Detection of Victoria lineage influenza B viruses with K162 and N163 deletions in the hemagglutinin gene, South Africa, 2018

Abstract Background A group of Victoria lineage influenza B viruses with a two amino acid deletion in the hemagglutinin (HA) at residues K162 and N163, was detected during the 2016 to 2017 Northern Hemisphere influenza season and continues to spread geographically. We describe the first identification of viruses with these deletions from South Africa in 2018. Methods Nasopharyngeal samples were obtained from the syndromic surveillance programs. Real‐time reverse transcription‐polymerase chain reaction was used for virus detection and lineage determination. Influenza genetic characterization was done using next‐generation sequencing on the MiSeq platform. The duration of virus circulation was determined using thresholds calculated using the Moving Epidemic Method; duration was used as an indicator of disease transmissibility and impact. Results In 2018, 42% (426/1015) of influenza‐positive specimens were influenza B viruses. Of 426 influenza B‐positive samples, 376 (88%) had the lineage determined of which 75% (283/376) were Victoria lineage. The transmissibility of the 2018 South African influenza season was high for a few weeks, although the severity remained moderate through most of the season. The sequenced 2018 South African Victoria lineage influenza B viruses clustered in sub‐clade V1A.1 with the 162‐163 deletions. Conclusions We report the first detection of the 162‐163 deletion variant of influenza B/Victoria viruses from South Africa in 2018, and suggest that this deletion variant replaced the previous circulating influenza B/Victoria viruses. These deletions putatively affect the antigenic properties of the viruses because they border an immune‐dominant region at the tip of the HA. Therefore, close monitoring of these newly emerging viruses is essential.

Influenza B viruses were first described in 1940 and since the 1980s, two lineages of influenza B viruses have circulated with distinct antigenicity and transmission dynamics. 2,3 Historically there has been less focus on influenza B because earlier studies suggested that influenza B caused less disease than influenza A during annual influenza epidemics and it does not pose a pandemic threat. 2 The epidemiology of influenza B Victoria and Yamagata lineage viruses differ, including the younger average age of persons infected with Victoria lineage viruses, but no differences by clinical presentation. [4][5][6] Typically, the South African influenza season starts with influenza A activity and ends with an influenza B "tail". 7 The mean onset of the South African influenza season is week 23. 8 South Africa used the trivalent influenza vaccine in 2018, which contains two influenza A viruses subtypes, A(H3N2) and A(H1N1)pdm09, and one influenza B virus lineage (Yamagata), based on the Southern hemisphere recommendations from the World Health Organization. 9 Since 2020, a quadrivalent vaccine, which contains both of the influenza B lineages is also available for use in South Africa.
Both influenza A and B are prone to antigenic drift, although influenza B viruses evolve more slowly than influenza A. Influenza B viruses are known for insertions and deletions in the hemagglutinin and neuraminidase genes during evolution. 10

| Study procedures
The procedures for syndromic surveillance for pneumonia have been previously described. 8,[13][14][15][16] In brief, study staff completed case report forms and collected combined nasopharyngeal and oropharyngeal (NPOP) swabs for all enrolled inpatients. For the Viral Watch programme, general practitioners completed a case report form and collected NPOP swabs for all enrolled outpatients. 8

| Sample collection and real-time RT-PCR influenza virus detection
Combined NPOP swabs were placed in a universal transport medium (Copan, Murrieta, CA), stored at 4 C to 8 C, and transported within 72 hours to NICD for testing. All specimens were extracted using   Biolabs, Massachusetts, USA). 19 Specimens were then processed for next-generation sequencing on the Illumina MiSeq instrument.
Illumina sequencing library was prepared from amplified PCR products using Nextera XT Sample Preparation kit (Illumina, California, USA). Samples were multiplexed and sequenced using Illumina Miseq v3 kit with 300-bp paired-end reads.
Reads were de-multiplexed following the removal of tags and adaptors. Following importation in CLC Genomics Workbench version 11 (Qiagen, Netherlands), reads were paired, trimmed, and merged using the following parameters: Trim quality score limit = 0.01, Trim ambiguous nucleotides with a maximum number of ambiguities = 1. After trimming, reads shorter than 50 nucleotides were discarded. The resultant sequenced reads were analysed using a reference-based mapping approach with the following parameters: mismatch cost = 10, insertion cost = 3, deletion cost = 3, and length fraction = 0.5, Similarity fraction = 0.95. The consensus sequence was extracted and exported for further analysis. Consensus were 100Â in coverage and did not include date under Q30.

| Phylogenetic analysis
The genetic diversity of full-length NA and HA consensus sequences were determined by alignment with international reference strains Maximum-likelihood trees were constructed using the Tamura-Nei model with 1000 bootstraps using MEGA version 7. 22 The 2018 sequences downloaded from GISAID (all geographical locations) 23 were used to construct phylogenetic trees, but only sequences clustering together with South African viruses were kept in the trees for better visualization.

ACKNOWLEDGMENTS
We thank all members involved in Viral Watch and pneumonia surveillance programmes for the collection of specimens and data management.

Cheryl Cohen reports grants from US CDC, grants from Sanofi
Pssteur, non-financial support from Parexel, during the conduct of the study.
Anne von Gottberg, Nicole Wolter, and Sibongile Walaza report grants from US CDC, during the conduct of the study.