A total of 348 stool samples were collected from children that presented with diarrhea with or without vomiting to eight Canadian pediatric hospital emergency departments that were part of the Canadian Immunization Monitoring Program, Active (IMPACT) network
 from 2007 to 2010. Five pediatric hospitals were included in the study during the first year (2007: Edmonton, Winnipeg, Ottawa, Quebec City, Halifax), three hospitals for the years 2008 and 2009 (Ottawa, Quebec City, Halifax) and six in 2010 (Vancouver, Saskatoon, Winnipeg, Toronto, Ottawa, Halifax). The total number of specimens collected at the sites were as follows: Ottawa, 128; Halifax, 90; Quebec City, 63; Edmonton, 24; Vancouver, 20; Saskatoon, 11; Winnipeg, 9; Toronto, 3.
The study was approved by the research ethics boards of each hospital, in accordance with the Helsinki Declaration on ethical principles for medical research involving human subjects. Specimens were collected from children under 5 years of age that presented with diarrhea with or without vomiting at emergency departments during 2007, 2008 and 2009
. Research nurses at the emergency department of each centre approached parents or guardians of children presenting with acute diarrhea and asked them to consent to rotavirus testing on the child’s stool sample. In 2010 the samples were from children under 16 years of age who were hospitalized for laboratory confirmed rotavirus infection at six hospitals (Vancouver, Saskatoon, Winnipeg, Toronto, Ottawa, Halifax).
At each hospital stool samples underwent testing for rotavirus by enzyme immunoassay (Premier rotaclone EIA kit, Meridian Bioscience), chromatographic immunoassay (bioMerieux Vikia Rota-Adeno, or Coris Bioconcept Combi strip) or by electron microscopy (EM). All specimens were then sent to the National Microbiology Laboratory for further testing and genotyping.
Genomic RNA was extracted from the stool samples using a Nuclisens Easymag magnetic silica extraction method (Biomerieux, France). Hemi-nested multiplex PCR assays were used to P and G genotype the extracted RNA
[32, 36–40]. PCR products were initially designated to genotype based on size comparison after direct visualization after electropheresis on agarose gels. During the course of genotyping work we replaced a previously described primer specific for P genotypes, 1T-1
, with the degenerate primer 1T-1DCDN since it was found to be complementary to a wider range of P strains. The 1T-1DCDN primer was designed with the sequence 5’ TCT ACT GGR TYR ACR TG 3’ using the primer 1T-1D
 as a model, and an alignment of the VP4 variable region using the Clustal W algorithm in MEGA 5.0
. The alignment included sequences from 66 previously untypable Canadian P samples (determined by genetic distance after sequencing), one Canadian P sample, and 15 P reference strains. The 1T-1DCDN primer binds at nt 340–356, a similar position, but one nucleotide shorter than that of 1T-1 (nt 339–356).
The new 1T-1DCDN primer was compared with 1T-1 primer for genotyping using a panel of 65 Canadian rotavirus samples that included 47 samples that were sequence-positive for P, 4 specimens that were P genotype and 14 that were rotavirus negative. The panel was representative in that it was composed of specimens that came from 5 different sites in 2007, 3 sites in 2008 and 2 sites in 2009.
All strains that were positively genotyped by hemi-nested multiplex PCR were confirmed by sequencing of the VP4 and VP7 regions, using the Con3/Con2 and Beg9/End9 primer sets, respectively
[36, 37], or a suitable alternative. PCR products were purified using Montage PCR Centrifugal Filter Devices (Millipore, USA) and then cloned for sequencing in Top10 chemically competent E. coli cells (Invitrogen, USA) using Invitrogen 5’TopoTA cloning kits or sequenced directly. Sequencing of plasmids was carried out by the Genomics Core section of the National Microbiology Laboratory using T3 and T7 plasmid-specific primers.
Phylogenetic analysis was used for confirmation of genotypes and for further analysis of the ORFs of Canadian G9 strains. The Canadian G9 ORF sequences of the strains RVA/Human-wt/CAN/RT034-07/2007/G9P through to RVA/Human-wt/CAN/RT088-09/2009/G9P (GenBank Accession numbers JF964998-JF965010) were aligned, along with reference strains. Phylogenetic trees were constructed using the Clustal W algorithm in MEGA 5.0 software package
, using the Maximum Likelihood method for phylogenetic analysis, with 1000 bootstrap replicates. Lineage designation was based on similarity to previously defined lineage reference strains