Rotaviruses form a genus rotavirus within the Reoviridae family (Estes and Kapikian 2007). Group A rotaviruses (RVA) are the major cause of severe dehydrating diarrhoea. Every year RVA infects approximately 114 million children that leads to 453,000 childhood deaths of which most occur in Africa and Asia (Tate et al. 2012). In South Africa, approximately 17,644 to 25,630 children are hospitalised of which 2,882 die annually due to rotavirus diarrhoea (WHO, 2012). An estimated 224 to 318 children die due to rotavirus disease in Pretoria and neighboring Brits areas of South Africa (Mapaseka et al. 2010).

RVA are non-enveloped tripled-layered enteric viruses that contain 11 double-stranded RNA GSs. The genome is encased within the VP2 (encoded by GS 2), inner VP6 (encoded by GS 6) capsid shell, and the outer glycosylated VP7 (encoded by GS 9) capsid that contains VP4 (encoded by GS 4). VP6 constitutes more than half of the mass of the rotavirus particle (Estes and Kapikian 2007) and contains rotavirus group antigens that are used to sub-classify RVA into subgroups (SG) I, II, I + II, and “non-I, non-II”) based on their reactivity to monoclonal antibodies (Greenberg et al. 1983).

Molecular methods have an edge in RVA diagnosis and classification over serological methods due to problems associated with the availability of monoclonal antibodies raised against specific rotavirus antigens. Of late, 16 genotypes (I1-I16) have been determined using a nucleotide sequence-based whole genome classification system (Matthijnssens et al. 2011). There is a correlation between the serological and genotype classifications where SG 1 and II corresponds to genotype 2 and 1, respectively (Iturriza-Gómara et al. 2002; Kerin et al. 2007; Matthijnssens et al. 2011).

GS 6 and its encoded VP6 play a major role in RVA detection. VP6 is the primary target antigen for RVA routine diagnostic serological techniques such as enzyme-linked immunosorbent assay (ELISA), immunofluorescence, and immunochromatography (Greenberg et al. 1983; Estes and Kapikian 2007). On the other hand, broadly reactive molecular diagnostic assays such as Reverse Transcriptase-Polymerase Chain Reaction (RT)-PCR protocols targets the ends or internal conserved regions of GS 6 nucleotide sequences by employing hybridization methods which uses sequence specific primers and/or probes (Iturriza-Gómara et al. 2002; Lin et al. 2008). The first molecular assays that were developed in the early 1990s utilised GS 6 nucleotide sequences of rotavirus strains circulating from early 1970s to late 1980s (Grinde et al. 1995). More recently, RT-PCR assays based on GS6 have also been developed using old reference and a few new human rotavirus GS 6 sequences. This resulted in a more reliable scheme for conveying various VP6 genogroups of human rotaviruses (Lin et al. 2008; Matthijnssens et al. 2012).

The availability of many complete GS 6 sequences in the GenBank database will ease the prior challenges of improving, developing and validating rotavirus characterisation methods that utilises VP6 encoding GS (Lin et al. 2008). It is important to continuously evaluate the nucleotide sequences for GS 6 of strains circulating particularly in African countries owing to the wide rotavirus G and P genotype variations that have been consistently reported (Mwenda et al. 2010; Kiulia et al. 2014; Seheri et al. 2014; Tsolenyanu et al. 2014). Such studies would inform in advance, potential evolutionary variations that would eventually affect efficiency of VP6 based rotavirus characterisation methods. In this report, the genetic variation of the GS 6 of rotavirus strains collected recently in Pretoria, South Africa, were characterised and compared against those of strains reported from other parts of the world.

The project was approved by the Medunsa Research Ethics Committee (MREC/P/168/2008; MREC/P/108/2013; PG). Eleven human stool specimens were selected randomly from previously screened ELISA rotavirus-positive samples from children presenting with diarrhoea at Lancet private pathology laboratories, Pretoria in 2008. These samples included two commonly characterised G1P[8], six uncommon G and P genotype combinations (G1P[4] and G2P[6]), and three G9P[8]/G9P[6] strains that have emerged in the last two decades. One rotavirus ELISA rotavirus-positive porcine sample obtained from the archive of the Medical Research Council-Diarrhoeal Pathogens Research Unit (MRC-DPRU) laboratory that could not be assigned G and P genotypes was also analysed. Nucleotide sequences for human and animal rotavirus strains containing G2, G8, G9 and G12 were selected from our previous studies (Jere et al. 201120122014; Nyaga et al. 2013) to cover frequently characterised strains around this region.

Rotavirus dsRNA was extracted from 10% fecal supernatant using the QIAamp viral extraction kit by following the manufacturer’s instructions (Qiagen, Valencia, CA). Synthesis of cDNA and genotyping of the GS 4 and 9 were performed using RT-PCR as described previously (Góuvea et al. 1990; Gentsch et al. 1992; Iturriza-Gómara et al. 2001; Iturriza-Gómara et al. 2004). In-house designed primer set VP6F (nt 1–20, 5’GGCTTTAAAACGAAGTCTTC3’) and VP6R (nt 1336–1356, 5’TGTAGTGAGAGGATGTGACC 3’) was used to amplify the entire GS 6 (1356 bp). In brief, RNA was denatured at 95°C for 5 min followed by reverse transcription at 50°C for 30 min. Thirty PCR cycles of 95°C for 30 sec, 50°C for 30 sec and 72°C for 45 sec, followed by an elongation step at 72°C for 7 min was performed. The PCR products were purified with the QIAquick PCR Purification Kit (Qiagen, Valencia, CA) by following the manufacturer’s instructions. The PCR amplicon were run on a 1% Tris-acetate- ethylenediaminetetraacetic acid- agarose gel stained with 1% ethidium bromide. The GS 6 cDNA amplicons were sequenced using the Sanger ABI 3130xl at Inqaba Biotechnical Industries (Pty) Ltd, Pretoria, South Africa. RotaC [http://rotac.regatools.be/] (Maes et al. 2009) was used to assign genotypes to GS 6.

Sequences were aligned using the MUSCLE program within MEGA version 5 (Tamura et al. 2011). Once aligned, the JModelTest 2 program (Posada 2008) was used to identify the optimal evolutionary model that best fitted the sequence datasets. Using corrected Akaike Information Criterion (AICc) the following model; TN93 + G + I was found to best fit the sequence data for the VP6 gene. Using these models, maximum likelihood trees were constructed using MEGA version 5 with 500 bootstrap replicates to estimate branch support. Nucleotide and amino acid distance matrixes were prepared using the p-distance algorithm in MEGA version 5 (Tamura et al. 2011).