Dysregulation of M segment gene expression contributes to influenza A virus host restriction

The M segment of the 2009 pandemic influenza A virus (IAV) has been implicated in its emergence into human populations. To elucidate the genetic contributions of the M segment to host adaptation, and the underlying mechanisms, we examined a panel of isogenic viruses that carry avian- or human-derived M segments. Avian, but not human, M segments restricted viral growth and transmission in mammalian model systems, and the restricted growth correlated with increased expression of M2 relative to M1. M2 overexpression was associated with intracellular accumulation of autophagosomes, which was alleviated by interference of the viral proton channel activity by amantadine treatment. As M1 and M2 are expressed from the M mRNA through alternative splicing, we separated synonymous and non-synonymous changes that differentiate human and avian M segments and found that dysregulation of gene expression leading to M2 overexpression diminished replication, irrespective of amino acid composition of M1 or M2. Moreover, in spite of efficient replication, virus possessing a human M segment that expressed avian M2 protein at low level did not transmit efficiently. We conclude that (i) determinants of transmission reside in the IAV M2 protein, and that (ii) control of M segment gene expression is a critical aspect of IAV host adaptation needed to prevent M2-mediated dysregulation of vesicular homeostasis.

66 cells with avian IAVs, and which is overcome in human-adapted influenza A lineages. Critically, 119 the mechanism leading to this host-restricted phenotype occurs at the level of gene expression 120 and is driven by host-specific differences in viral nucleic acid sequence, not amino acid changes.

121
This mechanism constitutes a novel paradigm in RNA virus host adaptation, and reveals a new 122 species barrier for IAV, which may be highly relevant for the emergence of avian IAVs into    (Figure 1D and 1E). Nonetheless, at no time did the 149 NL09 M-encoding virus grow to significantly higher titers than the avian M-encoding viruses in 150 any avian host substrate. In contrast, in mammalian cells, human-adapted IAV M segments 151 were found to support improved growth relative to avian-adapted M segments. Viral growth was 152 monitored from low and high MOI in human A549 cells and canine MDCK cells (Figure 2), and 153 additionally from high MOI in human 293T cells (Suppl. Figure 2A). In each case, PR8-based 154 viruses carrying the NL09 M segment grew with more rapid kinetics and to higher titers than 155 isogenic viruses carrying avian M segments, although the differences in growth from low MOI 156 did not reach statistical significance. To extend our findings to an independent human lineage of 157 IAV, we evaluated the growth from high MOI in A549 cells of PR8 Pan99 M and PR8 Beth15 M 158 viruses, each of which have an M segment derived from the human seasonal H3N2 lineage 159 (Suppl. Figure 1C, D). We found that both of these human-adapted M segments supported 160 significantly faster kinetics and higher magnitude of growth than the dkAlb76 89G M segment 161 (Suppl. Figure 2B). Thus, results from multiple different avian and mammalian culture systems

207
In A549 cells, the PR8 NL09 M virus expressed M2 at ~20% of the total M protein, similar to the 208 expression levels observed in infected DF-1 cells (Figure 5A). In contrast, avian M segments 209 yielded markedly higher levels of the M2 protein than were observed for the same viruses grown 210 in avian cells. Here, M2 levels were approximately equal to those of M1 ( Figure 5B, 5C, 5D).

211
Moreover, when compared to the NL09 M segment in human cells, levels of M2 protein were 212 significantly higher for avian M segments (P<0.0001) (Figure 5C, 5D). Thus, compared to 213 matched virus-host pairings in both human and avian systems, M2 was markedly 214 overexpressed when avian-adapted M segments were introduced into human cells.

216
Overexpression of M2 protein in mammalian cells corresponds to increased levels of the 217 M2 mRNA (mRNA 10)

218
The M segment vRNA is transcribed by the viral polymerase to give rise to a co-linear mRNA 219 (mRNA 7 ), which can be spliced by cellular splicing factors to yield mRNA 10 or mRNA 11 [18,56].

220
The M1 protein is translated from the unspliced mRNA 7 , while M2 is expressed from the spliced 221 mRNA 10 . To date, no polypeptide corresponding to the short mRNA 11 ORF has been identified.

234
For all viruses, mRNA 10 was expressed at low levels (10-40% of total segment 7 mRNA), and in

242
In A549 cells, the PR8 NL09 M virus showed relatively high abundance of mRNA 7 and low 243 abundance of mRNA 10 , as was seen in DF-1 cells (Figure 7A). By contrast, for viruses carrying 244 avian-adapted M segments, mRNA 10 levels exceeded those of mRNA 7 in A549 cells ( Figure 7B, 245 7C). Both decreases in mRNA 10 levels and increases in mRNA 7 levels contributed to the altered 246 relative abundance. Again, no significant differences in relative mRNA 11 levels among the 247 viruses were noted ( Figure 7D) suggesting that changes in splicing to produce mRNA 11 do not 248 contribute to changes in M segment mRNA and protein expression observed in A549 cells.

249
Overall, our data point to increased splicing to produce mRNA 10 as the underlying mechanism 250 leading to heightened M2 expression from avian M segments in A549 cells, and suggest that         Figure 7). Overall, these results 315 suggest that expression of avian M2 protein at high levels interferes with the turnover of 316 autophagosomes and with the localization of LC3B-GFP. In addition, the results reveal a 317 reduction in the tight perinuclear localization of M2 upon LC3B overexpression.

318
As an additional test designed to differentiate autophagy induction from autophagy block in cells 319 infected with IAV carrying avian M segments, we used chloroquine treatment followed by 320 Western immunoblotting for LC3B I and LC3B II. We hypothesized that, if autophagic flux is

349
To ensure that the amantadine was working to prevent LC3B II accumulation through inhibition

425
To verify that the levels of M1 and M2 protein exhibited by the chimeric M segment-encoding 426 viruses are modulated at the level of mRNA expression, we conducted RT primer extension 427 assays. These data confirmed that differences in the levels of segment 7 mRNA expression 428 (Suppl. Figure 11) mirrored the observed differences in M1 and M2 protein expression levels.

429
Overall, these data support the premise that splicing signals are at least partly encoded in non-

511
While the previously reported block in autophagy induced by IAV has generally been interpreted 512 as a viral defense against an antiviral mechanism, a pro-viral role for autophagy is not without 513 precedent. In Dengue virus infection, autophagy has been reported to support viral replication.

514
Specifically, autophagic turnover of lipid droplets in infected cells is thought to provide energy

606
For each hemagglutinin subtype (H1Nx to H16Nx), we aligned the available sequences to 607 generate individual consensus M1 and M2 amino acid sequences (Supplemental Table 1

706
Cells were treated with DAPI, and coverslips were mounted on slides using Vectashield (Vector 707 Labs) mounting medium. Images were obtained with on a Nikon FV1000 confocal microscope at 708 the Emory Integrated Cellular Imaging core facility.