H5N1 Evolution Via Recombination in China
Recombinomics Commentary
October 11, 2006
H5N1 sequences from northern China are being released by the Beijing Genome Institute. Some of these sequences are repeats of earlier sequences released 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” and these sequenced confirm the clear cut examples of homologous recombination.
The new sequences are generally full sequences of all eight gene segments and provide valuable information of the evolution and transmission of H5N1 in China. Many of the sequences are from poultry from various Northern provinces in China in 1997. H5N1 was first detected in 1996 in a goose from Guangdong. The following year, the first human H5N1 were reported in Hong Kong. The human isolates had the characteristic HA polybasic cleavage site, RERRRKKR, found in most of the HPAI throughout Asia. However, the human NA sequences have a 19 amino acid deletion in the stalk portion of the protein, which was not in the Guangdong goose. The human H5N1 also had internal genes (PB2, PB1, PA, NP) which matched sequences from H9N2 And H6N1. This constellation of genes was also found in bird isolates.
This constellation disappeared after Hong Kong culled over 1.5 million birds in 1997. However, the same constellation was recently found in smuggled eggs from Vietnam, indicating the genes continued to circulate in a reservoir outside of Hong Kong.
Reported human H5N1 re-appeared in Hong Kong in a family that visited Fujian province in 2003. The constellation of these genes was designated as the Z+ genotype. The following year the Z genotype emerged in H5N1 exploding out of China, including human cases in Vietnam and Thailand.
However, the NA of the Z genotype had a 20 amino acid deletion in the stalk region that overlapped the earlier 19 amino acid deletion. Many of the 1997 sequences that were just released have this 20 amino acid deletion, and therefore provide an early history of the NA from the Z genotype.
Analysis of these sequences indicates that they have moved into wild birds in eastern Asia. These sequences are distinct from the Qinghai sequences that were first detected at Qinghai Lake in 2005 and subsequently migrated to Russia, Mongolia, Kazakhstan, Europe, the Middle East, and Africa.
The sequences recently released link isolates in northern China and Japan to sequences in southeast Asia, including Indonesia. This linkage is largely via a number of polymorphisms acquired via recombination.
The new sequences also have large regions of identity in sequences isolated years apart. The first 1141 positions of NA from the 1997 Hubei isolate, A/chicken/Hubei/wj/1997(H5N1), are identical to the 2003 Hubei isolate, A/chicken/Hubei/wo/2003(H5N1). Similarly, positions 43-709 of the 2004 blackbird isolate, A/blackbird/Hunan/1/2004(H5N1), are also identical.
The conservation of these sequences raise serious questions about the role of random mutations in the seasonal changes of pandemic H5N1. As` noted earlier, strong conservation is also seen in Hong Kong isolates from the late 70’s, indicating copy errors are not a major source of the seasonal changes.
The same type of evolution was seen in Canadian swine isolates from 2003-2004 which had large portions of exact copies of 1977 PA and PB2 genes. Similarly, the positions 593 and 2033 the PB1 of the 1997 isolate, A/chicken/Hubei/wj/1997 is IDENTICAL to the 2004 isolate,
A/chicken/Henan/wu/2004, even though the sequences are 7 years apart.
Thus, the current data shows that H5N1 sequences can be faithfully copied over long time periods, and the portions of genes can be exchanged by homologous recombination.
These examples are most easily found in multiple isolates collected over a limited time frame. The sequences show identity and then abruptly diverge. These regions of conservation are spread across the gene and do not represent immutable regions or regions that consistently evolve at different rates.
The detection of such recombination in H5N1 in Asia has been hampered by the release of partial sequences or the release of a limited number of genes. NIAID has a program that will generate complete sequences of all eight gene segments. Filling in the database holes created by the partial sequences would be useful.
Media Sources
Phylogenetic Trees