Expanding the tolerance of segmented Influenza A Virus genome using a balance compensation strategy

Reporter viruses provide powerful tools for both basic and applied virology studies, however, the creation and exploitation of reporter influenza A viruses (IAVs) have been hindered by the limited tolerance of the segmented genome to exogenous modifications. Interestingly, our previous study has demonstrated the underlying mechanism that foreign insertions reduce the replication/transcription capacity of the modified segment, impairing the delicate balance among the multiple segments during IAV infection. In the present study, we developed a “balance compensation” strategy by incorporating additional compensatory mutations during initial construction of recombinant IAVs to expand the tolerance of IAV genome. As a proof of concept, promoter-enhancing mutations were introduced within the modified segment to rectify the segments imbalance of a reporter influenza PR8-NS-Gluc virus, while directed optimization of the recombinant IAV was successfully achieved. Further, we generated recombinant IAVs expressing a much larger firefly luciferase (Fluc) by coupling with a much stronger compensatory enhancement, and established robust Fluc-based live-imaging mouse models of IAV infection. Our strategy feasibly expands the tolerance for foreign gene insertions in the segmented IAV genome, which opens up better opportunities to develop more versatile reporter IAVs as well as live attenuated influenza virus-based vaccines for other important human pathogens.


Introduction
The balance of the multiple segments during IAV infection is vulnerable to artificial modifications. For example, it was reported that specific mutations in the promoters can enhance the levels of vRNA and mRNA [18,19], however, upon incorporation of these promoter-upregulating mutations, the recombinant IAV attenuates in replication, highlighting the critical need for the delicate balance between replication, transcription and protein expression of IAV segments [19]. In addition, many other factors including the segment-specific non-coding regions (NCRs), the length of coding regions, as well as the inherent activity and template preference of viral RdRp are involved in the competition [16,17,20].
In our previous report, we demonstrated that foreign insertions can drastically reduce the replication/transcription capacity of the modified segment due to inevitable increase in segment length, adversely impacting the balance of the multiple segments in the levels of vRNA and mRNA during IAV infection [21]. For example, when the imbalance occurred to the NS segment: First, inadequate vRNAs of modified NS are available for incorporation into progeny virions, producing a large amount of NS-null noninfectious particles [21]; Second, the expression of NS1 proteins decreased accordingly and may no longer counteract the host immune responses efficiently [22]; Third, the accumulation of NEPs which play a pivotal role to mediate nuclear export of viral ribonucleoproteins was greatly affected [23,24]. Together these mechanisms lead to a reduced replication and attenuation of the reporter IAVs.
In this study, we propose that incorporation of compensatory enhancing mutations during initial construction of a reporter IAV can rebalance the multiple segments, and subsequently reduce or eliminate the mechanism of attenuation. As a proof of concept, a reporter influenza PR8-NS-Gluc virus which carries a Gaussia luciferase (Gluc) gene fusing to NS1 was used as a starting point [25], and the aforementioned enhancing mutations were introduced into the promoter elements of the modified NS-Gluc segment to rectify the imbalance. We show that an optimized reporter IAV that shows improved replication kinetics and virulence was generated. Moreover, the tolerance of IAV genome can be further expanded using this "balance compensation" strategy. In conjugation with stronger compensatory enhancements, a recombinant IAV carrying a much larger firefly luciferase (Fluc) gene was subsequently generated.

Evaluation of replication-enhancing mutations of IAV vRNA segments
Diverse structures could be formed from the 3'-and 5'-terminal nucleotides of influenza vRNA segments, including cockscrew, panhandle, fork and hook-like structures, dynamically regulating the transcription and replication of vRNA during virus life cycle [26]. Interestingly, it has been well documented that the panhandle structure is involved in the initiation of replication/transcription, while panhandle-stabilizing mutations within the conserved 3'-NCR can enhance the promoter activity [16,18]. These panhandle stabilizing mutations have potential to be employed as compensatory enhancements for reporter IAVs.
We first evaluated how panhandle-stabilizing mutations enhanced vRNA replication, transcription and subsequent protein expression, especially in competition with its natural counterpart. To this end, a dual-template reporter RdRp assay was employed. In brief, the firefly and renilla luciferase (Fluc and Rluc) genes were flanked by 3'-and 5'-NCRs of the NS segment and inserted in a negative sense orientation under the control of the human RNA polymerase I (Pol-I) promoter, generating IAV minigene expressing plasmids pNCR NS -Fluc and pNCR NS -Rluc, respectively. Mutations of interest were then introduced into pNCR NS -Fluc, producing various mutants (pNCRx-Fluc), while pNCR NS -Rluc was set as a constant competitive control. When pNCRx-Fluc and pNCR NS -Rluc were co-transfected into 293T cells expressing IAV RdRp components and NP, mini vRNA reporters NCRx-Fluc and NCR NS -Rluc were expected to compete with each other to bind RdRp complex. We thus measured the relative expression of vRNA and mRNA from the dual-templates by qRT-PCR as a measure of the replication and transcription efficacy of mutated versus wildtype vRNA promoter. Fig 1A shows a schematic overview of the dual luciferase reporter-based competition assay.
As described previously, three pan-handle stabilizing mutations at 3'-NCR (G3A, U5C, C8U) and the varied combinations were introduced into NCR NS -Fluc separately by PCR using primers listed in S1 Table, and dual competition assay was conducted using minigenome NCR NS -Rluc as a competitive control ( Fig 1B) [16,18]. As shown in Fig 1C, double mutations G3A/C8U could significantly enhance both vRNA replication and transcription, while triple mutations G3A/U5C/C8U exhibited stronger enhancement, providing two sets of potent compensatory enhancement (CE). Here we will refer the G3A/C8U and G3A/U5C/C8U mutations as CE1 and CE2, respectively.

PLOS PATHOGENS
Balance compensation expands the tolerance of segmented IAV genome indicating that paired G3/C8 or A3/U8 is essential for the formation of transcription favored structures, e.g., the corkscrew like structure (Fig 1B).

Compensatory promoter enhancements can rebalance vRNA segments of reporter IAV and restore the wildtype-like fitness in-vitro and in-vivo
To determine whether the compensatory enhancements can restore the replication kinetics and virulence of the reporter influenza PR8-NS-Gluc virus, the two sets of promoter-upregulating mutations were incorporated into the modified NS-Gluc segment (Fig 2A) [25]. Through a modified dual-template RdRp assay with wildtype M segment as a competitive control template, the expression of the mutated NS-Gluc vRNA was compared to the original NS-Gluc and natural NS by qRT-PCR analysis. As anticipated, both sets of mutations CE1 and CE2 significantly increased the replication of NS-Gluc vRNA ( Fig 2B). Moreover, it is noteworthy that segment NS CE1 -Gluc which carries G3A/C8U mutations could replicate similarly in efficacy to natural NS, while segment NS CE2 -Gluc carrying G3A/U5C/C8U mutations replicated more efficiently. Similar compensatory enhancement was observed for vRNA transcription (S1 Fig).
Next, reporter influenza PR8-NS-Gluc viruses harboring additional CE1 or CE2 mutations at 3'-NCR of the modified NS-Gluc segment were rescued using reverse genetics system and were designated as PR8-NS CE1 -Gluc and PR8-NS CE2 -Gluc, respectively. Interestingly, the PR8-NS CE1 -Gluc virus showed improved replication kinetics as compared to original

PLOS PATHOGENS
Balance compensation expands the tolerance of segmented IAV genome PR8-NS-Gluc, although decreased replication was still observed compared to the parental wildtype PR8 virus (PR8-WT); however, PR8-NS CE2 -Gluc replicated even less efficiently than the original PR8-NS-Gluc virus ( Fig 2C). Together, these results suggest that proper compensation to the modified segment can restore the wildtype-like in-vitro replication kinetics of reporter IAVs.
To determine whether CE1 could rescue the attenuated virulence of reporter PR8-NS-Gluc virus, the lethality of PR8-NS CE1 -Gluc to mice was assessed alongside with PR8-WT and PR8-NS-Gluc viruses as controls. At a dose of 100 TCID 50 , PR8-WT infection caused rapid body weight loss of mice, while delayed and mild body weight loss were observed for PR8-NS-Gluc infected mice, indicating drastic attenuation of PR8-NS-Gluc virus ( Fig 2D). Interestingly, PR8-NS CE1 -Gluc virus induced more rapid and severe body weight loss compared to the original PR8-NS-Gluc virus, although it was still not as virulent as the parental PR8-WT virus (Fig 2D). Similar enhancement could also be reflected by the lethality data, as PR8-NS-Gluc virus was sublethal at indicated dose, PR8-NS CE1 -Gluc caused 60% lethality ( Fig 2E). These data clearly demonstrate that the in vivo fitness of reporter IAVs can be, at least partially, restored by rebalancing the levels of modified genome segments.
Our previous study has demonstrated that the ratio of genome copies to infective virus of PR8-NS-Gluc is much higher than the parental PR8-WT virus [21], suggesting an increased production of defective interfering (DI) particles. We therefore assessed the effect of CE incorporation on DI particles production. Consistently, the genome copies to infective virus ratio of PR8-NS-Gluc was about 11.6-folds compared to PR8-WT, while the PR8-NS CE1 -Gluc virus showed a 2.5-folds increase only, which was not statistically significant ( Fig 2F). These results suggest that CE1 incorporation can reduce the abnormal production of DI particles caused by foreign gene insertion. However, the PR8-NS CE2 -Gluc virus produced 11.4-folds more DI particles than PR8-WT virus (Fig 2F), which is in accordance with the overcompensation of CE2 incorporation to genome balance and the attenuated replication of PR8-NS CE2 -Gluc virus (Fig 2B and 2C).

Compensatory enhancement augments reporter protein expression
Considering CE1 could restore the replication capacity of the modified segment NS-Gluc, we reasoned that it was likely that the reporter protein expression could also increase correspondingly. To validate this, the kinetics of Gluc expression during PR8-NS CE1 -Gluc replication was monitored ( Fig 3A). Further, the relative reporter protein expression level was determined by normalizing the bioluminescence signal to viral titer. As expected, Gluc expression in the PR8-NS CE1 -Gluc infected cells was significantly augmented by more than 10-folds ( Fig 3B).
The reporter influenza PR8-NS CE1 -Gluc was next adapted for high-throughput antiviral screening. The signal-to-noise (S/N) ratio with PR8-NS CE1 -Gluc was boosted from 24 to 1109 compared to the original PR8-NS-Gluc based screening, while the coefficient of variation (CV) and Z' score were not affected ( Fig 3C). Considering S/N ratio is one of the most important factors of quality control for high-throughput screens, and a higher S/N value usually means higher sensitivity, the optimized PR8-NS CE1 -Gluc virus described here can be used as a powerful platform for antiviral screening.
In-vivo bioluminescence imaging (BLI) of viral infection provides a valuable tool to investigate virus-host interactions and evaluate antiviral therapies. Our previous study revealed that PR8-NS-Gluc infected mice failed for in-vivo imaging, likely due to the combined effects of low expression level of Gluc and low transmission light generated by Gluc through tissues. Since PR8-NS CE1 -Gluc exhibited enhanced virulence and augmented Gluc expression compared to the original PR8-NS-Gluc, we asked whether PR8-NS CE1 -Gluc virus could be used for in-vivo imaging in mice. To test this, female BALB/c mice (4 weeks old) were infected with 100 TCID 50 of influenza PR8-NS CE1 -Gluc and PR8-NS-Gluc separately, and the kinetics of Gluc expression in the lung tissues were monitored. As shown in Fig 3D, although similar trend of rising and declining of Gluc expression were observed in both groups, which reflected similar kinetics of virus spreading and clearance, the bioluminescence level in lung tissues from the PR8-NS CE1 -Gluc infected mice was roughly 20-fold higher than those from the PR8-NS-Gluc infected mice throughout the course of infection. Moreover, ex-vivo imaging revealed similar bioluminescence enhancement of PR8-NS CE1 -Gluc compared to PR8-NS-Gluc (Fig 3E and  3F). However, live imaging in the PR8-NS CE1 -Gluc infected mice was not successful after several rounds of attempt.

Development of a recombinant IAV carrying a large firefly luciferase gene
Considering our success in directed optimization of the reporter influenza PR8-NS-Gluc virus, we asked whether the "balance compensation" strategy would allow IAV genome to tolerate

PLOS PATHOGENS
Balance compensation expands the tolerance of segmented IAV genome larger foreign insertions, e.g., the Fluc gene. Fluc possesses unique advantages in live imaging, including the relative ease of substrate administration (intraperitoneal injection), longer duration of signal, low cost, and more importantly, the longer peak emission wavelength (~612 nm), which contributes to its high transmission ability through tissues, providing a high level of sensitivity. However, since Fluc is large in size (~1,650 bp) and the IAV genome might not tolerate such a large insertion.
In this study, we replaced the Gluc gene of the modified NS-Gluc segment with Fluc, generating NS-Fluc. An additional 2A protease sequence was inserted between NS1 and Fluc to avoid potential interfere of Fluc, due to its larger size, to the function of NS1 (Fig 4A) [27]. In addition, NS-Fluc constructs incorporating CE1 or CE2 mutations at 3'-NCR were also generated, designated as NS CE1 -Fluc and NS CE2 -Fluc, respectively. The replication capacities of the NS-Fluc and two mutants were determined using dual-template RdRp assay. As shown in Fig 4B, the replication of NS-Fluc was greatly impaired, with a signal roughly 0.2% of the parental NS, which was even worse than NS-Gluc (10% of the parental NS). In contrast, both CE1 and CE2 mutations at 3'-NCR improved the replication efficacy of vRNA NS-Fluc, by 3.8 and 9.4-fold, respectively.
Next, we tried to incorporate the original NS-Fluc and two mutants into recombinant IAVs by reverse genetics. Not surprisingly, no recombinant IAV harboring the original NS-Fluc segment was generated, because the replication capacity of NS-Fluc segment was severely compromised. In contrast, both recombinant IAVs harboring NS CE1 -Fluc or NS CE2 -Fluc were rescued successfully. While the NS CE1 -Fluc segment appeared to be unstable since the signal of exogenous reporter was lost rapidly within three passages when propagated in chicken embryo, the PR8-NS CE2 -Fluc virus exhibited good stability for at least five passages ( Fig 4C  and S2 Fig). Interestingly, deep sequencing of the fifth passage of PR8-NS CE2 -Fluc virus identified a U1G mutation with 20% frequency at the 3'-NCR of NA segment (S3 Fig).
As shown in Fig 4D, PR8-NS CE2 -Fluc replicated efficiently in MDCK cells, although the peak titer was 1-log lower than the parental PR8-WT virus. In a mouse model, PR8-NS CE2 -Fluc was shown to be infectious and pathogenic, causing rapid body weight loss and leading to 100% lethality at high challenge doses (Fig 4E).

Establishment of a robust live imaging animal model of IAV infection
We next evaluated the potential use of PR8-NS CE2 -Fluc in a live imaging mouse model of IAV infection. A group of mice were inoculated intranasally with PR8-NS CE2 -Fluc at a sub-lethal dose (1,000 TCID 50 ), and the real-time bioluminescence was performed every day for nine days to monitor temporal changes in virus replication and tissue distribution over time ( Fig 5A). It could be observed that viral infection was initiated in nasal-related tissues and both lobes of the lungs, as previously described [6,9,28]. The dynamics of IAV infection was also clearly delineated. As shown in Fig 5B, the BLI signal could be detected as early as day 1 post infection (p.i.), then elevated and peaks at day 2 or 3 p.i., while the signal decreased greatly after 6 days p.i., indicating virus clearance. In a separate experiment, cohorts of mice were imaged and necropsied at 1, 3, 5, 7, 9 days p.i. (3 mice per time point), and the high correlation between BLI signal intensity and virus load in lungs was observed (r 2 = 0.68, p < 0.001; Fig 5C), indicating that BLI measure can accurately reflect the virus replication in vivo.
To further test if our BLI-based mouse model of IAV infection could be used to evaluate efficacy of antiviral therapeutics, the mice were administrated with 10 or 30 mg/kg/day oseltamivir phosphate per oral twice a day starting at 2 hours prior to virus challenge. Mice treated with vehicles only were used as controls. BLI was acquired at day 2 and 5 p.i.. Both doses of oseltamivir reduced the signal intensities greatly (Fig 5D and 5E). These results demonstrate

PLOS PATHOGENS
Balance compensation expands the tolerance of segmented IAV genome

PLOS PATHOGENS
Balance compensation expands the tolerance of segmented IAV genome

PLOS PATHOGENS
Balance compensation expands the tolerance of segmented IAV genome that our PR8-NS CE2 -Fluc based live-imaging mouse model accurately reflects the in vivo efficacy of oseltamivir.

Development of a live imaging animal model of H3N2 IAV infection
To investigate whether we could develop a similar live imaging mouse model with another IAV strain, we developed and evaluated an A/X31(H3N2)-based live imaging strain. The X31 virus is a reassortant virus harboring six internal segments from A/PR8 as backbone and the HA and NA segments from a mouse adapted virus of the H3N2 Hong Kong epidemic strain, and it has been broadly used to establish mouse models of H3N2 infection [29].
First, we generated a reporter X31-NS CE2 -Fluc virus containing the six internal segments from the PR8-NS CE2 -Fluc as backbone by reverse genetics. Subsequently, the X31-NS CE2 -Fluc virus was characterized both in vitro and in vivo. Predictably, the in vitro replication of X31-NS CE2 -Fluc was reduced compared to wildtype X31 virus, with peak virus tier of more than 2-log lower (Fig 6A). In addition, the X31-NS CE2 -Fluc virus was attenuated in mice compared to the X31 virus ( Fig 6B). Nonetheless, the X31-NS CE2 -Fluc virus exhibited similar lethality as the recombinant PR8-NS CE2 -Fluc virus.
The live imaging of X31-NS CE2 -Fluc in mice was evaluated. BALB/c mice were grouped and inoculated intranasally with X31-NS CE2 -Fluc at a series of doses from 10 2 to 10 5 TCID 50 , followed by monitoring real-time bioluminescence at day 1, 2, 3, 5 and 8 post challenge ( Fig 6C). As shown in Fig 6D, the kinetics of BLI signal clearly showed the initiation, spreading

PLOS PATHOGENS
Balance compensation expands the tolerance of segmented IAV genome and clearance of virus infections. Moreover, the signals at day 2 from different groups correlated well to the challenge doses (Fig 6D), suggesting this IAV strain can be adapted to be used for live imaging.

Discussion
The limited tolerance of IAV genome to foreign gene insertion has hampered the development of recombinant IAVs as tools, e.g, reporter viruses [5,6,30], and live-attenuated influenza virus-vectored vaccines bearing foreign antigens [31][32][33]. Previously, we have unraveled an underlying mechanism that foreign insertions may cause reduced transcription/replication capacity of the modified vRNA segment in competition with other wildtype segments, impairing the balance of the segmented genome during infection. Thus, the balanced profiles of the eight segments at the levels of vRNA, mRNA and protein expression are all adversely affected, leading to defective genome packaging into progeny virions [21]. As a consequence, the replication and virulence of recombinant IAVs are attenuated, and the foreign insertions may be lost rapidly during virus passaging. In the present study, we developed a "balance compensation" strategy to expand the tolerance of the segmented IAV genome to facilitate genome manipulations. As depicted in Fig 7, directed optimization of recombinant IAVs is achieved by incorporating compensatory enhancements (CEs) to rectify the reduced replication of modified segment and genome imbalance.
In this study, a well-studied set of replication-enhancing mutations CE1 at the promoter element was introduced into the modified NS-Gluc segment of the reporter influenza PR8-NS-Gluc virus, and it was shown that CE1 could correct the reduced replication/transcription of modified NS-Gluc segment. Further, CE1 incorporation not only restored the wildtype-like fitness of the reporter virus (see Fig 2), but also significantly enhanced the reporter Gluc expression (see Fig 3). Importantly, we successfully generated stable replicationcompetent recombinant IAVs integrating a much larger Fluc gene coupled with a much stronger set of replication-enhancing mutations CE2 (see Figs 4 and 6).
Our work demonstrates there is delicate balance of the segmented IAV genome for genome manipulations. For instance, since CE1 can compensate the reduced replication/transcription efficacy of NS-Gluc to the level of natural NS, and CE2 leads to overcompensation (see Fig 2B), it seems reasonable to assume that CE1 but not CE2 was more suitable for directed optimization of the recombinant PR8-NS-Gluc virus. However, since Fluc is much larger and its insertion to the genome could impair the balance of segmented genome more drastically, the CE2 but not CE1 was more optimal for generating a more stable recombinant PR8-NS-Fluc virus (see Fig 4). It is noteworthy that CE2 is likely not optimal enough to compensate the reduced replication of NS-Fluc segment, and additional compensation(s) is likely required to generate more stable IAVs (see Fig 4B). In addition, since PR8 strain was laboratory-adapted and likely harbors increased tolerance for genome rearrangement, generation of recombinant IAVs from clinical isolates may require more optimization for genome balance.
Our "balance compensation" strategy, described here, is also segment dependent. Besides NS, other segments including PB1, PB2, PA, NA and NP have also been engineered to create reporter IAVs [5][6][7]30,34]. Considering the promoter sequences is highly conserved among all segments, the aforementioned promoter enhancing mutations should also apply to optimize these reporter viruses. In addition, many segment-specific enhancement candidates should be considered. First, as a U/C polymorphism at position 4 exists within the 3'-NCR of IAV segments, U4 contributes to a higher transcription/replication capacity compared to C4 [35]. For those segments that carry original C4 at the 3'-NCR, a C4 to U4 mutation should be employed as compensatory enhancement. Second, the segment-specific NCR sequences are also involved

PLOS PATHOGENS
Balance compensation expands the tolerance of segmented IAV genome in vRNA transcription/replication [36][37][38]. For example, it was reported that a U13 to C13 mutation in the 3 0 end of the NA gene promoted the expression of viral RNA and protein, while mutation of other sites within the UTR could differentially regulate viral genomic transcription and translation [36]. We speculate that a U13 to C13 mutation at the variable NCR sequence of the modified NA would compensate to the attenuation [7]. Third, since the accumulation of vRNA and mRNA during IAV infection is dynamic and segment-specific, both the inherent activity and template preference of viral RdRp may be involved in the regulation [17,20]. As mentioned above, Katsura et al. generated a mouse-adapted reporter influenza Venus-PR8 virus carrying a Venus gene within NS segment, and identified a PB2-E712D substitution that could stabilize the foreign gene insertion and restore wildtype-like replicative ability and virulence in mice [10,22]. Mechanistic studies revealed that the polymerase fidelity was not affected by PB2-E712D substitution [22], and the inherent polymerase activity of For directed optimization of the reporter IAV, proper compensatory enhancement was incorporated during initial construction. The replication/ transcription of NS-derived vRNA was specifically increased, while rebalance of the segmented genome could be achieved, restoring the wildtype-like fitness. Of note, the multiple segments in wildtype PR8 virus infected cells were shown in equal molar ratio for conceptual illustration only. https://doi.org/10.1371/journal.ppat.1010756.g007

PLOS PATHOGENS
Balance compensation expands the tolerance of segmented IAV genome PB2-E712D is even lower than that of wildtype PB2 [10]. Notably, considering the inherent polymerase activity was determined using an NP-derived template [10], and it was demonstrated that PB2-E712D enhanced the transcription/replication efficiency of the modified NS as compared to NP [22], we speculate that the mutated polymerase may possess an increased preference to NS segment over NP. Further studies are needed to investigate the preference profile of the mutated polymerase to the eight segments. It is possible that these enhancement candidates can be used independently or in combination to achieve appropriate compensation, i.e., to rebalance the segmented genome.
In summary, we have developed a "balance compensation" strategy for generation of reporter IAVs. Our strategy allows us to further expand the tolerance of IAV genome to foreign insertions. The success of the present study not only provides diverse valuable reporter IAVs and robust live-imaging mouse models of IAV infection, but also encourages generation of novel reporter viruses with more versatile capabilities, e.g., the bi-or tri-reporter viruses that express two or three foreign reporters from different segments [39]. Moreover, the feasibility of our strategy opens up better opportunities to develop live attenuated influenza virusvectored vaccines for other highly pathogenic viruses and bacteria.

Ethics statement
All animal experiments within this study were approved by the Institutional Animal Care and Use Committee (IACUC) of Shandong University of Traditional Chinese Medicine (Approval: SDUTCM20211230001).

Plasmids
(i) Construction of the IAV minigene expressing plasmids. The plasmids expressing NSderived reporter vRNAs were constructed by replacing the NS1-NEP ORFs with Fluc or Rluc encoding ORFs, under the control of the human RNA polymerase I (Pol-I) promoter. The NCR NS -Fluc fragment was amplified with primers NCR NS -Fluc-Forward/NCR NS -Fluc-Reverse using the pISRE-Luc plasmid as template, while the NCR NS -Rluc fragment was amplified with primers NCR NS -Rluc-Forward/NCR NS -Rluc-Reverse from pRL-TK plasmid. Both NCR NS -Fluc and NCR NS -Rluc fragment were then cloned into the Sap I linearized pPol-I vector using In-fusion cloning kit (Takara, Beijing, China) following the manufacturer's protocol, generating pNCR NS -Fluc and pNCR NS -Rluc respectively.
Next, NCR NS -Fluc fragments carrying panhandle-stabilizing mutations were amplified using indicated primer NCR NS -x-Forward paired with NCR NS -Fluc-Reverse from pNCR NS -Fluc template, and then cloned into pPol-I as described above, generating pNCR NS -x-Fluc mutants.
(iii) Construction of plasmids expressing wildtype and mutated NS-Fluc vRNAs. The modified NS-Fluc was constructed using similar strategy for construction of NS-Gluc, except that an additional 2A protease sequence was inserted between NS1 and Fluc to avoid fusion. The NS-Fluc construct was divided to 3 fragments for initial amplification. The left 3'-NCR-NS1 and right NEP-5'-NCR fragment were amplified using pDZ-NS-Gluc as template with primers 3'-NCR-Forward/NS1-2A-Reverse and 2A-NEP-Forward/5'-NCR-Reverse, respectively. The middle 2A-Fluc fragment was amplified using primers 2A-Fluc-Forward/Fluc-2A-Reverse from the template pNCRns-Fluc. Adjacent fragments overlap for at least 15 nts. All the three fragments and the Sap I linearized pPol-I vector were then ligated using the In-fusion cloning kit, generating wildtype pPolI-NS-Fluc plasmid.

The dual-template RdRp assays
The dual-template reporter RdRp assay was conducted as previously described with slight modifications [16]. Briefly, the minigene expressing plasmids pNCR NS -Fluc (or its mutants pNCRx-Fluc) and pNCR NS -Rluc were co-transfected into 293T cells with IAV RdRp constituent expressing plasmids pFlu-NP, pFlu-PB1, pFlu-PB2, and pFlu-PA using Lipofectamine 2000 (Invitrogen, USA) according to manufacturer's instructions. At 24 h.p.t., the cells were harvested, and a proportion of the cells were removed for luciferase assays using Dual-Glo Luciferase Assay System (Promega, Madison, WI USA), while the left cells were extracted for total RNAs using Simply P Total RNA Extraction Kit (Bioflux, Zhejiang, China). The vRNA and mRNA were reverse transcribed using PrimeScript RT reagent Kit with gDNA Eraser (Takara, China) with the NS-specific primer RT-vRNA-NS and oligo(dT), respectively, followed by quantitative-PCR using Fluc-specific and Rluc-specific qPCR primers. The relative level of Fluc were normalized to Rluc for vRNA, mRNA as well as luciferase activity were calculated to reflect the efficacy of vRNA replication, transcription and protein expression of indicated minigene constructs.
Alternatively, the natural or modified NS segments expressing plasmids were separately cotransfected with natural M-expressing plasmid (pPolI-M) into 293T cells expressing RdRp constituents. After 24h incubation, the cells were harvested for total RNA extraction. The vRNA and mRNA were reverse transcribed as described above, except for vRNA M were reverse transcribed using a universal 3'NCR primer RT-3'NCR, followed by qPCR using NSspecific and M-specific qPCR-primers. The vRNA and mRNA level of NS-derived segments were normalized to those of M to reflect their replication and transcription efficacy respectively.
The primers used for reverse transcription and qPCR were shown in supplementary S2 Table. In vitro growth curves MDCK cells growing in 6 well plates were infected by indicated viruses at a multiplicity of infection (MOI) of 0.01 TCID 50 /cell. After 1 h incubation at 37˚C, cells were washed and fresh Opti-MEM containing 2 μg/ml TPCK-trypsin were added. Aliquots were removed at various time points for viral titration and luciferase assays.

Luciferase assays
To determine the activity of Gluc, 50 μL of viral culture medium or lung tissue homogenate (appropriate dilution applied to avoid over range) were mixed with 50 μL of luciferase substrate using Pierce Gaussia Luciferase Flash Assay Kit (Thermo Scientific, Rockford, IL, USA) according to the manufacturer's instructions. The luminescence was detected immediately using Sirius L Tube Luminometer (Berthold, Germany).
Fluc assays were performed using a Britelite plus Reporter Gene Assay System (PerkinElmer, Waltham, MA, USA) according to the manufacturer's instructions. In brief, MDCK cells growing in 96 well plates were infected by indicated viruses at an MOI of 0.01. After 1 h incubation at 37˚C, cells were washed and fresh Opti-MEM containing 2 μg/ml TPCK-trypsin were added. At 24 h post infection (p.i.), the culture medium was discarded, followed by sequentially adding 50 μL PBS and 50 μL substrate. After incubation for 10min, the luminescence was detected immediately using BioTek SYNERGY neo2 Microplate Reader (BioTek, Winooski, VT, USA).

Genome stability analysis
The indicated recombinant IAVs were serially passaged in chicken embryos for at least five passages. The viruses of each passage were tittered and used to infect MDCK cells grown in white 96-well plates (10,000 cells/well) at an MOI of 0.1. The cells were harvested at 24 h.p.i. for luciferase assays.

PLOS PATHOGENS
Balance compensation expands the tolerance of segmented IAV genome

Animal models
Female BALB/c mice (4 to 6 weeks old) were used in this study. All animals were maintained under specific pathogen-free conditions and all efforts were made to minimize any suffering and the number of animals.
To determine the lethality of the viruses, five mice from each group were inoculated intranasally under isoflurane anesthesia with 10-fold serial dilutions containing 10 0 to 10 5 TCID 50 (30 μl) of virus. Body weight and survival were monitored daily for 14 days.
To measure virus replication in mice, three to six mice in each group were inoculated intranasally under isoflurane anesthesia with indicated sublethal dose of viruses. At indicated time points, mice were subjected to in vivo imaging, ex-vivo imaging or determination of viral load/ luciferase activities in lung tissues.
For antiviral treatments, 10-30 mg/kg/day of oseltamivir phosphate or vehicle only (PBS) were administered via intraperitoneal (i.p.) injection. The treatments were given twice daily for 5 days starting at 2 h before virus inoculation.

Ex-vivo imaging
Mice infected with sublethal doses of PR8-NS-Gluc, PR8-NS CE1 -Gluc or mock infected were euthanized on day 3 p.i. and the trachea and lungs were excised. A syringe needle was inserted into the opening of the trachea and 0.5 ml of coelenterazine (50 μg/ml, NanoLight Technology, AZ, USA) was injected into the lung, followed by imaging immediately using IVIS200.

In-vivo imaging
Mice infected with varies doses of PR8-NS CE2 -Fluc or X31-NS CE2 -Fluc were anaesthetized and the substrate D-Luciferin (PerkinElmer, Waltham, MA, USA) was injected intraperitoneally at 150 mg/kg. At 10 min after substrate administration, images were acquired with the Xenogen IVIS 200 and analyzed using the Living Image software (version 4.4).
To measure virus replication, live imaging was conducted daily, while for antiviral determination, the imaging were conducted on days 2 and 5 p.i.

Statistical analysis
For PR8-NS CE1 -Gluc based high-throughput screening approach, the quality was assessed by evaluation of the signal-to-noise (S/N) ratio, coefficient of variation (CV) and Z' factors. (1) S/ N = mean signal of negative control / mean signal of positive control; (2) CV = SD of negative control / mean of negative control; (3) Z' = 1-3 × (SD of positive control + SD of negative control) / (mean of negative control-mean of positive control). SD represents the standard deviation. A Z' value between 0.5 and 1.0 is considered robust enough for an HTS assay.
Statistical significance was determined using unpaired Student's t-test with two-tailed analysis and the GraphPad Prism 5 software package (GraphPad Software). Data are considered significant when P values are<0.05.  China). The complementary DNA was prepared using PrimeScript RT reagent Kit with gDNA Eraser (Takara, China) and NS segment specific primer (5'-CAGGGTGACAAAGACATAATG-3'). Then PCR analysis was performed using the 2xTaq MasterMix (Cwbio, China) and primers covering the full length of firefly luciferase gene (NS-Fluc-specific-Forward:5'-ACGTCGAGGAGAATCC CGGGCCCATGGAAGACGCCAAAAA-3'; NS-Fluc-specific-Reverse: CAGGCTAAAG TTGGTCGCGCCGCTGCCCAATTTGGACTTT). The PCR product was analyzed using 1% agarose gel electrophoresis. The plasmids pPolI-NS CE1 -Fluc and pPolI-NS CE2 -Fluc were used as positive controls, while pPolI-NS was used as the negative control.