, and WSN viruses used throughout the present study. Infection of HEK293T cells with the hvPR8 virus produced RNAP II degradation to the same extent as the VIC or the WSN strains (Fig. (Fig.8),8), whereas lvPR8 did not induce degradation. These results indicate that the degradation of the main component of the cellular transcription machinery, the largest subunit of the RNAP II, is a process induced by nonattenuated influenza virus strains.
Previous reports have shown that individually expressed PA subunit induces a protease activity, appearing to be a general characteristic since PA from both human and avian strains shows this phenotype (48, 60). Moreover, a recombinant virus containing a PA subunit with a decreased induction of proteolysis has defects in RNA replication, a delayed RNAP II degradation, and a reduced pathogenicity in mice (31, 58). Recently, it has been described that PA contains an endonuclease activity (14, 73) that has been mapped in its N-terminal part coinciding with the domain where its induction of proteolysis resides (61). Having taken these data into account, we tried to identify the underlying genetic mechanisms involved in the RNAP II degradation, using the model of low- and high-virulence PR8 viruses by generating reassortants containing different genes of these two viruses. The recombinants contained the three polymerase genes of the hvPR8 or each of its individual subunits in the lvPR8 genetic background (Table (Table1)1) . All of the recombinants contained the HA and NA genes from the hvPR8 virus. The results in HEK293T cells infected with these reassortants are shown in Fig. Fig.9.9. The recombinant virus containing the lvPR8 polymerase (lvPPP) did not induce the proteolytic process, whereas the opposite recombinant, the lvPR8 virus containing the hvPR8 polymerase (hvPPP), did produce degradation (Fig. (Fig.9A).9A). The analysis of the individual contribution of the different polymerase subunits showed, that whereas PB1 subunit from hvPR8 (hvPB1) does not induce proteolysis of the RNAP II, the exchange of either the PA or the PB2 subunits (hvPA and hvPB2), clearly reduces the RNAP II accumulation (Fig. (Fig.9B).9B). In Table Table11 it is also shown that hvPA and hvPB2 recombinants increase the virulence of the corresponding viruses in infected mice, whereas hvPB1 does not. These results indicate that PA and PB2 subunits individually contribute to the RNAP II degradation in influenza A virus-infected cells, and this effect correlates with an increase in virulence in mice.
Influenza A viruses cause annual outbreaks in humans and are responsible for several hundred thousand deaths worldwide (68). Periodically, antigenically novel transmissible strains emerge in humans that cause pandemics such as the one in 1918 that infected hundreds of million people and resulted in the death of an estimated 20 million (68). Among the viral factors that play a key role in virulence, the surface glycoproteins HA and NA, the polymerase complex, and the nonstructural protein NS1 are particularly important (50).
Influenza virus infection causes important alterations in the cellular transcription apparatus in cultured cells, including RNAP II degradation and inhibition of mRNA synthesis. Interestingly, the PR8 and AA attenuated strains are unable to degrade the main component of the RNAP II complex. Indeed, PR8 infection is associated with less inhibition of cellular mRNA transcription. Thus, we would like to discuss the possible contribution of the main influenza virus virulence factors to these processes.
HA and NA mediate virus entry into the host cell and dissociation of progeny particles and hence cooperate to virulence. The viral strains used in the present study belong to the H3N2 (VIC, A/Wyoming/3/2003), H1N1 (WSN, PR8, A/New Caledonia/20/99, A/England/1/51), H2N2 (AA), H5N1 (A/Vietnam/1203/04), and H9N2 (A/Turkey/Wisconsin/66) subtypes. Even though they do not share the same HA and NA constellation, all of the strains behave similarly in terms of RNAP II degradation, except for PR8 and AA (Fig. (Fig.66 to to8)8) (58). In addition, the PR8 and WSN strains share the same HA and NA subtype but clearly differ in their RNAP II degradation capacity. Moreover, the reassortant viruses of lvPR8 and hvPR8 viruses that contain the same HA and NA proteins induce or do not induce the degradation process, independently of their HA and NA composition (Fig. (Fig.9).9). These results indicate that despite the important role that viral glycoproteins have for influenza viruses in nature, they do not seem to be involved in the RNAP II degradation process that takes place in infected cells.
The nonstructural NS1 protein is a multifunctional protein with two main roles. (i) It works as an IFN-antagonist, and (ii) it modulates the cellular RNA metabolism. The hvPR8 and lvPR8 viruses used in the present study are indistinguishable with regard to inhibition of IFN synthesis in infected cells (22). Indeed, NS1 sequence comparison between hvPR8 and lvPR8 has shown that the high virulent PR8 strain only contains a conservative amino acid change at residue 101. However, while lvPR8 does not degrade RNAP II the hvPR8 virus induced an efficient degradation. In addition, the PR8 virus lacking NS1 fails to degrade the RNAP II since the rest of the PR8 viruses analyzed (Fig. (Fig.6)6) and the reassortant viruses containing the same NS1 protein (presented in Fig. Fig.9)9) behave differently in terms of RNAP II degradation. Therefore, NS1 is not involved in the proteolytic process that produces RNAP II degradation triggered by influenza A viruses in infected cells.
Numerous reports have pointed out the importance of the polymerase complex in viral pathogenesis. Recent findings strongly implicate the vRNA polymerase complex as a major determinant of the pathogenicity of the 1918 pandemic virus (71). Indeed, a total of 10 amino acid changes in the polymerase proteins differentiate the 1918 virus sequence from avian consensus sequences, and a notable number of the same changes have been found in recently circulating highly pathogenic H5N1 viruses that have caused deaths in humans (69).
It has been suggested that a specific gene constellation at the level of the genes coding for the three polymerase subunits is responsible for the consistent attenuation of the PR8 and the AA strains (17, 26, 63). Accordingly, it has been reported that the three polymerase subunits contribute individually to the attenuated phenotype of the AA strain (64), as well as to its temperature-sensitive phenotype (35). Using reassortant viruses, it has been previously described that among the key virulence factors of the hvPR8 strain, the polymerase genes largely contribute to its pathogenicity (22). Further work has identified that PA and PB2 subunits from hvPR8, but not PB1, promote the high virulence of influenza A virus in mice (Table (Table1)1) (59).
We present data here indicating that the entire polymerase complex of the hvPR8 confers the ability of the reassortant lvPR8 virus to degrade RNAP II and that both PA and PB2 subunits individually contribute to this process (Fig. (Fig.9).9). Furthermore, the degradation of the RNAP II observed in influenza virus-infected cells correlates with enhanced virulence in mice. Altogether, these results support the notion that the viral polymerase and especially the PA and PB2 subunits play an important role in virulence in agreement with their contribution to the RNAP II degradation.
Since PA expression induces a proteolytic activity, it is tempting to speculate that this function can be involved in RNAP II degradation, being modulated by the PB2 subunit. It should be pointed out that a specific interaction between the N-terminal 100 amino acids (aa) of PA and PB2 has been recently detected (25). This domain of PA-PB2 interaction is contained in the region of PA that presents endonuclease activity (aa 1 to 197) (14, 73) and induction of proteolysis (aa 1 to 247) (61). Therefore, PB2 could have a role regulating some of these PA activities in order to trigger the proteolytic process. However, as we previously described, the expression of PA-PB2 dimers does not induce the proteolytic process, since the entire polymerase complex is necessary to obtain RNAP II degradation (58). Therefore, although PB2 can modulate PA activity, this effect should be performed within the entire polymerase complex and with the possible contribution of specific cellular factors.
Virus-induced RNAP II degradation seems to be a general feature of influenza A viruses, since of nine different strains analyzed, including those of avian and human origin, mouse-adapted strains, and laboratory or natural isolates, all except the attenuated strains induced the degradation. Influenza virus induces profound alterations in the host cell, due to several mechanisms that include (i) cap-snatching of cellular pre-mRNAs (41), (ii) inhibition of cleavage and polyadenylation of cellular pre-mRNAs (10, 49), (iii) nuclear retention of poly(A)-containing cellular mRNAs (18), (iv) degradation of cytoplasmic cellular mRNAs (5, 33, 75), and (v) preferential utilization of the translation machinery by the viral-specific mRNAs (39). The reported RNAP II degradation exerted by the nonattenuated viruses could represent one strategy to avoid the competition of the RNAP II, once active synthesis of cellular pre-mRNA is no longer required. This mechanism would then contribute to highjack the metabolism of the infected cell and to suppress the establishment of the host antiviral defense against viral pathogens.
We are indebted to J. Ortín, O. Haller, C. Rivas, E. Yángüez, and L. Ver for their criticisms on the manuscript. The technical assistance of Y. Fernández and N. Zamarreño is also gratefully acknowledged.
The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention or the Agency for Toxic Substances and Disease Registry.
This study was supported by Ministerio de Educacion y Ciencia, Plan Nacional de Investigacion Científica, Desarrollo e Innovacion Tecnologica (BFU2005-02834), Comunidad de Madrid (S-SAL-0185-2006), and Ciber de Enfermedades Respiratorias. A.R. and A.P.-G. were supported by Ciber de Enfermedades Respiratorias.
Published ahead of print on 19 August 2009.