2. Referentes teóricos
2.1.6 Modelo de análisis Audy Salcedo (2015)
To confirm the importance of NS1 evolutionary markers K186 and C-terminal truncation, for efficient viral replication in the presence of type I IFN, E.Derm cells were pre-treated with Universal IFN for 24 h prior to infection (0.1 FFU/cell). As done previously, viral replication and percentage of infected cells were monitored for 72 h (Chapter 3, section 3.5.7 and 3.5.8).
While all viruses exhibited a reduction in viral growth (Figure 5-16-A), only the WT virus managed to infect an increasing number of cells (Figure 5-16-A) and produce an increasing quantity of infectious particles (Figure 5-16-B) in IFN-primed cells.
Figure 5-16: Growth kinetics of O/03 WT and NS1 mutant viruses in universal IFN-primed equine cells.
E.Derm cells were treated for 24h with universal IFN prior to infection (0.1 FFU/cell) with O/03-WT, - K186E, -230, and -K186E-230. Cells and supernatants were collected at different times post-infection. (A) Percentage of infected cells and cell viability was determined by flow cytometry. (B) Viral titres were measured as described in Methods. Bars correspond to mean of three independent experiments and error bars represent SEM. Significance was calculated as for Figure 5-1. ****, p<0.0001 for indicated virus at indicated time post infection versus all other conditions at the same time post infection.
4h 8h 12h 16h 24h 48h 72h 100 101 102 103 104 105 106 107 108 109 1010 1011 4h 8h 12h 16h 24h 48h 72h 0 20 40 60 80 100 4h 8h 12h 16h 24h 48h 72h 0 20 40 60 80 100
A
B
N P s ta in e d c e lls (% )Time post-infection (hpi) Time post-infection (hpi)
C e ll v ia b ili ty (% ) PF U /m l (L o g10 )
Time post-infection (hpi)
inoculum **** ** **** **** O/03-230 O/03-K186E-230 4h 8h 12h 16h 24h 48h 72h 100 101 102 103 104 105 106 107 108 109 1010 1011 O/03 WT O/03 K186E O/03 230 O/03 K186E 230 4h 8h 12h 16h 24h 48h 72h 100 101 102 103 104 105 106 107 108 109 1010 1011 O/03 WT O/03 K186E O/03 230 O/03 K186E 230 O/03-WT O/03-K186E 4h 8h 12h 16h 24h 48h 72h 0 20 40 60 80 100 O/03 WT O/03 K186E O/03 230 O/03 K186E 230 Mock Mock
5.3. Discussion
In this chapter, the effect of NS1 on EIV infection phenotype was evaluated. First, it was demonstrated that the NS segment, encoding the NS1 protein, affected EIV replication efficiency and cell-to-cell spread in mammalian cells. To do so, viral growth kinetics and plaque phenotype of WT and NS reassortant of U/63 and O/03 viruses were compared in MDCK and E.Derm cells. It was shown that the WT viruses displayed differences in growth kinetics (Figure 5-1 & -3) in both cell lines and plaque phenotype in MDCK cell (Figure 5-2). Although U/63-WT grew poorly in E.Derm cells (Figure 5-3), it replicated at high rate in MDCK cells and reached its peak of infection earlier (although at a lower titre) than O/03- WT (Figure 5-1). In contrast, O/03-WT grew more progressively and reached high titres in both cell lines (Figure 5-1 & -3). In addition, O/03-WT displayed larger plaques than U/63- WT in MDCK cells (Figure 5-2). Interestingly, reassortment of U/63 and O/03 NS segments modified these infection phenotypes. Furthermore, the effect was dependent on the viral context- and the cell type. Indeed, introduction of O/03 NS segment into U/63 backbone only changed U/63 growth pattern in MDCK cells (Figure 5-1), while its plaque phenotype (Figure 5-2) and growth kinetics were mostly unchanged in E.Derm cells (Figure 5-3). In contrast, O/03 growth pattern in E.Derm cells (Figure 5-3) and plaque phenotype in MDCK cells (Figure 5-2) were strongly affected by the NS segment reassortment, while its growth kinetic in MDCK cells was maintained (Figure 5-1). These results are in accordance with Shelton et al.’s work, who studied the replication capacity of NS reassortants between the human 2009pH1N1 virus and a human seasonal H3N2 virus, two viruses that express a short non-CPSF30 binder NS1 protein and a full-length NS1 that interacts with CPSF30, respectively (Shelton et al., 2012). They showed that the introduction of the H3N2 NS segment into the p2009pH1N1 virus reduced the replication efficiency of the latter in an IFN competent model. In addition, Twu and colleagues showed that the replication efficiency and plaque phenotype of an IAV expressing an NS1 protein that binds CPSF30 was strongly affected in MDCK cells when this binding was altered (Twu et al., 2006). Other reports have also described that for other IAVs, the two NS1 proteins could be swapped without affecting viral growth between (Kim et al., 2014). This further reinforce the hypothesis of species- specific functions of NS1.
To study in more details the effect of NS1 evolutionary markers on EIV infection phenotype, three isogenic mutants for U/63 and O/03 viruses carrying a mutation at codon 186 and/or codon 220 (Figure 5-4 & -5) in their respective NS segment –were generated. By comparing the replication efficiency and plaque phenotype (Figure 5-8 & -9) of mutant and WT viruses in MDCK (Figure 5-6 & -7) and E.Derm cells (Figure 5-10 & -11), it was confirmed
that the introduced mutation affected viral growth and cell-to-cell spread in a viral context- and cell type-dependent manner. Indeed, in MDCK cells substitution E186K resulted in a strong reduction of U/63 replication rate, a delay in the peak of infection, and an increase in virus titre at the peak of infection (Figure 5-6). However, in the context of O/03, K186E substitution did not change viral growth kinetic (Figure 5-7).
Furthermore, the plaque phenotype of U/63 did not seem to be affected by residue 186 (Figure 5-8). In contrast, K186E substitution significantly reduced O/03 plaque size (Figure 5-9). Furthermore, in E.Derm cells E186K substitution did not improve U/63 viral growth kinetics (Figure 5-10), but significantly reduced O/03 growth abilities (Figure 5-11). These findings are in accordance with the results obtained earlier with the NS reassortant viruses (Figure 5-1). Taken together, these data seem to indicate that in cells that do not possess an efficient antiviral response, like MDCK cells (Seitz et al., 2010), an avian-like virus such as U/63 can tolerate mutations in the NS1 gene that alter important functions of the protein, such as the CPSF30 binding. It is possible that other functions of NS1 or other viral proteins may compensate for the loss of CPSF30 binding of U/63 and allow the virus to maintain fitness in cells permissive for influenza virus. However, the results shown above also highlight that by losing the CPSF30 binding, U/63 lost the capacity to quickly take control of infected cells and rapidly produce new virions, a characteristic that would likely provide a great advantage when jumping hosts. In contrast, for a mammalian-adapted virus, like O/03, viral fitness seems to strongly depend on NS1 and on the maintenance of its evolutionary markers, E186K and C-terminal truncation. The gain of a new NS1 activity (CPSF30 binding) may have disrupted the subtle virus-host equilibrium, leading to attenuation of the virus in interferon competent E.Derm cells. This hypothesis is reinforced by the work of others showing that destabilisation of NS1-CPSF30 complex may arise during the adaptation process of IAV strains to certain hosts (i.e. duck to quail, human to mouse) (Hayman et al., 2006, Kochs et al., 2007a, Twu et al., 2007, Brown et al., 2001, Hossain et al., 2008).
Interestingly, as commented earlier (Chapter 1 section 1.3.1), the surface involved in NS1 ED dimerization is also involved in CPSF30 binding (Hale et al., 2008), and previous work has reported that the disruption of ED-ED interaction by substitution of W187 (direct neighbour of residue 186) resulted in attenuation of a recombinant PR8 virus in vivo (Ayllon et al., 2012). The ED dimerization has been shown to play a key role in reinforcing the dsRNA binding of NS1, hence reinforcing NS1’s control of the host IFN response. It is possible that restoring the CPSF30 binding in O/03 NS1 has disrupted more important functions that are crucial to support viral replication in IFN competent cells. Additionally, the presence of E186 in O/03 NS1 seemed to have a high fitness cost in equine cells when introduced in the
context of a short NS1 protein, as O/03-K186E replicated poorly in equine cells compared to the O/03-K186E-230 mutant (Figure 5-11). This could indicate that the potential new functions provided by the extension of O/03 NS1 C-terminal tail, e.g. PABPII binding or interaction with PDZ-containing proteins, could partially compensate for fitness cost associated with the reintroduction of CPSF30 binding.
The modification of NS1 C-terminus on its own seemed also to affect EIV infection phenotype, although to a lesser extent than substitution 186. Indeed, the U/63 NS1 truncated mutant virus (U/63-219) displayed a reduced growth rate compared to WT in MDCK cells (Figure 5-6). The growth pattern of U/63-219 resembled the one of U/63-NS- O/03 reassortant virus (Figure 5-1). Furthermore, as for residue 186 substitution, the truncation of NS1 did not seem to affect U/63 plaque phenotype (Figure 5-8), nor improve viral replication in E.Derm cells (Figure 5-10).
In the context of an adapted virus, like O/03, NS1 C-terminal extension alone strongly affected viral replication in interferon competent cells (Figure 5-11), as well as viral spread between neighbouring cells (Figure 5-9). This extension had even more dramatic effects when introduced in association with K186E substitution (Figure 5-7, -9 & -11). Taken together, these data suggest that introduction or suppression of the CPSF30 binding had bigger consequences on viral fitness than a PBM or PABPII binding sites.
When the response of equine cells to O/03 WT and mutant virus infection was assessed, it was observed that only the WT virus was able to replicate to high titres (Figure 5-11) in the presence of high levels of antiviral cytokines (Figure 5-14-A). Moreover, the O/03-230 and O/03-K186E-230 mutant viruses seemed to be able to control better than WT the production of antiviral cytokines upon infection (Figure 5-14-C & -D). These results are in accordance with a previous report from Daly & colleagues, who showed that ponies infected with an EIV from 2003 expressing a truncated version of NS1 (A/equine/Newmarket/5/2003) was inducing larger amounts of type I IFN than an EIV expressing a full length version of NS1 (A/equine/Sussex/1/1989) (Daly et al., 2011, Wattrang et al., 2003). Of note, similar results were found with swine and turkey influenza viruses (Cauthen et al., 2007, Solorzano et al., 2005). Additionally, others reported that decreasing NS1-mediated inhibition of host gene expression correlated with increased innate immune responses after infection (Twu et al., 2007, Hale et al., 2010).
The O/03-WT virus was also able to infect an increasing number of cells and replicate in equine cells primed with universal IFN (Figure 5-16). The ability to replicate in the presence of type I IFN is a significant fitness trait, as it renders an important arm of the host antiviral response ineffective. Moreover, K186 and C-terminal truncation seemed to be detrimental in this process. Interestingly, no Mx1 and ISG15 expression could be detected upon O/03-WT infection in equine cells (Figure 5-13) despite the presence of large amounts of antiviral cytokines secreted (Figure 5-14). This suggests that O/03-WT was able to repress ISG expression, although further work would be needed to determine the mechanism involved. Furthermore, the role of the JAK/STAT pathway in limiting viral replication of O/03-NS1 mutant viruses in E.Derm cells was confirmed with the use of Ruxolitinib prior to viral infection (Figure 5-15).
Finally, when looking at protein production upon infection it was observed that O/03- K186E-230 induced a cellular protein shutdown, particularly at early times post infection (Figure 5-12). This was associated with a premature induction of apoptosis compared to WT (Figure 5-13). This was followed by expression of Mx1 and ISG15 starting at 24 hpi and increasing at 48 hpi (Figure 5-13). The extension of NS1 C-terminus on its own also increased ISG proteins expression compared to WT and resulted in early induction of apoptosis in infected cells (Figure 5-13). However, similarly to WT no protein shutdown was detectable for the O/03-230 mutant virus (Figure 5-12). These results are in accordance with data obtained by others (Golebiewski et al., 2011, Javier and Rice, 2011, Liu et al., 2010) showing that full length NS1 proteins are involved in early induction of apoptosis, likely due to interaction with PDZ-domain containing proteins important in the maintenance of cellular homeostasis (Harris and Lim, 2001, Hung and Sheng, 2002, Kim and Sheng, 2004). It is also possible that the premature induction of apoptosis by O/03-K186E-230 was an indirect consequence of the many cellular-disturbing functions due to CPSF30 binding. Finally, we cannot rule out the possibility that premature apoptosis is due to a lack of control of PKR (Takizawa et al., 1996, Fujimoto et al., 1998, Takizawa et al., 1995, Wada et al., 1995), or a lack of activation of the PI3K/Akt-pathway. Indeed, the presence of viral dsRNA activates PKR and starts a cascade of events eventually leading to protein shutoff (Bergmann et al., 2000, Hatada et al., 1999, Wang et al., 2000) and induction of apoptosis (Takizawa et al., 1996, Van Campen et al., 1989, Gil and Esteban, 2000). NS1 has been shown to inhibit PKR function by several mechanisms (Tan and Katze, 1998, Hatada et al., 1999, Talon et al., 2000a), and delay apoptosis by activation of the PI3K/Akt-pathway (Ehrhardt et al., 2006, Ehrhardt et al., 2007, Hale et al., 2006, Shin et al., 2007c). In the future, it would be interesting to test if E186K substitution and C-terminal truncation modified EIV NS1 blockade
of PKR function, or altered its interaction with the p85b subunit of PI3K, and subsequent activation of PI3K/Akt pathway.
Taken together, these data indicate that the genetic constellation of an equine- adapted virus, like O/03, could already be optimized to function efficiently in equine cells and any disruption in viral-host interaction has a high fitness cost. Interestingly, a similar evolution pattern has been observed for the North American ‘classical’ swine H1N1 lineage, whose NS segment is of avian origin (related to the 1918 pandemic H1N1 influenza A virus). This virus maintained a full length NS1 protein (230 amino acids) until the mid-1960s, before introducing of a stop codon at position 219 that resulted in the same 11-amino acid C- terminal truncation of than EIV NS1. These changes have subsequently been retained in the ‘classical’ swine H1N1 lineage until the present day (Hale et al., 2010). Thus, the loss of CPSF30 binding and C-terminal truncation of NS1 seem to be a common evolutionary trait between swine and equine Influenza A viruses. Interestingly, human influenza A viruses seem somewhat different, and preferentially select for NS1 proteins binding CPSF30. For example, the NS1 protein of H5N1 viruses presented a defect in inhibiting general gene expression when the transmission occurred from birds to humans in 1997, however the viruses isolated since 1998 have gained this function (Twu et al., 2007, Clark et al., 2017). It would be interesting to compare the advantages and downsides of losing CPSF30 binding and variations of the C-terminal in different mammalian hosts, such as horses, pigs, dogs and humans.