Mem Inst Oswaldo Cruz, Rio de Janeiro, 113(1) January 2018
Characterisation of a rare, reassortant human G10P rotavirus strain detected in Honduras
1Centers for Disease Control and Prevention, Gastroenteritis and Respiratory Viruses Laboratory Branch, Atlanta, Georgia, USA
2University of Ghana, Department of Biochemistry, Cell and Molecular Biology, West African Center for Cell Biology of Infectious Pathogens, Legon, Accra, Ghana
3China Center for Disease Control and Prevention, National Institute for Viral Disease Control and Prevention, Beijing, China
4Nacional Colonia La Campaña, Tegucigalpa, Honduras
5Pan American Health Organization, Washington DC, USA
BACKGROUND Although first detected in animals, the rare rotavirus strain G10P14 has been sporadically detected in humans in Slovenia, Thailand, United Kingdom and Australia among other countries. Earlier studies suggest that the strains found in humans resulted from interspecies transmission and reassortment between human and bovine rotavirus strains.
OBJECTIVES In this study, a G10P14 rotavirus genotype detected in a human stool sample in Honduras during the 2010-2011 rotavirus season, from an unvaccinated 30-month old boy who reported at the hospital with severe diarrhea and vomiting, was characterised to determine the possible evolutionary origin of the rare strain.
METHODS For the sample detected as G10P14, 10% suspension was prepared and used for RNA extraction and sequence independent amplification. The amplicons were sequenced by next-generation sequencing using the Illumina MiSeq 150 paired end method. The sequence reads were analysed using CLC Genomics Workbench 6.0 and phylogenetic trees were constructed using PhyML version 3.0.
FINDINGS The next generation sequencing and phylogenetic analyses of the 11-segmented genome of the G10P14 strain allowed classification as G10-P14-I2-R2-C2-M2-A3-N2-T6-E2-H3. Six of the genes (VP1, VP2, VP3, VP6, NSP2 and NSP4) were DS-1-like. NSP1 and NSP5 were AU-1-like and NSP3 was T6, which suggests that multiple reassortment events occurred in the evolution of the strain. The phylogenetic analyses and genetic distance calculations showed that the VP7, VP4, VP6, VP1, VP3, NSP1, NSP3 and NSP4 genes clustered predominantly with bovine strains. NSP2 and VP2 genes were most closely related to simian and human strains, respectively, and NSP5 was most closely related to a rhesus strain.
Group A rotaviruseshave been a major cause of severe diarrhea worldwide among children youngerthan five years (Parashar et al. 2006 25 ) and account for a global mortality ofapproximately 215,000 annually (Tate et al. 2016 33 ). Rotaviruses belong to theReoviridae family of viruses. The viral genome of rotavirus, which iscomposed of 11 double-stranded RNA segments and encapsulated in a concentrictriple-layered protein structure, encodes for six structural and five or sixnon-structural proteins. A binomial classification system has been used traditionallyfor genotyping rotaviruses based on antigenic or genetic characterisation ofthe outer capsid proteins VP7 (G-type) and VP4 (P-type). However, a classificationsystem which makes use of the open reading frame sequences of all the genesis currently widely used as the standard nomenclature (Matthijnssens et al.2008). The viral proteins and non-structural proteins in the order VP7-VP4-VP6-VP1-VP2-VP3-NSP1-NSP2-NSP3-NSP4-NSP5are represented by the genotypes Gx-P[x]-Ix-Rx-Cx-Mx-Ax-Nx-Tx-Ex-Hx, respectively.At least 35 G-types, 50 P-types, 26 I-types, 21 R-types, 19 C-types, 19 M-types,30 A-types, 20 N-types, 21 T-types, 26 E-types, and 21 H-types have been detectedto date (https://rega.kuleuven.be/cev/viralmetagenomics/virus-classification/newgenotypes).With respect to the G and P genotypes in humans globally, G1-4, G9, P andP are still the most common; and the most frequently detected strain combinationsinclude G1P, G2P, G3P, G4P and G9P (Doro et al. 2014). Rotaviruseshave also been classified into Wa-like, DS-1-like and AU-1-like genogroups basedon cross hybridisation studies (Nakagomi et al. 1989). The genogroups classificationwas also later shown in genomic studies (Matthijnssens & Van Ranst 2012).
In developing countries,detection of uncommon strains and unusual genotype combinations are more frequentthan in developed ones, but are still a rare event compared to the detectionof the common human genotypes (Bourdett-Stanziola et al. 2010 2 , Seheri et al.2014). Latin America is one region in the developing world where many countrieshave introduced rotavirus vaccination as part of their national immunisationprograms (Desai et al. 2011 8 ). Studies from Latin American countries suggestthat the epidemiology of rotavirus in the region is the same as the global epidemiology,with more than 70% of strains detected being those that are most frequentlydetected globally (Castello et al. 2004 3 , de Oliveira et al. 2009 6 ). There area significant number of mixed infections and non-typeable strains, but unusualG-P combinations such as G1P, G1P, G2P, G3P, G2P, G4P, G5P,G9P, G4P,G10P and G11P are less frequently, observed (Linhareset al. 2011, Quaye et al. 2013 28 ). Some of these uncommon strains were locallydominant and associated with severe disease (Araujo et al. 2007 1 ).
Until 2013, allreports of the rare G10P strains detected in humans were either on the VP7and VP4 genes characterisations, or the VP6 and NSP4 genes in addition to theVP7 and VP4 genes (Ghosh et al. 2007 13 , Steyer et al. 2010 30 ), and thus far, theG10P strain has not been detected in the Americas, making this study thefirst report in the region.
Stool samples collectedduring the 2010-2011 rotavirus season in Honduras (n = 50) that had tested positivefor rotavirus antigen by enzyme immunoassay (EIA) were sent to the RotavirusSurveillance Laboratory at the US Centers for Disease Control and Preventionfor EIA confirmation, genotyping, and nucleotide sequencing. Ethical approvalwas obtained from the National Health Services of Honduras. For all the samples,10% stool suspensions were prepared from the specimens using PBS, and the presenceof the rotavirus antigen was confirmed by EIA using the Premier™Rotaclone® Detection Kit (Meridian Diagnostics, Inc., Cincinnati, OH). Theimmunoassays were read spectrophotometrically at 450 nm on an MRX Revelationplate reader (Dynex Magellan Biosciences, Chantilly VA). Immunoassays with absorbancevalues greater than 0.15 were considered positive for rotavirus antigen.
Following the manufacturer'sprotocols for initial characterisation, rotavirus double-stranded RNA was extractedfrom the 10% fecal suspensions using the automated KingFisher extraction system(Thermo Fisher Scientific, Waltham, MA) with the Max 96 Viral RNA IsolationKit (Ambion, Inc., Austin, TX). The extracted RNAs were used as templates forreverse transcription polymerase chain reaction (RT-PCR), genotyping, and nucleotidesequencing as previously described (Hull et al. 2011 16 ) to identify the VP4 andVP7 genotypes. The sequence for each gene was compared to rotavirus sequencesin the nr/nt database using the BLASTN program at the National Center for BiotechnologyInformation website (http://www.ncbi.nlm.gov/BLAST/).
For a sample thatwas detected as G10P from an unvaccinated 30-month old boy with acute gastroenteritis,large volumes of the 10% stool suspension were prepared and used for RNA extractionand sequence independent amplification as recently described (Potgieter et al.2009, Jere et al. 2011 17 ) to amplify all the 11 gene segments. The amplicons weresequenced by next-generation sequencing using the Illumina MiSeq 150 pairedend method (Genomics Lab, Hudson Alpha Institute for Biotechnology, Huntsville,Alabama).
Illumina sequencereads were analysed using CLC Genomics Workbench 6.0. A combination of denovo assembly and subsequent mapping to reference strain was used to obtainthe full-length genome of the strain. The assembled gene sequences were alignedwith reference rotavirus gene sequences using the ClustalW program within MEGA5.05 package (Tamura et al. 2011 32 ). Once aligned, the optimal evolutionary modelthat best fit each sequence dataset was identified using AICc criterion implementedin jModeltest2. The best models identified were TIM3+I (NSP1), TPM2uf+G (NSP2),TIM2+G (NSP3), TPM2uf+I (NSP4), TrN+I+G (NSP5, VP6), TIM2+I (VP1), GTR+I (VP2,VP3), GTR+G (VP4), and TIM3+I+G (VP7). Phylogenetic trees were constructed usingPhyML version 3.0 with aLRT statistics computed for estimation of branch support(Guindon et al. 14 2010), and p-distances were computed in MEGA 5.05 to determinethe similarities of the genes to reference strains in GenBank. The scale barson the phylogenetic trees represent proportion of substitution on a branch ofeach tree. The gene sequences were submitted to RotaC (http://rotac.regatools.be/)for genotype assignments. All the gene sequences of the G10P genome hasbeen deposited into the GenBank sequence database under accession numbers KU956006through KU956016.
The G10P strain,which was detected in a stool sample collected from an unvaccinated 30-monthold boy who reported at the hospital with severe diarrhea and vomiting in theDistrito Central of the Department of Francisco Morazan in Honduras during the2010-2011 rotavirus season, was amplified by sequence independent amplificationand sequenced by Illumina next-generation sequencing technology. The sequenceamplification products for the VP1-VP3 were not visible by agarose gel electrophoresis,even though the amplification was successful, and resulted in thousands of readswhen sequenced on the Illumina platform. The submission of the sequences toRotaC resulted in assignment of a G10-P-I2-R2-C2-M2-A3-N2-T6-E2-H3 constellation,showing that six of the genes (VP6, VP1, VP2, VP3, NSP2 and NSP4) were DS-1-like,NSP1 and NSP5 were AU-1-like, and NSP3 was the T6 genotype. The results suggestthe occurrence of multiple reassortment events in the evolution of the strain.Figure (A-K) shows the phylogenetic trees forall the 11 genes and Table I shows the percentagesimilarity of the genes to the most closely related reference strains in Genbank.The phylogenetic analyses and genetic p-distance calculations showed that theVP7 gene of the strain from Honduras clustered in a clade that contains bovine,equine, ovine, and human strains [Figure (A)].The VP4 gene is closest to the bovine strain Sun9, whereas the VP6 gene is closestto bovine UK-tc strain and the human strain ITA-tc/PA169 in a mixed clade ofhuman and bovine strains [Table I Figure(B-C)]. The VP6 gene occupied a basal position to the lineage that containshuman and bovine reference strains including closely-related RotaTeq® strains[Figure (C)]. VP1 is closely related to RotaTeq®strains and a human G6P strain from Australia [TableI Figure (D)], VP2 is most closely relatedto human strain A549 [Table I Figure(E)], and NSP2 is most closely related to a simian RRV strain [TableI Figure (H)]. The NSP5 gene was closestto a Rhesus tissue culture strain RTRV [Table IFigure (K)]. The VP3 gene belongs to a predominantlybovine clade, the NSP1 gene is in a clade with human and bovine strains, theNSP3 gene is in a clade with mixed strains, and the NSP4 gene belongs to a cladewith bovine and equine strains as shown in Figure(F, G, I, J), respectively, and Table I Whileit is a common phenomenon in such studies for genes to be closely related toreference strains with a percent similarity of over 99%, it was interestingto note that the percent similarities of the most closely related strains inGenBank to all the respective genes of the G10P strain from Honduras wereless than 99%. Table II is a comparison of thegenome constellation of the G10P strain from Honduras with other G10Pstrains that have been deposited in GenBank.
The detection ofrare rotavirus strains is a cause for concern due to the implications that theseuncommon strains may have for existing, and yet to be developed, rotavirus vaccines,as to whether the strains will be protected against vaccine-primed immunity.Due to the segmented genome of the virus, different combinations of the segmentscould be obtained from reassortment and result in the formation of new genomeconstellations (Doro et al. 2015 11 ). The detection of such strains during hospital-basedrotavirus surveillance programs, suggest that rare strains are also able tocause disease. Since first reported in animals 2009, the G10P rotavirusgenotype has not been detected frequently (Varshney et al. 2002 34 , Matthijnssenset al. 2009). Previous reports of the G10P strain have linked the originsof the uncommon genotype to animals, such as bovines, of the Mammalian orderArtiodactyla (Ghosh et al. 2007 13 , Steyer et al. 2010 30 , Cowley et al. 2013 5 , Mediciet al. 2015). The most common human rotavirus genotypes in the Latin Americanregion, including Honduras, are G1P, G9P, G2P, and recently G9P(de Oliveira et al. 2009 6 , Linhares et al. 2011 18 , Quaye et al. 2013 28 ).
The genome constellationof the G10P rotavirus strain from Honduras, which was characterised in thisstudy by whole genome sequencing, suggests the occurrence of interspecies transmissionactivities since the virus contains DS-1-like (genogroup 2), AU-1-like (genogroup3) strains, and a T6 genome segment coding for the non-structural protein thatfacilitates translation (NSP3). Strains with DS-1-like signatures are thoughtto be derived from bovine strains, whereas AU-1-like strains are thought tobe originated from dogs and cats (Yamamoto et al. 2011 35 , Doan et al. 2015 10 ). Previouscharacterisation of G10P strains from Slovenia suggested at least two-horizontalinterspecies transmission events that originated through a bovine-human interaction(Steyer et al. 2010 30 ). The VP7, VP6 and NSP4 genes in the Slovenian G10Pstrains were closely related to the corresponding genes for strains that havebeen found in bovines. The genome could have been formed during a co-infectionof a host with different strains of rotaviruses.
The Honduras G10Pstrain is different in gene constellation from the other human G10P strainsthat have been submitted to GenBank. The recently reported wild-type G10Protavirus strains from Vietnam (Do et al. 2017 9 ) and a tissue culture strainfrom Great Britain (Heiman et al. 2008 15 ) have a genomic constellation that issimilar to the G10P from Honduras (Table II.The strains from Vietnam, Great Britain and Honduras has the A3 genotype whereasother strains from other parts of the world including Australia (Cowley et al.2013), India (Mandal et al. 2016 19 ), Italy (Medici et al. 2015 24 ) has the A11 genotypefor the gene that codes for the interferon antagonising non-structural protein,NSP1 (Table II. The observed percent similaritiesof less than 99% for the relatedness of genes obtained in this study to GenBankreference strains suggest accumulation of mutations across all the 11 genesof the Honduras strain (Table I. To the best ofour knowledge, G10 and P strains combination have not been identified inHonduras. However, G10 strains associated with P (Linhares et al. 2011 18 ),and P strains associated with G4 (Tam et al. 2014 31 ) have been reported inthe Latin American region. The G10P strain from Honduras when compared toother G10P strains suggests that the uncommon genotype is novel. Howeverthe report is limited by the fact that there is not that much information ofgenomic data on rotaviruses from animals, and to also deduce if the Hondurasstrain was formed from strains that are present within the Americas or wereintroduced into Latin America from other parts of the world.
The emergence ofuncommon rotavirus strains, such as the G10P strain herein characterised,may have a negative implication for vaccine efficacy and effectiveness. Eventhough the licensed vaccines have been effective against rotavirus strains thatwere originally not targets, the effectiveness cannot be guaranteed for allstrains, especially the ones that are formed from interspecies transmissions.Rotavirus vaccination has been introduced into the national immunisation programsin multiple countries of Latin America and resulted in significant decreasesin childhood diarrhea morbidity and mortality (Desai et al. 2011 8 ). The introductionof the rotavirus vaccine in Honduras in 2009 has yielded a reduction in hospitalisationand mortality by 11-20% (de Oliveira et al. 2013 7 ) compared to observations madein relatively more developed countries which had reductions in the range of70-90% (Cortese et al. 2013 4 , Payne et al. 2013 26 ). The difference in reductioncould be attributable to the presence of rotavirus strains for which the vaccineintroduced is not as effective.
In summary, theresults of the whole genome characterisation of a G10P rotavirus strainfrom Honduras is consistent with the detection of an uncommon genotype witha novel genome constellation. This is the first report of a human G10P strainfrom the Latin American region, even though G10 and P strains associatedwith other VP4 and VP7 genotypes, respectively, have been detected. The occurrenceof the rare strain with gene segments that have animal origins suggest interspeciestransmission of rotavirus strains and may have resulted from a co-infectionof a common host. The continuous detection of uncommon strains in the post vaccineintroduction era emphasizes the need for monitoring the emergence of new rotavirusstrains.
OQ - Conceptualisedthe study, did laboratory analysis, analysed the data, and wrote the manuscript;SR - analysed the data and edited the manuscript; KR - did laboratory analysis,analysed the data, and edited the manuscript; MDE - did laboratory analysis,analysed the data and edited the manuscript; ZX - did laboratory analysis andanalysed the data; KIT - did laboratory analysis and analysed the data; DJCB- collected samples and did laboratory analysis; GRB - conceptualised the study,collected samples and did laboratory analysis; MDB - conceptualised the study,analysed the data, and edited the manuscript.
The authors ofthis study declare that they have no conflict of interest, financial or otherwise,related to this article. The findings and conclusions in this report are thoseof the author(s) and do not necessarily represent the official position of theCenters for Disease Control and Prevention. Names of specific vendors, manufacturers,or products are included for public health and informational purposes; inclusiondoes not imply endorsement of the vendors, manufacturers, or products by theCenters for Disease Control and Prevention or the US Department of Health andHuman Services.