The discovery has led advisers to the United States government, which paid for the research, to urge that the details be kept secret and not published in scientific journals to prevent the work from being replicated by terrorists, hostile governments or rogue scientists.
From the article:
“H5N1 remains a poor choice as a bioweapon, because it can’t be controlled once released, … , highlighting the need for an aggressive H5 vaccination effort.”
– Diverse Ferret H5N1 Transmission Pathways (Recombinomics, Feb. 22, 2012):
However, a paper that shows how to makes H5 readily transmissible had already been published in September last year by Reuben Donis and colleagues at the US Centers for Disease Control and Prevention in Atlanta, Georgia. His team equipped a different H5N1 with the same two mutations as Fouchier. He also added a third mutation, and switched other genes to make it transmissible among ferrets.
The above comments add considerable clarity for the requirements for efficient transmission of H5N1 in a ferret model. The above CDC paper has been published and therefore the receptor binding domain changes and associated influenza backgrounds are known. The transmitting H5N1 had 3 receptor binding domain changes (Q196R, Q226L, G228S), and two have been discussed in tandem for years, so it is virtually certain that the two RBD changes introduced by the Fouchier lab were Q226L and G228S. Similarly, the two viruses used in the above study, A/egret/Egypt/1162/2006 and A/Vietnam/1203/2004 have PB2 E627K, which is known to increase virulence and transmission in mammals, so the three changes introduced by the Fouchier lab are almost certainly Q226L and G228S in HA and E627K in PB2.
Thus, these two studies used very different approaches, but overlapped in at least three of the changes in the generation of a transmitting virus. The Fouchier study indicated the creation of a transmitting H5N1 was far less complex than the CDC study, which used a clade 2.2 H5, a seasonal H3N2 (A/Brisbane/10/2007) N2, and clade 1 internal genes. Fouchier used one H5N1 from Indonesia (clade 2.1) and only added three well known changes. Moreover, it is likely that the two additional changes were also well known, and may be among the other changes studied by the CDC including Q196R and S227N which synergize with each other, while Q196R in combination with Q226L and G228S drives the receptor binding to gal 2.6.
Similarly, the Kawaoka lab used a third approach, which took the H5 gene and put it on an H1N1pdm09 genetic background. The H5 was from Vietnam and was almost certainly another sub-clade (126.96.36.199), which demonstrated the versatility of the H5N1 H5, as well as the other genes which could support transmission. Thus, in addition to the H5, one lab used a different clade for internal genes (which also supplied E627K) and a human N2, while the other lab used H1N1pdm09 for the internal genes as well as N1.
It is unclear if additional receptor binding domain changes were introduced by the Kawoaka, but it is clear that three different H5 sub-clades can transmit when supported by diverse additional flu genes, highlighting the need for additional study in this area.
Moreover, it si also clear that the delay in publication of these studies servers no purpose, since the three changes introduced by Fouchier are clear, and passage of influenza in ferrets has been done for decades.
H5N1 remains a poor choice as a bioweapon, because it can’t be controlled once released, and the three experiments described above indicate efficient H5N1 transmission can be generated using a very wide array of combinations, highlighting the need for an aggressive H5 vaccination effort.