Comparison of different RT-qPCR assays for the detection of human and bovine group A rotaviruses and characterization by sequences analysis of genes encoding VP4 and VP7 capsid proteins

Authors


Correspondence

Pierre Ward, Agriculture and Agri-Food Canada, Food Research and Development Centre, 3600 Casavant blvd West, Saint-Hyacinthe, QC J2S 8E3, Canada. E-mail: Pierre.Ward@agr.gc.ca

Abstract

Aims

The aim of this study was to compare the performance of four RT-qPCR assays for the detection of human and bovine group A rotaviruses and to characterize the positive samples by sequence analysis of VP4 and VP7 genes.

Methods and Results

RNA extracted from eight human rotavirus strains, and a panel of 33 human and 25 bovine faecal samples was subjected to different RT-qPCR detection systems. Among these assays, only RT-qPCR primers and probe systems B and C were able to detect all human rotavirus strains from cell culture solutions and faecal samples. However, the results showed that the system C was generally more sensitive by one or two logs than the other RT-qPCR assays tested. With the bovine faecal samples, the most efficient RT-qPCR systems were B and A with the detection in 100 and 92% of samples tested, respectively. Human group A rotavirus G1P[8] and bovine G6P[11] were the most frequently used strains identified in this study. A G3P[9] strain, closely related to a feline rotavirus isolated in the USA, was also discovered in a human rotavirus infection.

Conclusion

The RT-qPCR system B was the only TaqMan assay evaluated in this study able to detect rotavirus RNA in all positive human and bovine faecal samples.

Significance and Impact of the Study

Utilization of only one RT-qPCR for the detection of human and bovine group A rotaviruses and the possibility of human infection by a feline rotavirus strain.

Introduction

Rotaviruses were first identified in 1973 as an important cause of gastroenteritis in children (Bishop et al. 1973). As is often the case with gastrointestinal diseases, rotaviruses are transmitted through the faecal–oral route. Rotaviruses are the most common agents of diarrhoeal illness in infants and young children worldwide, and in 2008, diarrhoea attributable to rotavirus infection resulted in 453 000 deaths worldwide in children younger than five years. More than half of all fatal rotavirus infections occurred in five countries: Democratic Republic of the Congo, Ethiopia, India, Nigeria and Pakistan (Tate et al. 2012). Rotaviruses are also an important cause of neonatal bovine diarrhoea (Maes et al. 2003; Swiatek et al. 2010) and are responsible for significant economic losses (Alfieri et al. 2004).

Rotaviruses are nonenveloped icosahedral viruses whose genome consists of 11 double-stranded RNA segments encoding six structural proteins (VP1, VP2, VP3, VP4, VP6, VP7) and six nonstructural proteins (NSP1-NSP6). The genome is enclosed in a triple-layered capsid of 70 nm in diameter. The outer layer is composed of VP4 and VP7 proteins; VP6 protein forms the inner layer; and VP1, VP2 and VP3 proteins form the core capsid layer (Min et al. 2006; Matthijnssens et al. 2008; Kottaridi et al. 2012). Viruses of the genus Rotavirus belong to the family Reoviridae and are classified into five defined species (A to E) and two tentative species (F and G), which are recognized by the International Committee on Taxonomy of Viruses (ICTV) (Raming et al. 2005). A potential new rotavirus species, ADRV-N, has been recently described and was tentatively assigned to species H (Matthijnssens et al. 2012). The rotavirus species are also commonly referred to as rotavirus groups. Groups A, B and C have been detected in human and animal samples, including swine and bovine (Maes et al. 2003), but groups D, E, F and G infect only animals (Rahman et al. 2005; Matthijnssens et al. 2012). Group A rotaviruses are the most commonly isolated rotavirus strains. These strains can be divided into three genogroups (Wa, DS-1 and AU-1) and have been classified into 27 G types and 35 P types based on the sequence diversity of the genes encoding the two outer capsid proteins, VP4 and VP7, respectively (Matthijnssens et al. 2011). In human group A rotavirus, G1P[8], G2P[4] and G4P[8] are generally the most prevalent genotypes (Gentsch et al. 2005; van der Heide et al. 2005; Ahmed et al. 2006). Among bovine rotavirus strains, the most prevalent genotypes worldwide are G10P[11], G10P[5] and G10P[1] (Steyer et al. 2010). Group C rotavirus has been associated with sporadic diarrhoeal illness in different parts of the world and could be an emerging pathogen in humans (Abid et al. 2007).

Group A rotavirus is widespread in wild and domestic animal species, and it has been suggested that zoonotic transmission plays a substantial role in the introduction of novel strains into the human population (Banyai et al. 2009). Evidence for zoonotic transmission of bovine rotavirus strains to humans and genetic reassortment between human and animal rotaviruses has been described in the literature (Khamrin et al. 2006; Martella et al. 2010; Steyer et al. 2010). Recently, a new complete genome classification system was developed for group A rotavirus strains. This nucleotide sequence-based system assigns a specific genotype to each of the 11 genome segments and has increased the recognition of homology between animal and human rotavirus strains (Matthijnssens et al. 2011).

It is known that raw food, treated water, untreated water and irrigation water can represent possible sources of contamination by rotaviruses. These viruses are very stable in the environment, can be spread by faecal material from sick people or animals and may remain infectious for many weeks (Brassard et al. 2005; Leung et al. 2005). Just a few viral particles appear to be sufficient to trigger infection in humans. To put this into perspective, there can be as many as 108 to 1011 particles ml−1 of stool in infected patients (Koopmans and Duizer 2004).

Transmission electron microscopy (TEM) and antigen detection kits, such as enzyme immunoassay and latex agglutination, are the most frequently used methods for the detection of rotavirus. However, these methods are generally less sensitive than molecular techniques such as RT-PCR or RT-qPCR. Approximately 106 viral particles ml−1 of sample are required for detection by TEM (Logan et al. 2006; Jothikumar et al. 2009). In the past ten years, many RT-qPCR tests using SYBR Green or hydrolysis probes targeting the NSP3, VP2, VP4, VP6 or VP7 genes have been developed for the detection of rotavirus RNA (Schwarz et al. 2002; Kang et al. 2004; Pang et al. 2004, 2011; Logan et al. 2006; Min et al. 2006; Freeman et al. 2008; Gutiérrez-Aguirre et al. 2008; Zeng et al. 2008; Jothikumar et al. 2009; Plante et al. 2011; Kottaridi et al. 2012). The sensitivity of these detection assays can be affected by the quality of the extracted RNA, by RNase contamination and by RT-PCR inhibitors in environmental and clinical samples, especially in faecal material (Escobar-Herrera et al. 2006; Rutjes et al. 2007; Scipioni et al. 2008), as well as by the genomic variability of rotavirus strains in the target regions used in the detection tests. Failure to amplify the viral RNA owing to these factors may result in false-negative results. The use of a sample process control artificially added to the samples prior to concentration of the viral particles and RNA extraction can be extremely useful for monitoring the quality of the extraction procedure and for identifying the potential presence of RT-PCR inhibitors that interfere with the amplification reactions (Jones et al. 2009; Mattison et al. 2009; Ward et al. 2009). In Canada, Health Canada's Technical Group on Virology has recommended the use of feline calicivirus (FCV) as a sample process control for detection methods (Houde et al. 2009).

Due to the low infectious dose, the zoonotic issues and the possibility of recovering rotavirus not only from clinical samples but also from environmental and food samples, it is important to have a sensitive detection method that can detect both human and animal strains. The aim of this study was to compare the performance of four previously published RT-qPCR assays, targeting the VP6, VP7 or NSP3 gene, for the detection of human and bovine group A rotavirus strains and to characterize the positive samples by sequence analysis of genes encoding VP4 and VP7 outer capsid proteins.

Materials and methods

Strains and faecal samples collected between 2005 and 2009

Eight human rotavirus strains from the American Type Culture Collection (ATCC) were used in this study [WA (ATCC VR-2018)/G1, 1-9-12/77/S (ATCC VR-1546)/G2, 89-12C2 (ATCC VR-2272)/G3, 408 (ATCC VR-2273)/G1, 248 (ATCC VR-2274)/G4, WISC2 (ATCC VR-2417)/G1, DS-1 (ATCC VR-2550)/G2, WI61 (ATCC VR-2551)/G9]. Viral strains were propagated to produce stock suspensions, as recommended by ATCC. A panel of 65 faecal specimens from Faculté de médecine vétérinaire, Université de Montréal, Saint-Hyacinthe (Canada), Chinook Regional Hospital, Lethbridge (Canada) and the Hospital for Sick Children, Toronto (Canada) was also used. This panel included 13 rotavirus-positive faecal samples collected from children, 10 human faecal samples from patients of different age groups with fulminant gastroenteritis and negative for rotavirus, 10 human faecal samples from healthy individuals and negative for rotavirus, 13 bovine faecal samples positive for rotavirus and 12 bovine faecal samples negative for rotavirus. Positive samples have been confirmed by transmission electron microscopy or by RT-PCR. In addition, the assays were tested for any cross-reactivity that may have occurred using a wide range of viral and bacterial pathogens present in the stool. These viral pathogens included strains of norovirus GI and GII, swine hepatitis E virus, hepatitis A virus and adenovirus 40/41. Bacterial DNA from Escherichia coli O157:H7, Campylobacter jejuni LSPQ 3234, Salmonella thyphimurium ATCC 14028, Staphylococcus aureus ATCC 25923 and Listeria monocytogenes ATCC 7644 were kindly provided by Evelyne Guévremont (Agriculture and Agri-Food Canada, Food Research and Development Center, Saint-Hyacinthe, Canada).

Viral RNA extraction

Faecal samples were diluted 1 : 5 (w/v) in sterile PBS, pH 7·2 (Life Technologies Inc., Burlington, ON, CA) before centrifugation for 20 min at 4000 g. The clarified stool suspensions or the stock solutions of each rotavirus strain were adjusted to reach 1% sodium dodecyl sulfate (Sigma-Aldrich, Oakville ON, Canada) and 100 μg ml−1 of Proteinase K (QIAGEN, Mississauga, ON, Canada). Mixtures were incubated at 37°C for 1 h. To monitor the RNA extraction process, 3·2 × 103 PFU of feline calicivirus (FCV) were added to 140 μl of the resulting suspensions as sample process control (Ward et al. 2009). Viral RNA was extracted with QIAamp® Viral RNA mini (QIAGEN) protocols adapted for the QIAcube robotic workstation (QIAGEN) using QIAamp® Viral RNA body fluid: manual lysis protocol. To protect the extracted RNA from exogenous RNases, RNase inhibitor (RNaseOut, Life Technologies Inc.) was added to the final QIAGEN AVE elution buffer. Recovered RNA was frozen at −80°C until further use.

Primers and probes

All primers (IDT, Coralville, IA, USA) and hydrolysis probes (IDT and Life Technologies Inc. for MGB probes) used in this study are listed in Table 1.

Table 1. Primers and probes used in this study
Molecular MethodPrimers or probeSequence 5′–3′Tm (°C)PolarityLocationReference
RT-qPCR FCV detectionFCV3-Q-AGACACCTCCGACGAGTTATGC57·6+ 299–319Mattison et al. (2009)
FCV3-Q-1CCGGGTGGGACTGAGTGG60·6 383–366
FCV3-QCy5 – CGCCTTACGGATATGAGCAGCCACATTAAC – IBRQ62·2 361–332
RT-qPCR Rotavirus A VP6 (system A)RotaA-F1GGATGTCCTGTACTCCTTGTCAAAA56·7+ 26–50Logan et al. (2006)
RotaA-F2GGAGGTTCTGTACTCATTGTCAAAAA55·3+ 26–51
RotaA-R1TCCAGTTTGGAACTCATTTCCA54·4 170–149
RotaA-R2TCCAGTTTGAAAGTCATTCCATT53·2 170–147
RotaA-P1FAM – ATAATGTGCCTTCGACAAT – MGB/BNFQ  93–75
RotaA-P2FAM – AATATAATGTACCTTCAACAAT – MGB/BNFQ  93–72
RT-qPCR Rotavirus A NSP3 (system B)ForwardACCATCTWCACRTRACCCTCTATGAG57·7+ 963–988Zeng et al. (2008)
ReverseGGTCACATAACGCCCCTATAGC57·31049–1028
ProbeVIC – AGTTAAAAGCTAACACTGTCAAA – MGB/BNFQ67·0+ 995–1017
RT-qPCR Rotavirus A NSP3 (system C)JVKFCAGTGGTTGATGCTCAAGATGGA57·2+ 17–39Jothikumar et al. (2009)
JVKRTCATTGTAATCATATTGAATACCCA50·3 147–123
JVKPFAM – ACAACTGCAGCTTCAAAAGAAGWGT – BHQ57·5+ 96–72
RT-qPCR Rotavirus A VP7 (system D)Q-Rota-ATGGATATCRATGGGATCATCATG53·3+ 598–620Plante et al. (2011)
Q-Rota-1TTTCGAATAGTACATGTCGTAGTTG52·8 791–767
Rota-VP7-QFAM – AAATTAGCTATAGTGGATGTCGTTGATGGG – BHQ58·3+ 715–744
RT-PCR Rotavirus VP7 (serotype G)Beg 9GGCTTTAAAAGAGAGAATTTCCGTCTGG57·8+ 1–28Gouvea et al. (1990)
End 9GGTCACATCATACAATTCTAATCTAAG51·71062–1036
RT-PCR Rotavirus VP4 (serotype P) humanCon 3TGGCTTCGCCATTTTATAGACA54·5+ 11–32Gentsch et al. (1992)
Con 2ATTTCGGACCATTTATATAACC47·7 887–868
RT-PCR Rotavirus VP4 (serotype P) bovineBov4Com 5TTCATTATTGGGACGATTCACA52·1+1067–1088Isegawa et al. (1993)
Bov4Com 3CAACCGCAGCTGATATATCATC53·51930–1909

Conventional RT-PCR

The full-length VP7 segment from human and bovine samples was amplified with Beg-9 and End-9 primers, according to the parameters described by Gouvea et al. (1990). The partial VP4 segment from human samples was amplified with Con 2 and Con 3 primers with cycling conditions described by Gentsch et al. (1992). Primers used for bovine samples were Bov4Com 5 and Bov4Com 3 with conditions described by Isegawa et al. (1993). RT-PCR was performed using a QIAGEN® OneStep RT-PCR Kit (QIAGEN), and the amplified products were separated on a 1% agarose gel with amplicons visualized with ethidium bromide staining.

Construction of plasmid DNA standards for RT-qPCR reactions

Conventional RT-PCR were carried out in a total volume of 20 μl using the QIAGEN® OneStep RT-PCR kit (QIAGEN), according to the manufacturer's recommendations in an Eppendorf Mastercycler gradient system (Brinkman Instruments Canada Ltd., Mississauga, ON, Canada). Amplifications were performed using group A rotavirus strain WA (ATCC VR-2018), and the different primer sets described in Table 1. RT-PCR fragments of 145, 87, 131 and 194 bp corresponding to TaqMan amplification primer system A (Logan et al. 2006), B (Zeng et al. 2008), C (Jothikumar et al. 2009) and D (Plante et al. 2011), respectively, were excised from the gel and purified using the QIAquick® Gel Extraction kit (QIAGEN). PCR products were cloned into pCR® 2·1 TOPO® vector using TOPO TA Cloning® kit (Life Technologies Inc.) with One Shot® TOP10 electrocompetent cells in accordance with the manufacturer's recommendations. The recombinant plasmid stocks were quantified using the NanoDrop spectrophotometer ND–1000 according to the manufacturer's instructions (NanoDrop Technologies Inc., Wilmington, DE, USA) and converted into copy number. The copy number of plasmid was calculated as: copy number = [(concentration of linearized plasmid)/(molar mass)] × (6·023 × 1023). These DNA plasmids were used for the generation of standard curves and as positive controls.

RT-qPCR assays

The RT-qPCR assays were carried out in 25 μl of a reaction mixture comprising 2·5 μl of extracted RNA and 22·5 μl of master mix. Master mix was prepared using the OneStep Brilliant II QRT-PCR core reagent kit (Agilent Technologies Canada, Mississauga, ON, Canada) and contained 5·0 mmol l−1 of MgCl2, 600 nmol l−1 of both forward and reverse primers and 250 nmol l−1 of hydrolysis probe for system A (Logan et al. 2006); or 400 nmol l−1 of both forward and reverse primers and 200 nmol l−1 of hydrolysis probe for system B (Zeng et al. 2008); or 250 nmol l−1 of forward and reverse primers and 100 nmol l−1 of hydrolysis probe for system C (Jothikumar et al. 2009); or 300 nmol l−1 of forward and reverse primers and 200 nmol l−1 of hydrolysis probe for system D (Plante et al. 2011). For the FCV assay, 5·0 mmol l−1 of MgCl2, 300 nmol l−1 of forward and reverse primers and 200 nmol l−1 of hydrolysis probe were included in the master mix (Mattison et al. 2009). RT-PCR amplifications were run in a Stratagene Mx3005P system (Agilent Technologies Canada) in a 96-well format under the following conditions: 30 min at 50°C for reverse transcription, 95°C for 10 min for initial denaturation then followed by 45 cycles of amplification with denaturation at 95°C for 15 s and annealing and extension at 60°C for 1 min. A standard curve for each system was generated using 10-fold serial dilution (108 to 100 genomic equivalents) in a 5 ng ml−1 salmon sperm DNA solution of appropriate purified DNA plasmid.

Cloning and sequencing of RT-PCR product

RT-PCR amplicons for VP4 and VP7 segments were excised from the gel and purified using the QIAquick Gel Extraction kit (Qiagen). Purified PCR products were cloned into pCR 2·1 TOPO vector using TOPO TA Cloning kit (Life Technologies Inc.) with TOP10 electrocompetent cells in accordance with the manufacturer's recommendations. Sequencing was performed on recombinant plasmids in both directions using a CEQ™ 8000 Genetic Analysis System (Beckman Coulter, Fullerton, CA, USA) and a CEQ Dye Terminator Cycle sequencing kit (Beckman Coulter) with M13 forward and reverse primers. Nucleotide alignment was undertaken with the CLUSTAL W (http://www.ebi.ac.uk/clustalw) program. The phylogenetic tree was created by the neighbour-joining method using CLC sequence viewer 6 software (http://www.clcbio.com). Bootstrap analysis was employed to determine the statistical confidence of the phylogenetic relationships. All sequences were deposited in GenBank under accession numbers JX470485JX470523.

Results

Efficiency evaluation of different RT-qPCR assays

A standard curve was established for each RT-qPCR system using the corresponding cloned amplicon, which was serially diluted from 1 × 108 to 1 × 100 copies and amplified in triplicate. The quantification cycle number values (Cq) were plotted against genomic equivalent copies. The standard curves obtained showed an efficiency of 99·1%, a regression coefficient of 0·993, a slope of −3·344 and an intercept of 39·01 for system A; an efficiency of 98·3%, a regression coefficient of 0·999, a slope of −3·363 and an intercept of 39·79 for system B; an efficiency of 99·1%, a regression coefficient of 0·996, a slope of −3·343 and an intercept of 39·32 for system C; and an efficiency of 100·2%, a regression coefficient of 0·996, a slope of −3·318 and an intercept of 40·99 for system D (data not shown). These standard curves indicated that the four assays could detect 2·5 copies per reaction.

Detection of rotavirus RNA by conventional RT-PCR and RT-qPCR assays

All extracted RNA samples were first individually tested for FCV. The 3·2 × 103 PFU of FCV added to the clarified stool suspensions as a sample process control before RNA extraction were detected in all samples with a mean Cq of 25·75. These results were correlated with the Cq of 26·16 obtained for the extraction control with FCV alone (data not shown). These results showed the efficiency of the sample genome extraction process and showed that the RT-PCR reactions were not affected by inhibitors. The different rotavirus molecular detection assays were evaluated and compared in parallel using the same RNA extracts. Each molecular assay included a negative control (RNAse-free water) and a positive control (cloned amplicon). All RT-PCR products obtained on ethidium bromide-stained agarose gel were of the correct size, and no bands were visible in the negative controls (data not shown). The detection results obtained with the four different RT-qPCR assays performed on human and animal rotavirus strains are presented in Table 2. RT-qPCR systems B and C detected rotavirus RNA in 100% (21/21) of the human rotavirus–positive samples, including eight strains from the American Type Culture Collection (ATCC) with G serotype, which are frequently detected in clinical samples. Systems A and D detected rotavirus RNA in 86% (18/21) and 81% (17/21) of the samples, respectively. With the bovine rotavirus–positive samples, system B was the best detection system, achieving 100% detection (13/13) compared with 92% (12/13), 46% (6/13) and 0% (0/13) for systems A, C and D, respectively. All negative samples were found to be negative with the different detection assays, except with system A, where a positive amplification result was observed in 20% of human rotavirus–negative samples. The positive fluorescence signal observed for these negative samples was in the 35–39 Cq range, which corresponds to the single-copy range for this detection system. These positive signals may be due to a nonspecific reaction or probe disruption at the end of the amplification process in the absence of target cDNA, because no amplification was observed with the other rotavirus detection systems tested. This phenomenon has also been reported in studies on norovirus and HAV RT-qPCR assays using hydrolysis probes (Loisy et al. 2005; Houde et al. 2007). Furthermore, with the different detection systems, no positive detection signals were observed for human norovirus GI (3 samples), norovirus GII (3 samples), swine HEV (3 samples), adenovirus 40/41 (3 samples), HAV HM-175, E. coli O157:H7, Salmthyphimurium (ATCC 14628), Camp. jejuni LSPQ 3234, Staph. aureus (ATCC 25923) or L. monocytogenes (ATCC 7644) or for the negative controls. In this experiment, system B showed a higher detection performance for human and bovine rotavirus RNA than the other RT-qPCR assays tested.

Table 2. Detection of rotavirus RNA with four different RT-qPCR assays
 RT-qPCR assays
Rotavirus System A (Logan et al. 2006)Rotavirus System B (Zeng et al. 2008)Rotavirus System C (Jothikumar et al. 2009)Rotavirus System D (Plante et al. 2011)
Human rotavirus strains from ATCC (%)6/8 (75)8/8 (100)8/8 (100)7/8 (88)
WA (ATCC VR-2018)/G1++++
1-9-12/77/S (ATCC VR-1546)/G2++
89-12C2 (ATCC VR-2272)/G3++++
408 (ATCC VR-2273)/G1++++
248 (ATCC VR-2274)/G4++++
WISC2 (ATCC VR-2417)/G1+++
DS-1 (ATCC VR-2550)/G2++++
WI61 (ATCC VR-2551)/G9++++
Children faecal samples positive for rotavirus (%)12/13 (92)13/13 (100)13/13 (100)10/13 (77)
Total (%)18/21 (86)21/21 (100)21/21 (100)17/21 (81)
Human faecal samples from individuals with fulminant gastroenteritis and negative for rotavirus (%)0/10 (0)0/10 (0)0/10 (0)0/10 (0)
Human faecal samples from healthy individuals and negative for rotavirus (%)4/10 (40)0/10 (0)0/10 (0)0/10 (0)
Total (%)4/20 (20)0/20 (0)0/20 (0)0/20 (0)
Bovine faecal samples positive for rotavirus(%)12/13 (92)13/13 (100)6/13 (46)0/13 (0)
Bovine faecal samples negative for rotavirus (%)0/12 (0)0/12 (0)0/12 (0)0/12 (0)

Sensitivity of the different RT-qPCR systems evaluated

The analytical sensitivity of the different RT-qPCR assays was evaluate using four human and four bovine samples known to be positive for rotavirus. The RNA extractions were serially diluted to 10−7 (Table 3). For each dilution, the same RNA extract was tested in duplicate using the four RT-qPCR assays. The limits of detection (LODs) for the four human rotavirus samples were generally 10 to 100 times higher with system C than with the other systems, except in the case of strain HRV 89-12C2 and sample STHY-125, where detection system D showed equivalent or greater sensitivity than system C. For the bovine samples, the limits of detection were also 10 to 100 times higher with RT-qPCR system C than with systems A and B. However, the LODs were identical for systems A and C and strains FMV 1094847 and FCV 1081508.

Table 3. Analytical sensitivity comparison of the four RT-qPCR in one human rotavirus collection strain, three human and four bovine faecal samples contaminated with rotavirus
RNA DilutionRT-qPCR assays
Rotavirus A System A Logan et al. (2006) (Cq)Rotavirus A System B Zeng et al. (2008) (Cq)Rotavirus A System C Jothikumar et al. (2009) (Cq)Rotavirus A System D Plante et al. (2011) (Cq)
Human rotavirus strain HRV 89-12C2 (ATCC VR-2272)
10−124·6224·729·529·3521·9421·9523·5823·29
10−227·6128·1833·4232·9825·3225·4326·6726·64
10−331·531·1537·4736·5828·5628·5429·6430·01
10−434·735·02No CqNo Cq31·7831·6833·5432·78
10−5No CqNo CqNo CqNo Cq34·6733·9036·5036·23
10−6No CqNo CqNo CqNo CqNo CqNo Cq37·6339·03
10−7No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
Human (children) faecal sample W44279 (HSC Toronto) STHY-125
10−125·2724·93030·2522·1822·2624·5424·43
10−228·4428·8333·0833·7625·2925·2328·5529·13
10−332·2631·4736·8341·5228·7628·4331·9131·92
10−435·5735·58No CqNo Cq31·5631·8235·9435·68
10−5No CqNo CqNo CqNo Cq37·2834·9737·6339·30
10−6No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
10−7No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
Human (children) faecal sample H71374 (HSC Toronto) STHY-126
10−123·9825·5629·0728·8321·0320·9825·2425·63
10−228·7529·0932·3932·4724·4224·529·5429·45
10−333·1232·8434·9535·2928·0928·0232·3933·02
10−437·5737·6438·3039·2431·431·3237·0335·66
10−5No CqNo CqNo CqNo Cq35·0434·89No CqNo Cq
10−6No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
10−7No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
Human (children) faecal sample (CRH Lethbridge) CHR-7
10−130·4130·7436·4335·1827·3227·4131·4131·6
10−233·7634·1No CqNo Cq30·2130·0135·2835·45
10−338·2135·97No CqNo Cq33·233·03No CqNo Cq
10−4No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
10−5No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
10−6No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
10−7No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
Bovine faecal sample (FMV St-Hyacinthe) FMV1094847
10−126·9927·0727·227·5123·2923·13No CqNo Cq
10−231·4330·4730·8131·526·5726·43No CqNo Cq
10−333·4833·6534·9635·6929·8329·66No CqNo Cq
10−436·6137·33No CqNo Cq33·3233·87No CqNo Cq
10−5No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
10−6No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
10−7No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
Bovine faecal sample (FMV St-Hyacinthe) FMV1081508
10−123·9824·3728·2428·1922·0122·02No CqNo Cq
10−227·4427·2431·5231·2525·4925·31No CqNo Cq
10−331·2731·2234·5935·0128·7628·63No CqNo Cq
10−434·7235·37No CqNo Cq32·3932·72No CqNo Cq
10−5No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
10−6No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
10−7No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
Bovine faecal sample (FMV St-Hyacinthe) FMV1075018
10−128·7328·335·2234·526·226·33No CqNo Cq
10−231·931·6137·3642·1829·6229·33No CqNo Cq
10−335·2935·81No CtNo Ct33·732·88No CqNo Cq
10−4No CqNo CqNo CqNo Cq36·5336·28No CqNo Cq
10−5No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
10−6No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
10−7No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
Bovine faecal sample (FMV St-Hyacinthe) FMV1089635
10−129·5228·8930·7530·2524·9624·91No CqNo Cq
10−232·9233·934·9834·4628·428·41No CqNo Cq
10−341·5538·2No CqNo Cq31·732·11No CqNo Cq
10−4No CqNo CqNo CqNo Cq34·3637·1No CqNo Cq
10−5No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
10−6No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq
10−7No CqNo CqNo CqNo CqNo CqNo CqNo CqNo Cq

Sequence analysis of partial VP4 and VP7 segments

The G genotypes were identified by comparison with the VP7 sequences with the reference strains. Among the 27 different G types, the human isolates from faecal material in this study were related to G1 (6 strains), G3 (3 strains) and G4 (1 strain) types (Fig. 1). One strain sequence, identified as a G3, showed similarity to two feline rotavirus strains, ITA/BA222 (GenBank accession number GU827411) and USA/Cat2 (GenBank accession number EU708961), having 98 and 95% similarity. The VP7 sequences of six bovine samples were related to G6 strains, while another four were related to G10 strains. To identify the P genotype, a partial VP4 segment of 887 nt was sequenced and compared with 35 established P genotypes. The human strains isolated from faecal material were associated with three different P genotypes: 2 strains with P[6], 6 with P[8] and 1 with P[9] (Fig. 2). The P[9] strain showed 95% homology with strain EU708961 USA/Cat2. The bovine strains revealed two P genotypes P[11] and P[5], with P[11] corresponding to six strains and P[5] to three strains (Fig. 3). The two targeted gene segments for G-P genotype combinations (Table 4) were successfully amplified and sequenced for only 8 of 13 human samples. Rotavirus G1P[8] was the most common combination observed in human infections (4/8; 50%), followed by G1P[6] (2/8; 25%), G4P[8] (1/8; 12·5%) and G3P[9] (1/8; 12·5%). In bovine samples, VP7 and VP4 genotypes were identified in 62% (8/13) of rotavirus-positive samples: G6P[11] (4/8; 50%) and G6P[5] (2/8; 25%), followed by G10P[11] (1/8; 12·5%) and G10P[5] (1/8; 12·5%).

Table 4. Rotavirus G and P genotypes identified in human and bovine faecal samples
 VP7VP4
  1. a

    ND: Not Determined.

Human faecal samples positive for rotavirus
Sample 8081 - STHY-21G1P[8]
Sample S66825 (HSC Toronto) – STHY-119ND aND
Sample T54210 (HSC Toronto) – STHY-120G3ND
Sample W45010 (HSC Toronto) – STHY-121G3ND
Sample T57043 (HSC Toronto) – STHY-122G1P[8]
Sample M43529 (HSC Toronto) – STHY-123G4P[8]
Sample F76031 (HSC Toronto) – STHY-124G1P[6]
Sample W44279 (HSC Toronto) – STHY-125G1P[8]
Sample H71374 (HSC Toronto) – STHY-126G1P[6]
Sample H73726 (HSC Toronto) – STHY-127NDND
Sample M45804 (HSC Toronto) – STHY-128NDP[8]
Sample CHR-7G1P[8]
Sample CHR-120G3P[9]
Bovine faecal samples positive for rotavirus
Sample FMV 1124422G10P[11]
Sample FMV 1122505NDP[11]
Sample FMV 1094847G10ND
Sample FMV 1089933G6P[5]
Sample FMV 1089635G6P[11]
Sample FMV 1087558G6P[11]
Sample FMV 1081508G6P[5]
Sample FMV 1081283NDND
Sample FMV 1078090G6P[11]
Sample FMV 1077415G10P[5]
Sample FMV 1075017G10ND
Sample FMV 1075018G6P[11]
Sample FMV 1087766NDP[11]
Figure 1.

Phylogenic tree of the VP7 nucleotide sequences of rotavirus strains in the study, analysed using the neighbour-joining algorithm for assignment of G genotype. Arrows indicate the strains identified in this study, and the bootstrap scores for branches are shown from 1000 replicates. The scale bar is proportional to 22% sequence divergence.

Figure 2.

Phylogenic tree of the partial VP4 nucleotide sequences of human rotavirus strains in the study, analysed using the neighbour-joining algorithm for assignment of P genotype. Arrows indicate the strains identified in this study, and the bootstrap scores for branches are shown from 1000 replicates. The scale bar is proportional to 25% sequence divergence.

Figure 3.

Phylogenic tree of the partial VP4 nucleotide sequences of bovine rotavirus strains in the study, analysed using the neighbour-joining algorithm for assignment of P genotype. Arrows indicate the strains identified in this study, and the bootstrap scores for branches are shown from 1000 replicates. The scale bar is proportional to 20% sequence divergence.

Discussion

The majority of rotavirus amplification assays target the VP4 and VP7 capsid genes and are generally used for genotyping, while most of the detection assays target the NSP3 or VP6 gene. NSP3 is a highly conserved nonstructural protein, and homologous sequences from this protein have been found in bovine, simian, porcine and human group A rotaviruses (Pang et al. 2004). This genomic region was found to be a good target region for PCR amplification of various rotavirus strains. RT-qPCR assays targeting the NSP3 gene have been proposed for the detection of rotavirus (Pang et al. 2004; Zeng et al. 2008; Jothikumar et al. 2009). RT-qPCR assays are now widely recognized as being very sensitive and rapid, automatable approaches that can provide a relative viral load. In the present study, four different proposed RT-qPCR primers and probe sets were evaluated for their ability to detect human and bovine rotavirus strains. Systems B and C target nucleotide sequences within the NSP3 gene, while system A and system D target the VP6 and VP7 proteins, respectively. These four RT-qPCR assays were principally designed for the detection of human group A rotaviruses; the two assays targeting the NSP3 gene have been validated with clinically important rotaviruses (Zeng et al. 2008; Jothikumar et al. 2009). In the present study, these four systems were used to detect animal rotavirus strains, primarily bovine strains, for the first time. Detection systems B and C showed the best performance for the detection of rotavirus in positive human faecal samples and rotavirus strains representing five major VP7 genotypes, G1, G2, G3, G4 and G9. In this study, system B was the only RT-qPCR assay able to detect rotavirus RNA in all positive human and bovine faecal samples. The primers and probe used in this RT-qPCR assay published by Zeng et al. consisted of a modified version of the primers and the TaqMan probe previously developed by Pang et al. 2004. System A could be a good method for the detection of human and bovine rotavirus; however, the positive amplification signal obtained for negative samples could produce false-positive results. System D, targeting the VP7 protein and previously tested only with the human rotavirus strain WA (ATCC CRL-2018) (Plante et al. 2011), was the least effective system, detecting only 81% of human rotavirus–positive samples, and none of the bovine rotavirus RNA material (0/13) included in this study. These results confirm that the VP7 capsid gene may not be a good target for the detection of a large diversity of rotavirus strains.

Serially diluted plasmid DNA standard was previously used to validate the primers and probes for amplification systems A, B and C and to estimate the limits of detection (Logan et al. 2006; Zeng et al. 2008; Jothikumar et al. 2009). The resulting limits of detection, ranging from 3 to 10 genome equivalent copies, are consistent with the results obtained in the present study. The sensitivity of system C, developed by Jothikumar et al. 2009, was estimated to be between 2 and 4 viral particles. In the present study, system C showed Cq values that were consistently lower than those for the other detection systems, for every sample. In addition, the analysis of rotavirus RNA dilution series revealed that the limit of detection (LOD) obtained for this assay was generally 10- to 100-fold more sensitive than that for the other assays. These results suggest that system C was more sensitive than the other three systems evaluated in this study for the detection of rotavirus in human and bovine faecal samples. The nucleotide composition of the targeted region, the RNA conformation or the availability of the targeted region for the reverse transcription step can explain the differences in efficiency and sensitivity among the RT-qPCR detection systems tested. As the four molecular detection assays were performed on the same RNA extracts, viral RNA extraction recovery, RNA integrity and PCR inhibitors from faecal samples can be excluded.

A new molecular typing system based on all 11 gene segments was developed recently (Matthijnssens et al. 2008). This system is based on sequencing and phylogenetic analysis of all fragments and uses the notations Gx-P[x]-Ix-Rx-Cx-Mx-Ax-Nx-Tx-Ex-Hx for the VP7-VP4-VP6-VP1-VP2-VP3-NSP1-NSP2-NSP3-NSP4-NSP5/6 genes. The system offers good discriminatory power and allows a very fine description of each strain's genotype. However, sequencing of the complete rotavirus genome is a time-consuming process, and this may limit the application of the new typing system. Rotavirus characterization based on G and P genotypes is still widely used. Four G and P genotypes were identified in this study in human samples, including G1P[8], G1P[6], G4P[8] and G3P[9], with G1P[8] being the most commonly detected strain. The G1P[8] genotype is predominant in various countries around the world (Gentsch et al. 2005; van der Heide et al. 2005; Ahmed et al. 2006; Pietruchinski et al. 2006; Lee et al. 2009) and is targeted by the two commercial vaccines approved in Canada, RotaTeqR and Rotarix™ (Lamhoujeb et al. 2010). The emerging rotavirus G9 and the rare G12 serotype were not detected in faecal samples from Canadian children included in this study.

The human rotavirus strain CHR-120 with the G3P[9] genotype was found to be similar to a feline rotavirus strain isolated in the United States. In a study by Tsugawa and Hoshino (2008), homology between human and feline rotavirus strains was demonstrated in all genome segments. Close contact between animals and humans in the domestic environment is a factor that promotes interspecies transmission and mixed infections. The G3 strain has not only been identified in humans but also in several other animal species, such as pigs, monkeys, dogs, cats, horses, mice, lambs, birds and rabbits (Khamrin et al. 2006; Martella et al. 2010).

Studies carried out in Argentina, the Netherlands, Australia, Brazil and Canada showed that the G and P genotypes commonly identified in bovine rotavirus strains are G6, G10, P[11] and P[5] (Alfieri et al. 2004; van der Heide et al. 2005; Lamhoujeb et al. 2010; Swiatek et al. 2010; Badaracco et al. 2012). A similar pattern was observed in the present study; the genotypes identified were G6P[11], G6P[5], G10P[11] and G10P[5]. No similarity has been observed between these bovine rotavirus strains and human strains.

Results obtained in this study indicate that the RT-qPCR assay developed by Jothikumar et al. 2009 (system C) performed very well for the detection of human rotavirus; it was generally more sensitive than the other RT-qPCR assays tested. However, system B is the only assay evaluated in the study that was able to detect rotavirus RNA in all positive human and bovine faecal samples. Among the P and G genotypes identified in this study, G1P[8] and G6P[11] were predominant in the human and bovine rotavirus strains, respectively. Sequence analysis of the VP4 and VP7 genes revealed that the human rotavirus strain CHR-120 was similar to a feline rotavirus strain isolated in the United States.

Acknowledgement

This research was supported by Agriculture and Agri-Food Canada Research Branch Project RPBI # 1485.

Ancillary