Viral RNA-binding ability conferred by SUMOylation at PB1 K612 of influenza A virus is essential for viral pathogenesis and transmission

Posttranslational modifications, such as SUMOylation, play specific roles in the life cycle of invading pathogens. However, the effect of SUMOylation on the adaptation, pathogenesis, and transmission of influenza A virus (IAV) remains largely unknown. Here, we found that a conserved lysine residue at position 612 (K612) of the polymerase basic protein 1 (PB1) of IAV is a bona fide SUMOylation site. SUMOylation of PB1 at K612 had no effect on the stability or cellular localization of PB1, but was critical for viral ribonucleoprotein (vRNP) complex activity and virus replication in vitro. When tested in vivo, we found that the virulence of SUMOylation-defective PB1/K612R mutant IAVs was highly attenuated in mice. Moreover, the airborne transmission of a 2009 pandemic H1N1 PB1/K612R mutant virus was impaired in ferrets, resulting in reversion to wild-type PB1 K612. Mechanistically, SUMOylation at K612 was essential for PB1 to act as the enzymatic core of the viral polymerase by preserving its ability to bind viral RNA. Our study reveals an essential role for PB1 K612 SUMOylation in the pathogenesis and transmission of IAVs, which can be targeted for the design of anti-influenza therapies.


Introduction
Influenza A virus (IAV) is an important zoonotic pathogen that causes frequent epidemics and occasional pandemics in humans. Based on the antigenicity of the surface glycoproteins hemagglutinin (HA) and neuraminidase (NA), IAVs are classified into 18 different HA subtypes and 11 different NA subtypes [1]. Since 1900, humans have suffered four influenza pandemics, among which the most recent 2009 pandemic H1N1 virus transmitted to over 215 countries and territories between April 2009 and August 2010 [2]. Moreover, humans are also constantly facing threats posed by avian influenza viruses (AIVs), with H5N1 and H7N9 human infections as two prime examples. The first human infection with H5N1 AIV occurred in Hong Kong in 1997 [3], and between 2003 and 2020, 861 human infection cases were reported, of which 455 were fatal [4]. The H7N9 low pathogenic AIVs identified in 2013 and the subsequently mutated H7N9 highly pathogenic AIVs in 2017 [5][6][7][8] led to 1568 human infections, including 615 fatal cases [9]. Therefore, continued efforts to elucidate the factors and mechanisms underlying the pathogenesis and transmission of IAVs remain critical for the development of novel anti-influenza therapies.
The viral ribonucleoprotein (vRNP) complex of IAVs is composed of individual viral RNA segments, three polymerase subunits [i.e., polymerase basic protein 1 (PB1), polymerase basic protein 2 (PB2), polymerase acidic protein (PA)], and nucleoprotein (NP) [10]. PB1 functions as the catalytic subunit and the assembly core of the RdRp [11][12][13], which is responsible for the transcription and replication of the IAV genome [14,15]. In addition to its N-and C-terminal extensions that make contact with PA and PB2, respectively, PB1 contains a central region comprising finger, fingertips, palm, and thumb domains [16], which are characteristics of RdRp. Terminal initiation during replication is dependent on the priming loop in the thumb domain of PB1 [17], which is also essential for transcriptional elongation, but not for transcriptional initiation [17,18].
SUMOylation is an important regulatory posttranslational protein modification mechanism in eukaryotic cells. SUMO (small ubiquitin-related modifier) is a ubiquitin-like protein that is covalently conjugated to a target protein. There are four SUMO forms in humans, SUMO1, SUMO2, SUMO3, and SUMO4, among which SUMO1-SUMO3 are ubiquitously expressed and covalently conjugated to the lysine residues of target proteins [19]. SUMOs are conjugated to target proteins via a three-step cascade, involving a heterodimeric E1 activating enzyme, SAE1/SAE2; a unique E2 conjugating enzyme, Ubc9; and several E3 ligases [19,20]. The SUMOylation of a particular substrate may interfere with or facilitate its interaction with binding partners, regulate its activity, function, subcellular localization, and can affect its stability [21,22].
IAV proteins are also targets of SUMOylation. To date, three IAV proteins [i.e., non-structural protein 1 (NS1), matrix protein 1 (M1), and NP] have been clearly identified as substrates of SUMO and the role of their SUMOylation in virus replication has been clarified. Among them, NS1 was the first identified SUMOylated IAV protein, the SUMOylation of which leads to enhanced protein stability and improved virus growth in cell culture [23,24]. SUMOylation of M1 has been shown to facilitate the nuclear export of vRNP complexes and the assembly of virus particles [25]; SUMOylation of NP is essential for the intracellular trafficking of NP and for virus growth in vitro [26]. More IAV proteins could be targets of SUMO modification, and IAV infection has been shown to induce a global increase in cellular SUMOylation [27,28]. However, the role of SUMO modification in the pathogenesis and transmission of IAV has not been explored.
Here, we demonstrated that PB1, the catalytic core of the viral RdRp, is SUMOylated by SUMO1 in both transfected and virus-infected cells. We found a highly conserved lysine residue at position 612 (K612) is the key SUMO acceptor site of PB1. The SUMOylation of PB1 at K612 was found to be critical for vRNP complex activity and virus replication in vitro. Mechanistically, the SUMOylation of PB1 at K612 was essential for PB1 to function in the binding of vRNA. Importantly, a SUMOylation-deficient PB1 K612R mutation significantly attenuated the virulence of H1N1, H5N1 and H7N9 viruses in mice. Furthermore, the replication and respiratory droplet transmission of the PB1 K612R mutant of a 2009 pandemic H1N1 virus in ferrets was reduced compared with the wild-type virus, which created a stress condition that drove the R612K reversion mutation in inoculated and exposed ferrets.

IAV PB1 is mainly SUMOylated by SUMO1
Given the importance of SUMOylation in the posttranslational modification of viral proteins, we first investigated whether IAV PB1, the core subunit of viral RdRp, is a target of the host SUMOylation system. HEK293T cells were cotransfected with Flag-tagged PB1 of A/WSN/33 (WSN, H1N1) virus (WSN-H1PB1), along with or without SUMO-conjugating enzyme Ubc9, SUMO1 or SUMO1 mut (a conjugation-deficient SUMO1 variant), and SUMO-specific protease SENP1 or SENP1 mut (SENP1 with loss-of-function mutations) [24,26]. At 48 h post-transfection, the cell lysates were immunoprecipitated with anti-Flag agarose, and then western blotted with an anti-PB1 monoclonal antibody (mAb). When WSN-H1PB1 was cotransfected with Ubc9 and SUMO1, a clear additional band for PB1 of~110 kDa was detected, reflecting the potential modification of PB1 (95 kDa) by SUMO1. In contrast, this 110-kDa band was absent from cells cotransfected with SUMO1 mut , indicating that WSN-H1PB1 is indeed SUMOylated (Fig 1A). To confirm these results, we introduced an expression construct for the de-SUMOylating enzyme SENP1 or SENP1 mut into the transfection. The high molecular weight band of SUMOylated PB1 at~110 kDa was markedly reduced when cotransfected with SENP1, whereas it was unaffected in the SENP1 mut -expressing cells ( Fig 1A). Of note, the same molecular weight bands of WSN-H1PB1 were also detected when the precipitated proteins were analyzed by using an anti-SUMO1 mAb, further confirming that WSN-H1PB1 was modified by SUMO1 ( Fig 1A).
Finally, we determined whether IAV PB1 was also SUMOylated by the other two forms of SUMO, SUMO2 and SUMO3. As shown in Fig 1D-1F, WSN-H1PB1, VN-H5PB1, and AH-H7PB1 were not SUMOylated when cotransfected with Ubc9 and SUMO2 or SUMO3.
Taken together, these results clearly demonstrate that IAV PB1 protein is a target for SUMOylation by SUMO1, but not SUMO2 or SUMO3.

The lysine residue at position 612 (K612) is the SUMOylation site of IAV PB1
To identify the SUMOylation site(s) of PB1 protein, we used the in silico prediction program GPS-SUMO [29]. Three lysine residues, K379, K612, and K736, were predicted to be potential candidate sites for SUMO-conjugation in WSN-H1PB1 (Fig 2A). Individual K-to-R mutant PB1 proteins for all three predicted lysine residues were then generated, and their effect on the SUMOylation of PB1 was examined by coimmunoprecipitation (co-IP) and western blotting analysis. The precipitated proteins were detected by using an anti-PB1 mAb and an anti-SUMO1 mAb, respectively. As expected, WSN-H1PB1 was SUMOylated to form a 110-kDa high molecular weight band when cotransfected with Ubc9 and SUMO1. The main SUMOylated band was also present in the lanes transfected with the PB1/K379R and PB1/K736R mutants, but not the PB1/K612R mutant. These results indicate that K612, but not K379 or K736, is the key SUMO acceptor site of WSN-H1PB1 (Fig 2B).

PLOS PATHOGENS
with SUMO1, whereas the indicated SUMOylated PB1 band disappeared when SUMO1 mut was used instead of SUMO1. However, when the WSN-H1PB1/K612R mutant was cotransfected with SUMO1 or SUMO1 mut , the 110-kDa PB1-SUMO1 band was not detected. These data further confirm that K612 is the authentic SUMOylation site of WSN-H1PB1.
We next determined whether VN-H5PB1 and AH-H7PB1, which both possess K612, were also SUMOylated at this site. Compared with wild-type VN-H5PB1 and AH-H7PB1, which were SUMOylated by SUMO1, analysis of the VN-H5PB1/K612R and AH-H7PB1/K612R mutants did not reveal the 110-kDa PB1-SUMO1 band (Fig 2D and 2E), indicating that the lysine residue at position 612 is also a major SUMOylation site of VN-H5PB1 and AH-H7PB1. By performing a systemic sequence analysis involving 14,226 PB1 sequences, we found that the K612 residue is highly conserved among different IAV subtypes (S1 Table). Collectively, our results indicate that the K612 residue is a common SUMO acceptor site of the IAV PB1 proteins.

SUMOylation at K612 does not affect the stability or cellular localization of PB1
After identifying K612 as the major SUMOylation site of IAV PB1, we investigated whether the stability of PB1 is affected by SUMOylation at K612. HEK293T cells were transfected with an expression plasmid encoding WSN-H1PB1 or WSN-H1PB1/K612R mutant. At 24 h posttransfection, protein synthesis was blocked with cycloheximide (CHX) and the steady-state levels of PB1 after CHX treatment were monitored by use of western blotting. Both PB1 and PB1/K612R protein levels remained stable at 0, 2, 4, 6, 8, and 12 h under CHX treatment ( Fig  4A), indicating that SUMOylation at K612 has no effect on the stability of IAV PB1. To further demonstrate that SUMOylation of PB1 at K612 does not alter protein stability, we cotransfected Flag-tagged WSN-H1PB1 or WSN-H1PB1/K612R mutant with Ubc9 and SUMO1 in HEK293T cells. As shown in Fig 4B, both native PB1 and SUMOylated PB1 were present in cells expressing WSN-H1PB1, and the level of total PB1 remained stable across different time points; native PB1, but not the 110-kDa PB1-SUMO1, was detected in cells expressing WSN-H1PB1/K612R, whose level remained unchanged. These results demonstrate that the stability of IAV PB1 is unaffected by the SUMOylation status of the K612 residue. We also found that SUMOylation at K612 did not affect the stability of VN-H5PB1 or AH-H7PB1 ( Fig  4C-4F). Together, these data show that SUMOylation at K612 does not affect the stability of IAV PB1.
Since SUMOylation could potentially affect the cellular localization of target proteins, we examined whether SUMOylation at K612 altered the intracellular distribution of IAV PB1. A549 cells were infected with WSN (H1N1) or WSN-PB1 K612R (H1N1) virus at an MOI of 5, and the cellular localization of the PB1 protein was analyzed by use of an immunofluorescence assay. As shown in Fig 5A and 5B, at 2 h post-infection (p.i.), the fluorescence signal of PB1 was very weak, and PB1 mainly localized to the cytoplasm of both WSN-and WSN-PB1 K612Rinfected cells. At 4 h p.i., PB1 had entered the nucleus and appeared to accumulate there. At 6 h and 8 h p.i., the PB1 signal was strong in the nucleus of the infected cells, and by 10 h p.i., PB1 was largely exported from the nucleus and had accumulated in the cytoplasm. Overall, no observable quantitative difference was found between WSN-and WSN-PB1 K612R -infected cells with respect to the localization of PB1 across the infection course ( Fig 5C). These results indicate that PB1 SUMOylation at 612K does not influence the cellular localization of PB1 during virus infection.

SUMOylation at PB1 K612 does not affect the assembly of the viral RNP complex
PB1 is the core subunit of the viral heterotrimeric polymerase complex, with its N-terminal region interacting with PA and its C-terminal region binding PB2 [12,13]. We therefore tested whether the SUMOylation of PB1 at K612 affects the PB1-PA and PB1-PB2 interactions. HEK293T cells were transfected for 48 h with WSN-H1PB1 or WSN-H1PB1/K612R mutant individually or simultaneously were transfected with WSN-H1PB2 or WSN-H1PA. Cell lysates were immunoprecipitated with anti-PB1 mAb and then western blotted to detect PB1, PB2, or PA. The amount of PB2 or PA that was immunoprecipitated by WSN-H1PB1 or WSN- H1PB1/K612R was comparable (Fig 6), indicating that the interactions between PB1-PB2 and PB1-PA were not affected by the SUMOylation at K612 of PB1. We also tested whether the SUMOylation at PB1 K612 affects the assembly of the viral RNP complex. Viral RNP complexes, bearing WSN-H1PB1 or WSN-H1PB1/K612R, were reconstituted in HEK293T cells by co-transfection of expression plasmids for PB2, PB1 or PB1/K612R, PA, and NP of WSN (H1N1) virus, together with pHH21-SC09NS F-Luc, which was used to provide a vRNA-like template RNA [30][31][32]. At 48 h post-transfection, the vRNP complexes were immunoprecipitated from cell lysates by using an anti-PB1 mAb, and the presence of PB2, PA, and NP in the precipitated vRNP complexes was detected by western blotting. No significant differences in the amounts of PB2, PA, and NP were observed between vRNP complexes bearing WSN-H1PB1 and those bearing WSN-H1PB1/K612R (S1 Fig). These results indicate that SUMOylation at K612 of PB1 has no effect on the assembly of the vRNP complex of IAV.

SUMOylation at K612 of PB1 enhances vRNP complex activity
The vRNP complex of IAV is responsible for the transcription and replication of the viral genome [14,15]. We therefore asked whether vRNP complex activity was affected by SUMOylation of PB1 at K612. To this end, a minigenome assay was performed in HEK293T cells, which were transfected with the four vRNP complex protein expression plasmids, PB2, PB1 or PB1/K612R, PA, and NP of WSN (H1N1), VN/1180 (H5N1), or AH/1 (H7N9) virus, together with a reporter plasmid, pHH21-SC09NS F-Luc, encoding a vRNA-like firefly luciferase gene [30][31][32]. Forty-eight hours later, the luciferase activity of the cell lysates was measured to reveal the vRNP complex activity. The results showed that the vRNP complex activity of WSN (H1N1), VN/1180 (H5N1), and AH/1 (H7N9) virus decreased by approximately 50%, 55%, and 95%, respectively, in cells expressing PB1/K612R compared with that of PB1 (Fig 7A-7C). These findings demonstrate that the SUMOylation of PB1 at K612 is essential for the vRNP complex activity of IAV.

SUMOylation-defective PB1/K612R mutation impairs IAV replication
To determine the effect of the SUMOylation-defective PB1/K612R mutation on the in vitro replication of IAV, we first compared the plaque morphologies of the wild-type WSN (H1N1), VN/1180 (H5N1), and AH/1 (H7N9) viruses with those of the SUMOylation-defective PB1/ K612R mutant viruses in MDCK cells. Compared with the wild-type viruses, which formed The results are expressed as the mean ± standard deviation (SD) of triplicates and the significance was tested with a Student's t-test (A-C) or a one-way ANOVA followed by t-test (G-I). � , P < 0.05; �� , P < 0.01; ��� , P < 0.001; ���� , P < 0.0001.
We next examined the replication kinetics of the PB1/K612R mutant viruses. MDCK cells were infected with one of the wild-type viruses or with the PB1/K612R mutant viruses at an MOI of 0.001. Supernatants were collected at 12, 24, 36, and 48 h p.i. and titrated by a plaque assay on MDCK cells. We found that the PB1/K612R mutation led to 25-to 76-, 2-to 7-, and 3-to 22-fold reductions in growth titers from 12 to 48 h p.i. for the WSN-PB1 K612R (H1N1), VN/1180-PB1 K612R (H5N1), and AH/1-PB1 K612R (H7N9) viruses, respectively, compared with the corresponding wild-type viruses. These data further demonstrate that the SUMOylation of PB1 at K612 is critical for optimal replication of IAV (Fig 7G-7I).
To further confirm that the K612R mutation affects the binding of the PB1 protein to vRNA, we conducted an RNA-protein pull-down experiment. Ten micrograms of biotinylated model vRNA was captured with streptavidin magnetic beads. Cell lysates containing 100 μg of total protein that had previously been transfected with a plasmid expressing truncated WSN-H1PB1 protein (PB1 494-757 or PB1 494-757/K612R ) or pCAGGS vector as a control were then added to the RNA-bound beads. The amount of both unbound and vRNA-bound proteins was analyzed by western blotting. We found that the same amount of vRNA bound more WSN-H1PB1 494-757 protein than WSN-H1PB1 494-757/K612R protein (Fig 8K). When we conducted this experiment in the VN/1180 (H5N1) and AH/1 (H7N9) backgrounds, we found that the amount of the PB1/K612R protein bound by vRNA was significantly lower than that of the wild-type protein (Fig 8L and 8M). Collectively, these results demonstrate that the K612R mutation in IAV PB1 impairs the ability of the PB1 protein to bind vRNA.

The virulence of SUMOylation-defective PB1/K612R mutant viruses is attenuated in mice
To determine the significance of SUMOylation at K612 of PB1 to the virulence of IAVs, we  in their organs were determined by using plaque assays on MDCK cells. The WSN (H1N1) and WSN-PB1 K612R (H1N1) mutant viruses replicated only in respiratory organs, and the titers of the WSN (H1N1) virus in the nasal turbinates and lungs of the mice were significantly higher than those in these tissues of the mice that were inoculated with the WSN-PB1 K612R (H1N1) mutant virus (Fig 9A). The VN/1180 (H5N1) virus caused systemic infection, with virus replication detected in the nasal turbinates, lungs, spleens, and kidneys, whereas the replication of the VN/1180-PB1 K612R (H5N1) mutant virus was restricted to the respiratory tract, with much lower titers in the nasal turbinates and lungs than the wild-type VN/1180 (H5N1) virus ( Fig 9B). Similarly, the viral titers of AH/1-PB1 K612R (H7N9) mutant virus in the nasal turbinates and lungs of mice were much lower than those inoculated with AH/1 (H7N9) virus ( Fig 9C). These results demonstrate that the SUMOylation-defective PB1/K612R mutation significantly attenuates the replication of IAVs in vivo.
We next determined the 50% mouse lethal dose (MLD 50 ) of the PB1/K612R mutant viruses. Groups of five mice were intranasally inoculated with serially diluted mutant viruses or the corresponding wild-type viruses, and their body weight, disease signs, and death were monitored daily for 14 days. The wild-type WSN (H1N1) virus caused severe disease and killed the mice at an MLD 50 of 10 3.5 PFU (Figs 9D and S2A). However, no deaths were observed among the mice inoculated with any dose of the WSN-PB1 K612R (Fig 9F and 9G). Similarly, the body weight loss of mice inoculated with AH/1 (H7N9) virus and AH/1-PB1 K612R (H7N9) mutant virus were observed in the 10 5 -10 7 PFU groups, with more severe body weight loss observed in mice inoculated with AH/1 (H7N9) virus than AH/1-PB1 K612R (H7N9) mutant virus (S2E and S2F Fig); and the MLD 50 was 10 5.5 PFU for AH/1 (H7N9) virus, whereas was >10 7.5 PFU for AH/1-PB1 K612R (H7N9) mutant virus, displaying an at least 100-fold reduction compared with the wild-type virus (Fig 9H and 9I). These data indicate that the virulence of the SUMOylation-defective PB1/K612R mutant viruses is highly compromised compared with that of the wild-type viruses in mice.
We also performed histopathological studies on lung sections from the wild-type-and PB1/ K612R mutant virus-inoculated mice. The WSN (H1N1) virus-infected mice exhibited more severe lung lesions (i.e. perivascular edema, infiltration of inflammatory cells, necrosis, and exfoliation of bronchiole epithelial cells) compared with the mice infected with the WSN-PB1 K612R (H1N1) mutant virus. Immunohistochemical analysis with an anti-NP pAb revealed more viral antigen-positive cells in the lungs of mice infected with the wild-type WSN (H1N1) virus than in mice infected with the mutant virus (Fig 9J-9M). Similar results were observed in the lungs of mice infected with the wild-type VN/1180 (H5N1), AH/1 (H7N9) and  (H-M). The results are expressed as the mean ± SD of triplicates and the significance was tested with a one-way ANOVA followed by t-test (H-J). � , P < 0.05; �� , P < 0.01; ��� , P < 0.001.
https://doi.org/10.1371/journal.ppat.1009336.g008  (Fig 9N-9U). These data show that the PB1/ K612R mutant viruses induce significantly less severe lung lesions compared with the wildtype viruses. Collectively, the above findings indicate that the SUMOylation at PB1 K612 plays a pivotal role in the pathogenesis of IAVs in mice.

SUMOylation at PB1 K612 may facilitate the replication and respiratory droplet transmission of IAV in ferrets
Efficient and sustained transmission among humans is the most prominent property of pandemic and epidemic IAVs [35]. To determine whether the SUMOylation of PB1 at K612 affects the transmissibility of IAVs, a 2009 pandemic H1N1 virus, A/Fuzhou/1/2009 (FZ/1, H1N1), and its SUMOylation-defective PB1/K612R mutant FZ/1-PB1 K612R (H1N1) were rescued by reverse genetics. We then tested the replication ability of these viruses in ferrets. Both the FZ/1 (H1N1) and FZ/1-PB1 K612R (H1N1) viruses were detected in the nasal turbinates, soft palate, tonsils, trachea, and lungs of the inoculated ferrets on day 4 p.i. Overall, the titers of the FZ/1-PB1 K612R (H1N1) virus in the organs tested were lower than those of the FZ/1 (H1N1) virus, especially in nasal turbinates, tonsils, and lungs (Fig 10A and 10B). We next tested the respiratory droplet transmissibility of the FZ/1 (H1N1) and FZ/1-PB1 K612R (H1N1) viruses in ferrets. Both the FZ/1 (H1N1) and FZ/1-PB1 K612R (H1N1) viruses were detected in nasal washes from inoculated ferrets between days 2 and 6 p.i. Notably, FZ/1 (H1N1) was present in the nasal washes of all three exposed ferrets on days 3, 5, and 7 post-exposure (p.e.), and was present in the nasal wash of one exposed ferret on day 9 p.e., with peak titers ranging from 6.0 to 7.3 log 10 PFU/mL (Fig 10C). By contrast, in the FZ/1-PB1 K612R (H1N1)-exposed group, the virus was detected in the nasal wash of only one ferret on day 3 p.e., and in the nasal wash of another ferret until day 7 p.e., with peak titers ranging from 3.9 to 6.3 log 10 PFU/mL ( Fig  10D). All of the inoculated ferrets and the exposed ferrets that shed viruses seroconverted on day 21 p.i. or p.e. (Fig 10E and 10F). These results suggest that the SUMOylation-defective PB1/K612R mutation is correlated with the impaired replication and transmission of the FZ/1 (H1N1) virus in ferrets.
Finally, we sought to determine whether the introduced PB1/K612R mutation was stably maintained during viral replication in ferrets. Viral RNAs extracted from nasal washes of FZ/ 1-PB1 K612R -inoculated and -exposed ferrets were reverse-transcribed into cDNA, and the fulllength PB1 gene was amplified and deep-sequenced. Strikingly, PB1/K612R was not stably maintained during virus replication in ferrets. The R612K reversion mutation in PB1 was found on day 2 p.i. in the inoculated ferrets, and all of the viruses contained the R612K reversion mutation in the inoculated animals on day 4 p.i. (S3 Table). In the virus-positive exposed ferret on day 3 p.e., 99% of the viruses in the nasal wash contained PB1 R612. However, the (n = 3). Data shown are mean ± SD for organ samples from three mice; significance was assessed with a one-way ANOVA followed by t-test. �  viruses mutated to K612 quickly, with 96% of the sequencing reads bearing K612 on day 5 p.e. Interestingly, the second virus-positive exposed ferret had already acquired the PB1 R612K reversion mutation when the virus was detected in the nasal wash on day 7 p.e. (Fig 10D and  S3 Table). These results suggest that the transmission of the FZ/1-PB1 K612R (H1N1) virus may have occurred or have been enhanced after it acquired the R612K reversion mutation in its PB1 protein.

Discussion
Posttranslational modifications, such as phosphorylation, acetylation, ubiquitination, and SUMOylation, play important roles in the regulation of protein activity, localization and stability [36]. SUMOylation is an important reversible posttranslational modification that is extensively involved in the regulation of viral protein functions [37]. Several viral proteins of IAV, for example, NS1, M1, and NP, are reported to be SUMOylated in transfected and infected cells [23,[25][26][27]38]. As the assembly core of the IAV polymerase complex, PB1 binds to the PA protein through its N-terminal domain and binds to the PB2 protein through its C-terminal domain in order to play a central role in the transcription and replication of the viral genome. In addition to the clear role of SUMOylation in the posttranslational modification of the specific IAV proteins mentioned above, IAV infection has been shown to induce a global increase in cellular SUMOylation levels and PB1 has been suggested to be a target of SUMOylation [27,28]. However, the molecular details of PB1 SUMOylation and its role in virus replication remain unclear. More importantly, the biological significance of SUMOylation for the pathogenesis and transmission of IAVs has never been investigated.
In the present study, we clearly demonstrated that IAV PB1 is SUMOylated by SUMO1 in transfected and infected cells, and we mapped the key SUMOylation site of the PB1 protein to residue K612 in the C-terminal region. Large-scale sequence analysis revealed that the lysine residue at position 612 of PB1 is highly conserved among different IAV strains. These findings reveal that SUMOylation at K612 of PB1 is a common posttranslational modification mechanism of IAVs. We speculate that the acquisition and stable maintenance of K612 in PB1 is the outcome of host adaptation of IAVs to achieve optimal viral fitness.
The functional consequences of SUMOylation are highly diverse depending on the target proteins [21,39,40]. We investigated the effect of SUMOylation of K612 on the biological properties of the PB1 protein, and found that SUMOylation at K612 did not affect the interaction of PB1 with the other two polymerase subunits, PB2 and PA, and also did not interfere with vRNP complex assembly. Unlike the viral NS1 protein, which is stabilized by SUMOylation [24], the SUMOylation of K612 had no effect on the stability of PB1, and also did not change the cellular localization of PB1 in infected cells. Functionally, we found that the vRNP complex activity decreased significantly when the PB1 protein could not undergo SUMO modification at the mutated K612R residue. In the WSN (H1N1) and VN/1180 (H5N1) virus backgrounds, the vRNP complex activity was reduced by about 50% due to the PB1/K612R mutation, and in the AH/1 (H7N9) virus background, this mutation led to an approximately 95% reduction in vRNP complex activity. Because the SUMO acceptor site K612 is located in the vRNA binding domain of the PB1 protein, we tested whether the K612R mutation affects the binding of PB1 to vRNA. We found that lack of SUMOylation of PB1 at the mutated K612R residue significantly impaired the ability of the PB1 protein to bind to vRNA, thereby severely limiting the replication and transcription of IAVs. These data clearly indicate that SUMOylation at K612 is indispensable to PB1 for the ability to bind to vRNA with high affinity, which is essential for PB1 to fulfil its function as the enzymatic core of the vRNP complex.
In characterizing the effect of SUMOylation at K612 of PB1 on the replication of IAVs, we found that SUMOylation-defective PB1 K612R mutant viruses formed much smaller plaques in MDCK cells than the wild-type viruses, and that their growth kinetics were also severely compromised in MDCK cells. These results are consistent with the role of SUMOylation of other IAV proteins, such as NP, M1, and NS1, for viral replication in vitro [24][25][26]. Consistent with the in vitro growth properties, the PB1/K612R mutant viruses replicated to much lower titers in mouse organs compared with the corresponding wild-type viruses; of note, the replication of the VN/1180-PB1 K612R (H5N1) mutant virus was restricted to respiratory organs, whereas the wild-type VN/1180 (H5N1) virus caused systemic infection.
The biological significance of SUMOylation at K612 of PB1 on the pathogenesis and transmission of IAVs was thoroughly explored. In the mouse model, we found that the virulence of the SUMOylation-defective PB1 K612R mutant viruses was significantly attenuated compared with that of the wild-type WSN (H1N1), VN/1180 (H5N1) or AH/1 (H7N9) virus, indicating that SUMOylation at K612 of PB1 is necessary for IAVs to be virulent in mice. Sustained respiratory droplet transmission is an essential property of pandemic and epidemic IAVs [35], and we found that the SUMOylation of PB1 at K612 may facilitate the transmission of IAVs. The wild-type 2009 pandemic H1N1 virus, A/Fuzhou/1/2009, transmitted to all three exposed ferrets via respiratory droplets, whereas the PB1-K612R mutant virus transmitted to two of the three exposed ferrets. Notably, the R612K reversion mutation was identified on day 2 p.i., and accounted for 100% of the residues at position 612 of PB1 on day 4 p.i., in all three inoculated ferrets; virus shedding in one of the two positive exposed ferrets was detected until day 7 p.e., and almost all shed viruses carried the R612K reversion mutation. These findings indicate that the stress condition, caused by the SUMOylation-defective PB1/K612R mutation, drives the reversion of the mutant virus to the wild-type phenotype. The SUMOylation at K612 of PB1 would, therefore, appear to be a prerequisite for the efficient replication and transmission of pandemic influenza viruses. Together, these data shed light on the importance of SUMOylation at K612 of PB1 in the pathogenesis and the transmission of IAVs.
In summary, here we found that IAVs can manipulate the host SUMOylation system to modify their PB1 protein. The highly conserved K612 residue of PB1 was pinpointed as the key acceptor site that conjugates with SUMO1. vRNP complex activity and viral replication were severely attenuated in the presence of the SUMOylation-defective PB1/K612R mutation, which were attributed to the impaired binding affinity to vRNA caused by this mutation. Strikingly, the SUMOylation at K612 of PB1 was important for the pathogenesis and transmission of IAVs in mammalian models. Collectively, our study revealed the biological significance of SUMOylation at K612 of PB1 in IAV infection, which could serve as a potential therapeutic target for the development of novel antiviral therapies.

Ethics statements
This study was carried out in strict accordance with the recommendations in the Guide for the

Facility
All experiments with live H5N1 and H7N9 viruses were conducted within the enhanced animal biosafety level 3 (ABSL3+) facility at the HVRI of CAAS, which is approved for such use by the Ministry of Agriculture and Rural Affairs of the People's Republic of China. All animal studies were approved by the Review Board of the HVRI, CAAS. The details of the facility and the biosafety and biosecurity measures used have been previously reported [31,41].

Cells and viruses
HEK293T cells were cultured in DMEM (Sigma-Aldrich) supplemented with 10% fetal bovine serum (Sigma-Aldrich), human lung carcinoma cells (A549) were cultured in F12K medium (Life Technologies) supplemented with 10% FBS, and Madin-Darby canine kidney (MDCK) cells were cultured in MEM (Life Technologies) containing 5% newborn calf serum (Sigma-Aldrich). All cells were maintained in a humidified incubator containing 5% CO 2 at 37˚C.

Generation of mutant viruses by reverse genetics
Mutant viruses with the PB1/K612R mutation (AAA at positions 1858-1860 of PB1 gene were mutated to CGA) in the WSN (H1N1), VN/1180 (H5N1), AH/1 (H7N9) and FZ/1(H1N1) background were generated by using the reverse genetics system as described previously [15]. The rescued viruses were fully sequenced to ensure the absence of unwanted mutations.

Co-IP and western blotting analysis
HEK293T cells were transfected for 48 h with the indicated plasmids by using Lipofectamine 2000 reagent (Invitrogen). To determine the SUMOylation of the PB1 protein during viral infection, HEK293T cells were first transfected for 24 h with Ubc9, SUMO1, or SUMO1 mut , followed by infection with WSN (H1N1) or WSN-PB1 K612R (H1N1), VN/1180 (H5N1) or VN/ 1180-PB1 K612R (H5N1), and AH/1 (H7N9) or AH/1-PB1 K612R (H7N9) mutant virus (MOI = 2) or mock infection for 18 h. The transfected or infected cells were then washed with cold PBS and lysed with NP40 buffer (Beyotime Biotechnology, China) containing complete protease inhibitor cocktails (Roche Diagnostics) for 30 min on ice. The cell lysates were centrifuged at 12,000 rpm at 4˚C for 10 min, and the supernatant was then subjected to immunoprecipitation with anti-Flag agarose at 4˚C for 6 h. The beads were washed four times with cold PBS (containing 1 mM PMSF), and the bound proteins were eluted with SDS loading buffer and separated by SDS-PAGE. The proteins were transferred onto nitrocellulose membranes (GE Healthcare), which were blocked in 5% skimmed milk in PBST for 1 h, and then incubated with the indicated primary and secondary antibodies. Blots were visualized with the Odyssey infrared imaging system (Li-Cor BioScience).

Immunofluorescence assay
A549 cells seeded in glass-bottom dishes were infected with WSN (H1N1) or WSN-PB1 K612R (H1N1) mutant virus at an MOI of 5. At 2, 4, 6, 8, and 10 h p.i., the cells were fixed with 4% paraformaldehyde for 1 h and permeabilized with 0.1% Triton X-100 in PBS for 15 min. The permeabilized cells were blocked with 5% bovine serum albumin in PBS for 1 h, and then were incubated with rabbit anti-PB1 pAb for 1 h. After three washes with PBS, the cells were incubated with the secondary antibody Alexa Fluor 488 donkey anti-rabbit IgG (H+L) for 1 h, followed by incubation with 4',6-diamidino-2-phenylindole (DAPI, Sigma-Aldrich) for 15 min to stain the nuclei. Images were acquired by using a Zeiss LSM880 confocal microscope.

Plaque assay
Plaque assays were performed on MDCK cells in 12-well plates as described previously [43]. Briefly, MDCK cells were infected with 10-fold serial dilutions of virus supernatants in 1×MEM (0.3% BSA) for 1 h at 37˚C. Then the cells were washed with PBS and overlaid with 1% SeaPlaque agarose (Lonza) in 1×MEM (0.3% BSA, 0.5 μg/ml TPCK-treated trypsin). After 48 to 72 h of incubation, the cells were fixed with formalin, and plaques were stained with 0.1% crystal violet and counted.

Generation of model vRNA
A 360-nucleotide model vRNA was transcribed in vitro from a cDNA containing the T7 promoter and the noncoding region of the NS gene of WSN (H1N1), VN/1180 (H5N1), or AH/1 (H7N9) virus by using the RiboMax Large Scale RNA Production Systems (Promega) as described previously [34]. The vRNAs transcribed in vitro were concentrated and purified by using a Spin Column RNA Cleanup & Concentration Kit (Sangon Biotech) and quantified by using NanoDrop (ThermoFisher).

Protein binding vRNA assay
A protein binding vRNA assay was carried out as described previously [34]. Briefly, HEK293T cells were transfected with a plasmid expressing Flag-tagged truncated PB1 494-757 or PB1 494-757/K612R or with pCAGGS-Flag vector as a control. The cells were lysed in RIPA lysis buffer (ThermoFisher) supplemented with protease inhibitor cocktail (Roche). Anti-Flag agarose was added to cell lysates containing 100 μg of total protein, followed by incubation for 4 h on a roller at 4˚C. After three washes with diethyl pyrocarbonate (DEPC)-treated 0.1M NaCl, vRNA (3 μg) transcribed in vitro was added to the beads and incubated on a roller for 4 h at 4˚C. The supernatant containing the unbound vRNA and the beads bearing the proteinbound vRNA were isolated and quantified by means of real-time RT-PCR (primer sequences available upon request).

RNA-protein pull-down assay
The RNA binding protein assay was performed according to the instructions of the Pierce Magnetic RNA-Protein Pull-Down Kit (20164, ThermoFisher). In brief, 10 μg of model vRNA was labeled with a single desthiobiotinylated cytidine bisphosphate at the 3' end by using T4 RNA ligase. Then, the labeled vRNA was captured with streptavidin magnetic beads. After the beads were washed twice with 20 mM Tris (pH7.5) and Protein-RNA binding buffer, cell lysates containing 100 μg of total protein were added to the RNA-bound beads and incubated for 1 h at 4˚C with rotation. The beads containing the biotinylated RNA/protein complexes were collected on a magnetic stand. The supernatants were harvested for subsequent detection and analysis. The beads were then washed three times with washing buffer and boiled in SDS loading buffer. Both the supernatants and the precipitates were resolved by means of 10% SDS-PAGE followed by western blotting.

Ferret study
Four-month-old female ferrets (Wuxi Cay Ferret Farm, China) that were serologically negative for influenza viruses were used in this study. The animals were anesthetized via intramuscular injection with ketamine (20 mg/kg) and xylazine (1 mg/kg). To investigate virus replication, groups of two ferrets were anesthetized and inoculated i.n. with 10 6 PFU of test virus in a 500μl volume (250 μl per nostril). The ferrets were euthanized on day 4 p.i. and their nasal turbinates, soft palate, tonsils, trachea, lung, spleen, kidneys, and brain were collected for virus titration. For the respiratory droplet transmission studies, groups of three ferrets were inoculated i. n. with 10 6 PFU of test virus and housed in specially designed cages inside an isolator as described previously [6,7,46]. Twenty-four hours later, three naïve ferrets were placed in an adjacent cage (4 cm away), separated by a double-layered net divider. Nasal washes were collected from each of the virus-infected ferrets on days 2, 4, 6, and 8 p.i. and from each of the exposed ferrets on days 1, 3, 5, 7, 9, and 11 p.e., and were titrated by use of plaque assays on MDCK cells. Sera were collected from all animals on day 21 p.i. or p.e. for HI antibody detection.

Deep sequencing
Viral RNA was extracted from the stock virus, nasal washes, or lung homogenates by using the QIAmp viral RNA mini kit (QIAGEN) and was reverse-transcribed into cDNA by use of Uni12 primer (5'-AGCRAAAGCAGG-3'). The genome of the viruses was amplified and sequenced as described previously [31]. Briefly, next generation sequencing libraries were constructed by using the NEBNext Ultra DNA Library Prep Kit for Illumina. For each sample, 1 μg of DNA was randomly fragmented to <500 bp by sonication. The fragments were then treated with End Repair Mix for end repairing and with A-Tailing Mix for dA-tailing, followed by a T-A ligation to add adaptors to both ends. Each sample was then amplified by PCR using P5 and P7 primers. The PCR products were cleaned and quantified by using a Qubit2.0 Fluorometer (Invitrogen). Then, libraries with different indices were multiplexed and loaded onto an Illumina HiSeq instrument. Sequencing was carried out using a 2×150 paired-end configuration; image analysis and base calling were conducted by the HiSeq Control Software (HCS) + OLB + GAPipeline-1.6 (Illumina) on the HiSeq instrument. The genomes of all of the samples sequenced yielded more than 5000-fold genome coverage depth (with raw sequencing data of approximately 2.0 Gb obtained per sample). The sequences were processed and analyzed by GENEWIZ (GENEWIZ Biotechnology).

Statistical analysis
Quantitative data are presented as means ± SD of at least three biological replicates. Data were statistically analyzed with a two-tailed unpaired Student's t-test or a one-way ANOVA followed by t-test by using GraphPad Prism 7.0 software. Statistical parameters are reported in the figures and figure legends. P values < 0.05 were considered statistically significant.