This study was designed to identify the molecular determinants that confer transmissibility of the pH1N1 virus and the mechanism by which they promote transmission. RD transmission can be modulated at the level of the infected donor, the environment, and the recipient. We established an RD transmission caging system that allowed for aerosol sampling of infected ferrets. In our system, the Rec pH1N1 virus transmitted to 100% of the naïve animals and replacement of the NA and M gene segments with the corresponding gene segments from TRS resulted in reduced transmission efficiency. These findings indicate that the Eurasian-origin NA and M gene segments contribute to the efficient transmission of the Rec pH1N1 virus. The fact that the Eurasian virus only transmitted to 50% of the naïve animals demonstrates that gene constellation may influence this phenotype as it does other properties such as virulence 
. Yen et al. have recently suggested that a balance between HA and the Eurasian-origin NA contribute to the transmissibility of the pH1N1 virus 
. Unlike our study, they used swine isolates that had not infected humans; therefore, any compensatory mutations that promote the initial transmission from an animal host to human were not taken into account. Based on our results, we believe that the biological properties of both Eurasian-origin gene segments influence particle release and thus efficient RD transmission. In our study, we found that susceptibility of the recipient ferrets to the specific virus, measured as the FID50
of the virus, did not correlate with transmission efficiency. Since environmental factors such as temperature and relative humidity were unaltered during the study, they did not contribute to the transmission phenotype. Therefore, we focused our attention on the release of influenza viruses by the infected donor ferrets. The viruses used in this study shared similar receptor specificity and replicated efficiently in the upper respiratory tract of ferrets. These two factors have been implicated in the transmissibility of other influenza viruses but they did not contribute to the enhanced transmission phenotype of the pH1N1 virus in our study. Using aerosol biosamplers to measure the release of virus into the air, we found that viruses containing the Eurasian-origin NA and M gene segments released influenza viral RNA-containing particles into the air consistently and this correlated with increased NA activity of these viruses. The Eurasian-origin gene segments also conferred the pleomorphic phenotype of the pH1N1 virus. Our observations extend our knowledge of the molecular determinants of RD transmission and provide an explanation for the epidemiological success of the pH1N1 virus.
An infected donor can generate aerosols during normal breathing or upon sneezing and coughing 
. In our study, we used ferrets as donors because they are highly susceptible to influenza viruses and can both transmit the virus to humans and acquire infection from humans 
. Ferrets infected with influenza viruses develop clinical symptoms such as weight loss, sneezing, and lethargy 
. Disease severity in ferrets and humans varies by strain, with highly pathogenic strains such as H5N1 avian influenza viruses causing more severe disease than seasonal influenza strains 
. We found that the 2009 pH1N1 virus and its precursor viruses caused similar disease severity in ferrets, defined by >10% weight loss and presence of clinical symptoms like sneezing and runny nose (Table S1
). However, we also found that one out of four ferrets infected with TRS or Eurasian viruses developed croup and were able to efficiently transmit the virus to their naïve partners. Upon further analysis, we found a correlation between infected ferrets that were observed sneezing or coughing and infection of their naïve neighbors, indicating that generation of aerosols by sneezing and coughing enhances RD transmission.
In this study, we examined the size of influenza viral RNA-containing particles released from ferrets infected intranasally (IN) with influenza viruses and found that the ferrets primarily released influenza viral RNA-containing particles greater than 4 µm in size into the air (). Consistent with our observations, Gustin et al. reported that anesthetized ferrets infected IN predominantly released large (>4.7 µm) infectious particles during normal breathing. However, they found that ferrets infected by aerosol released much smaller (0.65 to 4.7 µm) particles containing infectious virus into the air 
. We found that ferrets inoculated IN with pH1N1 and Eurasian viruses released large (>4 µm) and small (1 to 4 µm) influenza viral RNA-containing particles more consistently than ferrets infected with the TRS and 6
2 reassortant viruses ( and ). The viruses with more consistent release of virus had a higher NA activity than viruses that were associated with sporadic release of influenza viral RNA-containing particles (). Thus, NA activity correlates with the release of both large and small influenza viral RNA-containing particles. However, these particles are not sufficient for efficient RD transmission since the Eurasian virus, which consistently released large and small influenza viral RNA-containing particles, transmitted to only 50% of the naïve animals (). Additionally, in animals infected with the TRS virus, we only detected the presence of large particles containing influenza viral RNA in the air, yet this virus transmitted to 50% of the naïve animals. These data suggest that the large particles (>4 µm) may contribute to RD transmission of viruses in the ferret model system. Release of large particles containing influenza has been observed in human clinical studies 
. However, the relative importance of these particles in human transmission is unclear.
Interestingly, release of submicron influenza viral RNA-containing particles differed between pH1N1 and the Eurasian viruses (). The Rec pH1N1 infected ferrets consistently released submicron influenza viral RNA-containing particles while ferrets infected with the Eurasian virus did not. Given that the animal cages have a continuous air flow rate of 40 cubic feet per minute, it is also possible that we were unable to thoroughly capture the submicron particles released by the ferrets by sampling on the outside of the cage. Aerosol sampling in different environments suggests that humans predominantly release small, respirable particles that likely result in the respiratory or aerosol transmission of influenza viruses 
. Since the pH1N1 infected ferrets released more submicron particles than ferrets infected with any of the other viruses, it is possible that the submicron particles are responsible for the efficient aerosol transmission of the pH1N1 virus.
Previous studies have demonstrated a role for HA receptor binding specificity and specific amino acid residues in the PB2 protein on RD transmission of influenza A viruses 
. The emergence and transmissibility of the 2009 pH1N1 virus cannot be explained by these molecular determinants of transmissibility of the virus via RDs. Instead, our study illustrates the importance of the NA and M proteins in the transmissibility of the pH1N1 virus. We found that NA activity correlates with the release of particles greater than 1 µm in size and this may be necessary, but not sufficient, for RD transmission. Additionally, we found that viral morphology correlated with transmissibility of swine-origin viruses in the ferret model. The pleomorphic Rec pH1N1 virus was more efficiently transmitted than the spherical 6
2 reassortant, TRS, and Eurasian viruses, suggesting that this phenotype may be important for RD transmission of swine-origin viruses. While there are many examples of α2,6-specific receptor binding influenza viruses that do not transmit in animal models or in the human population 
, there are no reports of RD transmission of α2,3-specific receptor binding influenza viruses. Therefore, virus receptor binding specificity is also necessary, but not sufficient, for transmission.
Our data indicate that in order to more accurately assess pandemic threat potential, phenotypes that are important for transmission such as viral replication in the upper respiratory tract of ferrets, release of respirable influenza virus-containing particles, and receptor specificity of novel influenza viruses should be characterized.