Evolutionary changes between pre- and post-vaccine South African group A G2P[4] rotavirus strains, 2003-2017

Abstract

The transient upsurge of G2P[4] group A rotavirus (RVA) after Rotarix vaccine introduction in several countries has been a matter of concern. To gain insight into the diversity and evolution of G2P[4] strains in South Africa pre- and post-RVA vaccination introduction, whole-genome sequencing was performed for RVA positive faecal specimens collected between 2003 and 2017 and samples previously sequenced were obtained from GenBank (n=103; 56 pre- and 47 post-vaccine). Pre-vaccine G2 sequences predominantly clustered within sub-lineage IVa-1. In contrast, post-vaccine G2 sequences clustered mainly within sub-lineage IVa-3, whereby a radical amino acid (AA) substitution, S15F, was observed between the two sub-lineages. Pre-vaccine P[4] sequences predominantly segregated within sub-lineage IVa while post-vaccine sequences clustered mostly within sub-lineage IVb, with a radical AA substitution R162G. Both S15F and R162G occurred outside recognised antigenic sites. The AA residue at position 15 is found within the signal sequence domain of Viral Protein 7 (VP7) involved in translocation of VP7 into endoplasmic reticulum during infection process. The 162 AA residue lies within the hemagglutination domain of Viral Protein 4 (VP4) engaged in interaction with sialic acid-containing structure during attachment to the target cell. Free energy change analysis on VP7 indicated accumulation of stable point mutations in both antigenic and non-antigenic regions. The segregation of South African G2P[4] strains into pre- and post-vaccination sub-lineages is likely due to erstwhile hypothesized stepwise lineage/sub-lineage evolution of G2P[4] strains rather than RVA vaccine introduction. Our findings reinforce the need for continuous whole-genome RVA surveillance and investigation of contribution of AA substitutions in understanding the dynamic G2P[4] epidemiology.

Keywords: G2P[4] group A rotavirus strains; rotavirus; sub-lineages; whole-genome analysis.

Fig. 1.

VP7 Maximum likelihood phylogenetic tree. Maximum likelihood phylogenetic tree based on the ORF of the VP7 encoding genome segment. The evolutionary model used was Tamura-3 parameter + Gamma (T92 + G). South African pre-vaccine G2 sequences are indicated with black circles while post-vaccine strains are highlighted in red circles. Thirteen previously sequenced G2 strains (from 1984 - 1996) were included for reference and are indicated with black triangles. The reference sequences used to construct the tree are indicated with blue circles. The pie-chart summarizes the number of pre- and post-vaccine sequences. Lineages are indicated in roman numerals. Only bootstrap values > 70% are shown adjacent to each branch node. The scale bar indicates the number of NT substitutions per site.

Fig. 2.

Non-synonymous amino acid substitutions in South African G2 sequences. Non-synonymous amino acid substitutions shared in each sub-lineage of G2 lineage IV where South African G2 strains in this study clustered in respect to the outgroup G2 strain, RVA/Human-tc/USA/DS-1/1976/G2P[4], are shown. South African pre-vaccine G2 sequences are indicated with black circles while post-vaccine strains are highlighted in red circles. Nodes with defining non-synonymous amino acid substitutions are highlighted in blue rectangles.

Fig. 3.

VP4 Maximum likelihood phylogenetic tree. Maximum likelihood phylogenetic tree based on the ORF of the VP4 encoding genome segment. The evolutionary model used was Tamura-3 parameter + Gamma (T92 + G). South African pre-vaccine P[4] sequences are indicated with black circles while post-vaccine sequences are highlighted in red circles. The reference sequences used to construct the tree are indicated with blue circles. The pie-chart summarizes the number of pre- and post-vaccine sequences. Lineages are indicated in roman numerals. Only bootstrap values > 70% are shown adjacent to each branch node. The scale bar indicates the number of NT substitutions per site.

Fig. 4.

Non-synonymousamino acid substitutions shared in each sub-lineage of G2 lineage IV where South African P[4] strains in this study clustered in respect to the outgroup P[4] strain, RVA/Human-tc/USA/DS-1/1976/G2P[4], are shown. South African pre-vaccine G2 sequences are indicated with black circles while post-vaccine strains are highlighted in red circles. Nodes with defining non-synonymous amino acid substitutions are highlighted in blue rectangles.

Fig. 5.

Protein structural and free energy change analysis of VP7. The superimposed VP7 protein structure models of strain, RVA/Human-wt/ZAF/UFS-NGS-NICD18920/2017/G2P[4], representative of the current post-vaccine G2P[4] strains with that of an old G2 South African strain, RVA/Human/ZAF/SA405GR/87/1987/G2P[4]. The observed AA substitutions, A87T, D96N, I113T, N125T, V129M, S178N, N213D, N242S and I287V were shared by G2 strains in lineage IV, where all the South African G2 strains in this study clustered, in relation to the G2 strains in lineage I where a 1987 G2 strain RVA/Human/ZAF/SA405GR/87/1987/G2P[4] and also the prototype DS-1-like strain, RVA/Human-tc/USA/DS-1/1976/G2P[4], clustered. The free energy change difference after mutations was -1.592 kcal/mol. The structure of the pre-vaccine strain is in deep-teal while that of the post-vaccine strain is dark gray. The AA residues are represented in green and firebrick red to examine the structural conformation of the protein structure. The AA residues highlighted in firebrick red represents the pre-vaccine strain while that in green represents the post-vaccine strain.