H5N1 H9N2 Recombinants in Northern China
Recombinomics Commentary
October 26, 2006
Researchers from Beijing Genome Institute have released approximately 300 H5N1 sequences this month. The sequences are from poultry, wild birds, and swine and define the evolution of H5N1 in China. A series of sequences was deposited February 28, 2005 at Genbank under the title “A cohort of AIV H5N1 subtypes isolated from wild aquatic birds and domestic poultry revealed rapid transmission, frequent reassortment, and identifiable recombination events“. The recent sequences confirm the earlier clear cut examples of recombination, and extend and expand the earlier observations back to 1997 through isolates from 2004.
One striking example of recombination can be seen in A/chicken/Jilin/hl/2004(H5N1). Although the isolate has an H5 that has origin tarcing back to the first H5 in the 1996 Guangdong goose, and the N1 is similar to NA in in genotype G+, as seen in human isolates from 2003 from patients in Hong Kong who had visited Fujian Province, the remaining six gene segments have extensive regions of identity with H9N2 isolates from China the mid and late 1990’s (see deatils here). The large stretches of identity were similar to the identities found between Canadian swine 2004 PB2 and PA isolates and 1977 swine sequences from Tennessee. The sequence fidelity for over 25 years in Canada or 10 years in China, raises serious questions about the importance of random mutations in seasonal variation of seasonal and pandemic flu.
Similarly, the extensive regions of identity with H9N2 were not due to reassortment, which shuffles who gene segments but does not change the gene sequence. Each of the 6 gene segments had regions of identity with H5N1 isolates in addition to regions from H9N2 isolates, indicating that each of the six gene segments was generated by homologous recombination.
Although clear cut examples of recombination in H5N1 fro Hong Kong had been reported previously (see slides linked here), the sequences from China show that the frequency is markedly higher than indicated by the sequences in the databases at GenBank or Los Alamos. Most of the sequences from Hong Kong or China are partial sequences, which can generate strong bias in the database, especially if the withheld regions represent “unexpected” results based on the submitted sequences. These partial sequences would then under represent the frequency or importance of recombination in the evolution of H5N1 and would also impact the development of vaccine based on the recombination in H5N1 isolates.
Similarly, the withholding of new sequences such as the Qinghai sequences generated by Weybridge from European samples collected in 2005 or early 2006, also limits analysis.
To date, most analysis by WHO and consultants have focused on what was thought to be random mutations. These new data from China support the view that the polymorphisms in H5N1 are not due to recent mutations and are far from random. These polymorphisms can be readily found in the expanding database and are present in earlier isolates in subtypes and locations which would be likely to donate new polymorphisms via homologous recombination and there are predictable.
The new sequences from China with identifiable recombination, including recombination between H5N1 and H9N2, highlight the need for a robust database of full sequences. Recently full H5N1 Qinghai sequences have been generated through the NIAID influenza sequencing project. This project has also been used to generate full sequences from avian serotypes other than H5N1. the failure to generate full sequences of H5N1 in Asia generated by St Jude and Hong Kong University and release of H5N1 sequences hoarded by Weybridge remain a cause for concern and impeded the development of vaccines targeting immerging pandemic H5N1 strains.
Media Sources
Phylogenetic Trees