| H5N1 influenza viruses are important pathogens for both veterinary and public health. In 1997, highly pathogenic H5N1 avian influenza virus caused outbreaks in chickens in Hong Kong and was transmitted to humans, causing the deaths of 6 of 18 people infected. Up-to-data, about 500 human cases of H5N1 infections with greater than 50% mortality have been reported in many countries, posing huge challenge to the public health. To understand the genetic determinants of viral pathogenicity and their ability to cross species barriers to mammalian hosts is, therefore, a research priority on influenza viruses.The H5N1 viruses are divided into distinct phylogenetic clades based on their HA genes. The clade 2.3.4 is one of the widely distributed clade and the viruses in this clade continue to infect poultry and humans in China and several Southeast Asia countries. Our previous studies revealed clade 2.3.4 viruses isolated in China formed multiple genotypes due to the big diversity detected in their NA and internal genes. We tested 16 clade 2.3.4 viruses isolated from China in a mammalian mouse model. Our results indicated that these viruses are different in their replication and virulence in mice, and they could be classified into high, middle and low pathotypes.Based on the results described above, we selected two viruses as models to explore the virulence determinants of the clade 2.3.4 viruses in mammalian hosts. The two viruses, A/Anhui/2/2005(AH/2/05) and A/chickens/Sichuan/81/2005 (CK/SC/81/05), are genetically very similar and only differ by 27 amino acids in their entire genome. However, the two viruses are significantly different in their virulence to mice. The AH/2/05 virus is highly lethal in mice, with a 50% mouse lethal dose (MLD50) of 1.5 log10EID50 (50% egg infectious dose), while the CK/SC/81/05 is quite mild, and it mainly replicates in the lung of mice, with a MLD50 of 5.5 log10EID50. We used reverse genetics to generate a series of reassortants of these two viruses and tested them in mice. The results indicated that none of the single-gene-switch contributed to the observed difference of the two viruses in mice, but the switch of both the PA and PB1 genes could completely change the pathotypes of two viruses in mice. We found that the changes of two amino acids at position 622 and 709 of PB1 and one amino acid at position 315 of PA contributed together to the pathogenicity difference of the two viruses in mice. We further demonstrated that the changes in these sites affect the polymerase activities of the two viruses in mammalian cells, which partially elucidate the mechanism of their virulence in mammalian hosts. Our results here are the first to indicate that the mutations in PA and PB1 proteins are critical for the pathogenicity of H5N1 influenza in mammalian hosts, and the amino acids we identified here could be used for assessing the virulence potential of H5N1 field isolates. |
