Introduction Different names have been used in publications to describe emerging lineages of the highly pathogenic avian influenza A (H5N1) viruses. This has made discussion and comparison of the various lineages difficult. The H5N1 viruses have now appeared in at least 53 countries on three continents and continue to infect humans as well as to evolve and diversify at an alarming rate. Remarkably, the hemagglutinin protein has not been replaced in the various isolates since 1996, while other genes have undergone reassortment yielding many different genotypes. As such, following the evolution of the HA provides an initial constant by which the strains may be effectively compared. It was proposed to develop a clade nomenclature system based upon the evolution of the HA for several reasons: 1. To unify the system so that interpretation of sequence/surveillance data from different labs becomes easier. 2. To remove stigmatizing labelling of clades by geographical reference. 3. To provide for easy future expansion of the phylogenetic tree. 4. To provide a starting point for a more extensive system to follow based upon antigenic variation and reassortment into multiple genotypes. An international core group of 8 scientists and their collaborators were convened to initiate this process with the encouragement and approval of three international agencies: World Health Organization (WHO), World Animal Health Organization (OIE) and the Food and Agriculture Organization (FAO). Phylogenetic analysis was performed by a variety of approaches on all of the publicly available H5 HA sequences that have evolved in the A/Goose/Guangdong/96-like H5N1 lineage. The initial results support the idea that the currently circulating HPAI H5N1 viruses could be effectively grouped into numerous clades easily designated by a hierarchical numbering system. For example, the 'Fujian-like lineage' within the antigenically diverse Clade 2 of H5N1 would be designated Clade 2.3.4, with other distinct branches called and etc.., while the 'Qinghai lineage' would be designated Clade 2.2. If such a system could become universally accepted, it might serve as a model for other influenza lineages and genes, including seasonal influenza, which currently uses geographical names, and other avian influenza viruses that infect humans. The progress of the international working group and collaborators will be provided in this late breaker presentation. Methodology Sequence alignments Nucleotide sequences of the highly pathogenic H5N1 hemagglutinin (HA) (only nearly complete sequences) were collected from publicly available databases: GenBank (NCBI) and Influenza Sequence Database of Los Alamos National Laboratories (LANL) Large alignment consisted of 884 HA sequences each approximately 1,659 nucleotides - identical sequences/redundant isolates were removed from the alignment to remove bias Small alignment consisted of 109 HA sequences each approximately 1,659 nucleotides - isolates chosen include vaccine strains, reference strains, many human isolates, pathogenesis study strains, geographically diverse isolates Phylogenetic trees Neighbor-joining trees were generated using MEGA (Version 3.1) and PAUP (Version 4.0) using the Kimura 2-parameter. Large tree rooted to highly pathogenic historical Eurasian H5 isolates (turkey/England/91 and chicken/Scotland/59). Small tree rooted at the clade 0 node (gs/Guangdong/1/96 lineage) bootstrap replicates were performed to support tree topology. Maximum likelihood and maximum parsimony trees also generated to confirm clade topology. Conclusions The results support the hypothesis that the HPAI H5N1 viruses can be grouped into several clades designated by a numbering system that can continue to be expanded as the virus continues to evolve. Establishes a system and guidelines to name/number existing clades of highly pathogenic H5N1 avian influenza, including criteria for assigning a new clade name/number to a newly emerging group of variants. Universal acceptance of this nomenclature system will: - unify the classification of isolates so that interpretation of sequence/surveillance data from different laboratories becomes easier - remove stigmatizing labelling of clades by geographical reference - provide a system for future expansion of all-inclusive phylogenetic trees that can be used to follow the global patterns of viral evolution - provide a starting point for a more extensive nomenclature system that can be used to follow the ever-changing antigenic variation of H5N1 and the emergence of novel genotypes from reassortment of the viral internal genes - may serve as a model for other influenza lineages and genes, including seasonal influenza, which currently uses geographical names, and other avian influenza viruses that infect humans Large and small trees containing publicly available sequences will be posted on the WHO GIP website and OFFLU website and maintained as up-to-date (evergreen) evolutionary trees of the H5 hemagglutinin to keep an open forum for following H5N1 evolution - WHO GIP website OFFLU website – WHO/OIE/FAO H5N1 Evolution Working Group participants Mike Purdue, WHO, GIP, Geneva, Switzerland* Ian Brown, VLA-Weybridge, UK Hualan Chen, Harbin Veterinary Research Institute, CAAS, China Ruben Donis, Influenza Division, CDC, Atlanta, GA USA Ron A.M. Fouchier, Erasmus University, Netherlands Yoshi Kawaoka, U of Wisconsin, USA & Inst of Medical Sciences Tokyo, Japan John Mackenzie, John Curtin School of Medicine, Australia Gavin Smith, The University of Hong Kong, HK SAR, China Yuelong Shu, China CDC, China Collaborators Ilaria Capua, IZSVE, Padova, Italy Nancy Cox, Influenza Division, CDC,, Atlanta, GA USA Todd Davis, Influenza Division, CDC, Atlanta, GA USA Rebecca Garten, Influenza Division, CDC, Atlanta, GA USA Yi Guan, The University of Hong Kong, HK SAR, China Elizabeth Mumford, WHO, GIP, Geneva, Switzerland Collin A. Russell, Department of Zoology, University of Cambridge, UK Catherine Smith, Influenza Division, CDC, Atlanta, USA Derek Smith, Department of Zoology, University of Cambridge, UK Dhanasekaran Vijaykrishna, The University of Hong Kong, HK SAR, China Clade designation criteria 1) Maintain previously designated clade numbers when possible (i.e., Clade 2.2 remains 2.2 and Clade 1 remains 1) 2) New clade designations based on phylogenetic tree topology derived from the large tree - H5N1 progenitors (closest to gs/Guangdong/1/96) designated as Clade 0 - Subsequent clades numbered starting from Clade 3 (i.e., Clades 3-9) - Clades designated by the presence of a distinct common node shared by at least 4 isolates - Subclades/sub-subclades designated as a single clade evolves into more than one distinct lineage (based on sharing of a common node) 3) Average percentage pairwise distances between and within clades (using the Kimura 2-parameter): - Distinct clades should have 1.5% average distances between other clades - Distinct clades should have 1.5% average distances within the clade (may be slightly higher in clades that have highly evolved outliers (i.e., ck/Shanxi/2/06 in Clade 7) 4) Bootstrap support for clade defining node (based on 1000 replicates) - >60 at clade defining node 5) Antigenic properties as measured by the hemagglutination inhibition assay should be used as a correlate of clade designation when the data is available Results 9 unique clades defined from analysis of publicly available H5N1 sequences. Clade 2 is highly divergent with 5 subclades within the clade ( ). - subclades 2.1 and 2.3 further delineated into sub-subclades due to distinct evolution of more than one lineage Clade topology nearly identical regardless of phylogenetic analysis/algorithm used (i.e., NJ, ML, MP algorithms produce nearly identical trees) Some clade topology lost if partial HA sequences are used (i.e., < 800 ntds.) - partial HA sequences appropriate for checking clade designation - ideal alignments should have nearly compete HA sequence Isolates in the small tree match the clade designations of the large tree - overall tree topology from big tree to small tree varies because fewer isolates were used in the small tree, which distorts the hierarchical structure of the tree - despite this, all clade designations remain the same Clade descriptions 0 = early progenitors; predominately from Hong Kong (HK) and China (mostly avian, few human) 3 = from HK, China, Vietnam (all avian) 4 = 2002/2003 lineage from HK and China; 2005/2006 from Guiyang Prov. (all avian) 5 = from China and Vietnam; 2004 lineage from Guangxi Province (all avian) 6 = 2002/2004 from China (all avian) 7 = 2002/2004 from China; 2005/2006 from Yunnan, Hebei, Shanxi Provinces (all avian) 8 = from HK and China (all avian) 9 = from China (all avian) 1 = 2002/2003 progenitors from HK; from Vietnam, Cambodia, Thai, Laos, Malaysia (mixed A/H) 2.1 = from Indonesia (mixed avian/human) 2.2 = 2005 progenitors from Qinghai Lake outbreak and Mongolia; isolates from Eastern and Western Europe, the Middle East, and Africa (mixed avian/human) 2.3 = from China, HK, Vietnam, Thailand, Laos, and Malaysia (mixed avian/human) 2.4 = from China (predominately Yunnan and Guangxi Provinces) (all avian) 2.5 = 2003/2004 from Korea, Japan, China; 2006 lineage from Shantou Prov. (all avian) Contact information: Michael L. Perdue, Ph.D. WHO Global Influenza Programme Dept. of Epidemic and Pandemic Alert and Response 1211 Geneva 27, Switzerland Tel: or 3004; Fax: or 4498; Towards a Unified Nomenclature System for the Highly Pathogenic H5N1 Avian Influenza Viruses WHO/OIE/FAO H5N1 Evolution Working Group