Molecular characterization of rotavirus strains from pre- and post-vaccine introduction in South Africa

Abstract

Introduction: Group A rotavirus (RVA) persists as the infectious etiological agent for acute gastroenteritis (AGE) in children under the age of five years globally, despite the introduction of RVA vaccines. A major concern during the post-RVA vaccination period is the potential for vaccine-induced strain changes, which may lead to putative vaccine-escape strains that may affect the efficacy of RVA vaccines in the long-term. There is a paucity of longitudinal whole-genome studies to assess the impact of RVA vaccine introduction on the circulating RVA strains in South Africa. Aim: The study aimed to elucidate the impact of RVA vaccine introduction on the whole-genome of G1P[8] and G2P[4] RVA strains that circulated in South Africa, seven years before and seven years after RVA vaccine (Rotarix®) introduction. Materials and Methods Rotavirus positive stool samples genotyped conventionally as G1P[8] (n=103) and G2P[4] (n=98) as part of the routine World Health Organization (WHO) RVA surveillance were retrieved from the archival storage of Diarrhoea Pathogens Research Unit (DPRU) based in Pretoria and Center of Enteric Diseases, National Institute for Communicable Diseases (NICD), Johannesburg, South Africa. Rotavirus dsRNA was extracted from the stool samples. cDNA was synthesized from the purified extracted RNA using the Maxima Kit. The DNA libraries were prepared using the Nextera XT Kit and paired-end sequencing for 600 cycles (301 x 2) was performed on a MiSeq Illumina® platform. Results and Discussion Firstly, the study investigated the genotype variation and relative emergence of potential vaccine-escape strains. Two atypical DS-1-like G1P[8] strains were identified. The two atypical G1P[8] strains circulated prior to vaccine introduction in South Africa compared to reports in recent studies that detected this strain during the post-vaccination period. Phylogenetic analysis showed that these atypical G1P[8] strains emerged through reassortment mechanism involving locally circulating South African G1P[8] strains and the DS-1-like backbone of G2P[6] strains. Secondly, the study investigated the evolutionary dynamics of South African pre-and post-G1P[8] and G2P[4] strains. South African G1 strains clustered in G1 lineage-I and II while majority (84.2%) of P[8] strains grouped in P[8] lineage-III, highlighting prevalent co-circulation of the G1 and P[8] strains in these lineages before and after vaccine introduction. All the South African G2 strains segregated into the predominant G2-lineage IV, while 99.7% of the P[4] strains clustered in P[4]-lineage IV, underscoring the continued prevalence of these lineages over time. Thirdly, the study investigated the genetic changes between pre and post-vaccine strains, whereby, for G1P[8] strains, we identified amino acid (aa) differences that potentially explained the yearly clustering of post-vaccine strains into sub-lineages. However, we did not observe aa differences that were consistently conserved throughout the post-vaccine period. For G2P[4] strains, we identified aa residues that were consistently prevalent and appeared at a significantly higher frequency in 60-80-% of post-vaccine strains at various sites in the gene segments compared to their negligible proportions of 1-2% during the pre-vaccine period. Fourthly, we investigated the aa profiles within the mapped antigenic regions of the neutralization proteins. The aa profile in the antigenic profiles of pre- and post-vaccine G1 and P[8] strains was nearly the same except for the N147D substitution observed in eight post-vaccine G1 strains. Similarly, the overall aa profile in G2 and P[4] antigenic domains remained mostly unchanged between pre and post-vaccine strains. Conclusion: Rotarix® did not appear to impact the antigenic and non-antigenic profile of South African post-vaccine G1P[8] and G2P[4] strains. Furthermore, the atypical DS-1-like strains identified in this study appeared to have emerged through natural RVA evolutionary processes and not necessarily vaccine-induced. Continued long-term whole-genome surveillance remains necessary to assess the impact of RVA vaccination on circulating RVA strains.