Nomenclature updates resulting from the evolution of avian influenza A(H5) virus clades 2.1.3.2a, 2.2.1, and 2.3.4 during 2013–2014

Aim The A/goose/Guangdong/1/96-like hemagglutinin (HA) genes of highly pathogenic avian influenza (HPAI) A(H5) viruses have continued to rapidly evolve since the most recent update to the H5 clade nomenclature by the WHO/OIE/FAO H5N1 Evolution Working Group. New clades diverging beyond established boundaries need to be identified and designated accordingly. Method Hemagglutinin sequences deposited in publicly accessible databases up to December 31, 2014, were analyzed by phylogenetic and average pairwise distance methods to identify new clades that merit nomenclature changes. Results Three new clade designations were recommended based on division of clade 2·1·3·2a (Indonesia), 2·2·1 (Egypt), and 2·3·4 (widespread detection in Asia, Europe, and North America) that includes newly emergent HPAI virus subtypes H5N2, H5N3, H5N5, H5N6, and H5N8. Conclusion Continued global surveillance for HPAI A(H5) viruses in all host species and timely reporting of sequence data will be critical to quickly identify new clades and assess their potential impact on human and animal health.


Introduction
Highly pathogenic avian influenza (HPAI) A(H5N1) viruses continue to circulate in poultry and wild birds in parts of Asia, the Middle East, and Africa. The A(H5) hemagglutinin (HA) gene of these viruses, derived from the A/goose/ Guangdong/1/96 (gs/GD/96) H5 HA lineage, has continued to rapidly evolve since the most recent update to the H5 clade nomenclature by the WHO/OIE/FAO H5N1 Evolution Working Group. 1 The majority of circulating viruses detected since the beginning of 2013 shared the N1 neuraminidase (NA) gene and genotype V-or Z-like internal genes (or reassortants between them) from viruses that evolved in Asia since 1996. [2][3][4] Full genome analysis of viruses collected before 2010 revealed limited propensity of the genotype V-or Z-like H5N1 viruses to generate persistent reassortant genotypes with divergent avian influenza genes. 5 Recently, however, several HPAI A(H5) lineage viruses have acquired NA genes from unrelated avian influenza viruses via reassortment. A novel H5N5 reassortant subtype was detected in 2008, 6,7 and H5N2, H5N3, H5N6, and H5N8 reassortant subtype viruses were reported in China during or after 2010. [8][9][10][11][12][13] These viruses shared H5 genes originating from clade 2.3.4 H5N1 viruses that circulated in China and elsewhere since 2005. In 2014, H5N6 and/or H5N8 subtype viruses were detected in birds from 12 countries in Asia (China, Korea, Japan, and Russia: H5N8; China, Laos, and Vietnam: H5N6), Europe (Germany, Netherlands, UK, Italy: H5N8) and North America (Canada and United States of America: H5N2 and H5N8). [14][15][16][17] In January and February 2015, HPAI H5N1 viruses were detected for the first time in the United States and Canada, respectively, 18 although soon it became apparent that the N1 NA gene was of North American origin. 19 Several recent independent reports describe the evolution of these emerging clade 2.3.4 HPAI viruses and identify divergent viruses using interim clade nomenclature conventions 15,20,21 pending definitive designations by the WHO/OIE/FAO H5 Evolution Working Group. This report describes the phylogenetic analysis of H5 HA sequence data available since the last nomenclature update (data deposited in databases through December 31, 2012) 1 and nomenclature for the newly emerged clades. Clade designations for newly emerged phylogenetic groups were recommended according to previously established criteria. 4

Materials and methods
A total of 4294 H5 HA sequences from GISAID and GenBank databases with virus sequence deposit dates up to and including December 31, 2014, were used in this nomenclature analysis. Since the last H5 clade nomenclature revision, 567 sequences were newly submitted and an additional 16 gs/ GD/96-like H5 HPAI viruses with N2 and N5 NA subtypes were also included for the first time in the analysis. All new sequence names, their assigned clades, accession numbers, and data sources are provided in Supplementary Data S2-S4. Sequences were curated using custom Perl scripts and database filters as in the previous nomenclature updates whereby sequences were removed before further analysis if analysis detected signs of recombination, 16 duplicates, more than five ambiguous nucleotides, less than 60% alignment length, and frame-shifts.
Data were aligned via MAFFT v7.187 22 and trimmed to the beginning of the mature H5 HA protein gene sequence using JALVIEW v2.8.1. 23 Approximate maximum likelihood trees (GTR+GAMMA with 10 000 local support bootstraps) were constructed using FASTTREE v2.1.4. 24 Automated clade annotation of new sequences used LABEL v0.4.4/H5v2013 25 along with manual phylogenetic inspection. An H5 HA phylogeny visualization for a smaller dataset (Figure 1), as well as pairwise p-distance matrices, was calculated in MEGA 5.1, 26 and group averages were calculated with a custom Perl script. Figure 1 is a tree of 242 representative H5 HA genes rooted to Gs/GD/96 (maximum likelihood, 10 000 local support bootstraps, GTR+GAMMA, FastTree2).
New clade designations required the presence of viruses sampled during 2013 and/or 2014, formation of monophyletic clusters with bootstrap values ≥60%, and within-clade average pairwise distances of ≤1Á5%; however, this last upper limit is relaxed when phylogenetic and surveillance data are insufficient to support the split of the clade. 1 As before, new clades were evaluated in the context of five major phylogenetic groups identified: (i) clades 0, 1, and 3 through 9; (ii) clades 2.1 and 2.4; (iii) clades 2.2 and 2.5; (iv) clade 2.3 with all higher order clades except 2.3.2.1; and (v) clade 2.3.2.1 (see Figure S1 A-E). 1 As used for the previous update to the H5N1 nomenclature system, 1 fifth-order groups are designated using an additional letter to the right of the fourthorder clade (i.e., 2.3.2.1a).

Results
The reconstructed phylogenies revealed monophyletic groups in the majority of circulating clades. Groups containing new sequences were analyzed to determine the within-clade average pairwise nucleotide distances (APD; Figure S1A-E). Based on previously defined nomenclature criteria, we concluded that clades 2.  1Á59%, slightly exceeding the 1Á5% threshold; however, these isolates were not detected since 2010 and therefore did not meet the criteria for a new clade designation. Clades 1.  Figure 1, Table 1).

Discussion
Enzootic HPAI A(H5N1) viruses continue to cause outbreaks, predominantly in poultry, and were reported in over 11 countries including Bangladesh, Cambodia, China, Egypt, India, Indonesia, Laos, and Vietnam during 2014. 29,30 Since the initial spread of H5N1 viruses throughout Eurasia and into North Africa, the viruses became entrenched in poultry populations in geographically isolated areas causing sporadic outbreaks and human infections. 5  Lake, China, has been extensively documented. The recent detection of HPAI A(H5) gs/GD/96 lineage viruses in wild ducks in Canada and the western seaboard of the USA is not unprecedented considering previous evidence of gene flow between the Eurasian and American low pathogenicity avian influenza gene pools and overlapping migratory bird flyways. 32 H5N8 and H5N2 viruses have also recently been detected in poultry, and H5N2 viruses have been identified in wild birds from the central and Mississippi flyways of North America. 15,33,34 It remains to be seen whether these viruses become established in North America or what affect they may have on the influenza virus gene pool of aquatic bird populations worldwide.
In articles published between updated nomenclature releases, emerging clades such as 2.3.4.4 have provisionally been given names that may differ from future nomenclature revisions. 6,8,9,11,12,21,27,28 While the WHO/FAO/OIE H5N1 Evolution Working Group considers historical convention in its nomenclature revisions, authors are encouraged to add the word provisional to describe emerging clades in order to facilitate continuity within the literature in the event that such names are not later adopted by the community at large. Continued global surveillance, monitoring, and characterization of HPAI A(H5) viruses from poultry and wild birds, as well as those from sporadic human infections, will be critical to assess the prevalence and public health significance of these new clades in the future.
Data S1. Members of the World Health Organization/ World Organisation for Animal Health/Food and Agriculture Organization (WHO/OIE/FAO) H5N1 Evolution Working Group.
Data S2. List of viruses, clade designation, gene accession numbers, and database sources for sequences used in Figure 1.
Data S3. List of new sequence data since the previous clade nomenclature update, 1 including viruses, clade designation, gene accession numbers, and database sources for sequences used in the phylogenetic analysis.
Data S4. List of authors, originating and submitting laboratories of the sequences from GISAID's EpiFlu TM Database used in this report.