The Origins of G12P[6] Rotavirus Strains Detected in Lebanon

RESEARCH ARTICLE Reslan et al., Journal of General Virology DOI 10.1099/jgv.0.001535

The origins of G12P[6] rotavirus strains detected in Lebanon

Lina Reslan1,2†, Nischay Mishra3†, Marc Finianos1, Kimberley Zakka1, Amanda Azakir1,4, Cheng Guo3, Riddhi Thakka3, Ghassan Dbaibo1,2, W. Ian Lipkin3,* and Hassan Zaraket1,4,*

Abstract The G12 rotaviruses are an increasingly important cause of severe diarrhoea in infants and young children worldwide. Seven human G12P[6] rotavirus strains were detected in stool samples from children hospitalized with gastroenteritis in Lebanon during a 2011–2013 surveillance study. Complete genomes of these strains were sequenced using VirCapSeq-VERT, a capture- based high-throughput viral-sequencing method, and further characterized based on phylogenetic analyses with global RVA and vaccine strains. Based on the complete genomic analysis, all Lebanese G12 strains were found to have Wa-like genetic backbone G12-P[6]-I1-R1-C1-M1-A1-N1-T1-E1-H1. Phylogenetically, these strains fell into two clusters where one of them might have emerged from Southeast Asian strains and the second one seems to have a mixed backbone between North Ameri- can and Southeast Asian strains. Further analysis of these strains revealed high antigenic variability compared to available vaccine strains. To our knowledge, this is the first report on the complete genome-based characterization of G12P[6] emerging in Lebanon. Additional studies will provide important insights into the evolutionary dynamics of G12 rotaviruses spreading in Asia.

INTRODUCTION (glycoprotein), are found on the double-shelled capsid Group A rotavirus (RVA) infection is the global leading structure, and are used as determinants of serotype specificity cause of severe, dehydrating diarrhoea in children younger [3]. Currently, more than 60 G/P combinations have been than 5 years. Although the pathogenic burden of rotavirus identified in human infections globally with the following has decreased during the past decade, rotaviruses were genotypes accounting for the majority of the strains: G1, G2, nonetheless responsible for nearly 130 000 deaths annu- G3, G4, G9 and G12, in association with P[4], P[6], P [8] ally [1]. In Lebanon, RVA accounts for nearly one third of [4]. RVA’s segmented RNA genome and error-prone RNA- gastroenteritis-associated hospitalizations in children under dependent RNA polymerase contribute to its high diversity 5 years of age [2]. through reassortment, recombination and mutations. Reas- sortment between human and animal viruses reveal that RVA is a non-enveloped double-stranded RNA virus of the animal-to-human transmission contributes to the increased Reoviridae family. Its genome is composed of 11 segments, RVA diversity [5–7]. coding for six structural viral proteins (VPs) and six non- structural proteins (NSPs). Each segment codes for one G12 strains have been increasingly detected globally, viral protein, except for segment 11, that codes for two becoming the sixth major human G genotype [8, 9]. They proteins (NSP5 and NSP6). Two of the viral proteins, VP4 have emerged predominantly in combination with either P[6] (non-glycosylated protease-sensitive protein) and VP7 or P[8], and less commonly with P[4] or P[9] genotype. In

Received 21 August 2020; Accepted 18 November 2020; Published 17 December 2020 Author affiliations: 1Center for Infectious Diseases Research, American University of Beirut, Faculty of Medicine, Beirut, Lebanon; 2Department of Pediatrics and Adolescent Medicine, American University of Beirut, Faculty of Medicine, Beirut, Lebanon; 3Center for Infection and the Immunity, Mailman School of Public Health, Columbia University, NY 10032, New York; 4Department of Experimental Pathology, Immunology and Microbiology, Faculty of Medicine, Beirut, Lebanon. *Correspondence: Hassan Zaraket, hz34@ aub. edu. lb; W. Ian Lipkin, wil2001@ cumc. columbia. edu Keywords: Human Rotavirus A; G12P[6]; Genome; Lebanon; VirCapSeq-VERT. Abbreviations: MENA, Middle East and North Africa; NSPs, non-structural proteins; RVA, Group A rotavirus; USA, United States of America; VPs, viral proteins. The nucleotide sequences of the full-length genome of G12P[6] strains of this study were deposited in GenBank under accession numbers: M13 (MN746056-MN746066); M21 (MN746067-MN746077); M23 (MN746078-MN746088); R99 (MN746089-MN746099), R80 (MN746100-MN746110); H280 (MN746045-MN746055); and H270 (MN746034-MN746034). †These authors contributed equally to this work One supplementary table is available with the online version of this article.

Table 1. Demographic and clinical characteristics of patients with G12P[6]

Vomiting Diarrhoea Severity of dehydration

Patient Specimen collection Age (m/d)* Gender Temperature Episodes/day Duration Episodes/day Duration (%) Vesikari ID date (mm/yy)* (°C) (days) (days) score

M13 02/2011 12 d Male 38 3 1 3 4 1–5 % 11

M21 03/2011 12 m Male 39 5 3 4 3 1–5 % 18

M23 03/2011 6 m Female 39 5 2 3 7 1–5 % 17

R99 04/2013 31 m Male 38.5 8 3 4 2 1–5 % 15

R80 06/2012 11 m Female 39.5 4 3 4 5 1–5 % 17

H280 02/2013 8 m Male 38.5 4 2 6 1 1–5 % 14

H270 01/2013 35 m Male 40 3 4 5 7 1–5 % 17

*m, month; y, year; *d, day.

the Middle East and North Africa (MENA), G12 strains have gastroenteritis was assessed using the previously described been reported in several countries; Saudi Arabia [10], Yemen 20-point Vesikari scoring system manual [22]. [11], Iraq [12], Iran [13], Tunisia [14, 15], Egypt [16], Pakistan [17] and Turkey [18]. Whole-genome sequencing An extended classification and nomenclature system was VirCapSeq-VERT library preparation and sequencing were implemented by the Rotavirus Classification Working Group, performed as previously described [23]. Total nucleic acid to define genotypes of each genome segment. The notation was extracted from stool suspension using the NucliSENS Gx-P[x]-Ix-Rx-Cx-Mx-Ax-Nx-Tx-Ex-Hx (‘x’ representing easyMAG automated platform (BioMérieux, Boxtel, The the genotype number) denotes the VP7-VP4-VP6-VP1-VP2- Netherlands) followed by cDNA synthesis. The libraries VP3-NSP1-NSP2-NSP3-NSP4-NSP5 genes, respectively. To were prepared using the Hyper Prep kit (KAPA Biosystems, date, at least 35 G, 50 P, 26 I, 21 R, 19 C, 19 M, 30 A, 20 N, Boston, MA, USA) and unique barcodes and hybridized 21T, 26E and 21 H genotypes have been described (http:// with the VirCapSeq-VERT probe set prior to a final PCR rega.kuleuven.be/cev/viralmetagenomics/virus-classifica- and sequencing (Illumina HiSeq 4000). The obtained reads tion). Worldwide, the majority of human RVA strains possess were then trimmed to remove low-quality sequences and were either genotype 1, constellation Wa-like (I1-R1-C1-M1-A1- assembled using Geneious 11.0.2. Individual gene segments N1-T1-E1-H1), or the genotype 2, constellation DS-1-like were then genotyped using the Virus Pathogen Database (I2-R2-C2-M2-A2-N2-T2-E2-H2) [19]. However, strains of and Analysis Resource (VIPR) (https://www.viprbrc.org/ mixed Wa-like/DS-1-like constellations have also emerged brc/rvaGenotyper.spg?method=ShowCleanInputPage& [20]. decorator=reo). In the present study, we attempted to investigate the genomic diversity and origin of G12P[6] RVA strains that we previ- Phylogenetic analyses ously identified in Lebanon during a prospective, multicenter, Phylogenetic analysis was performed using the maximum- hospital-based surveillance study conducted between 2011 likelihood method after selecting the best-fit evolutionary and 2013 [21]. model for the data set, based on the value of the lowest corrected Akaike information criterion (AICc) using mega 6.06 software [24]. Models used in this study were Tamura METHODS 3-parameter (T92)+G+I (NSP1, NSP2, VP1, VP3, VP7), Sample selection T92+G (VP6, NSP3, NSP4, NSP5/6), Tamura-Nei parameter (TN93)+G (VP2), TN93+G+I (VP4). Statistical analyses were The seven G12P[6] RVA-positive samples included in this performed using bootstrap method with 1000 replicates. study were identified during a surveillance study that was conducted in Lebanon during years 2011 through 2013 to The nucleotide sequences of the full-length genome of G12P[6] determine the incidence of RVA-associated gastroenteritis in strains of this study were deposited in GenBank under hospitalized children under 5 years of age [21]. The G12P[6] accession numbers: M13 (MN746056-MN746066); M21 strains are M21, M13, M23, H280, H270, R80 and R99. (MN746067-MN746077); M23 (MN746078-MN746088); R99 The demographic and clinical features of the patients were (MN746089-MN746099), R80 (MN746100-MN746110); H280 obtained along with the specimens. The severity of RVA (MN746045-MN746055); and H270 (MN746034-MN746034).

2 Reslan et al., Journal of General Virology 2020

Prediction of glycosylation sites H270, R80 and R99 clustered closely with G12P[6] strains The N-linked glycosylation sites (N-X-S or T patterns, where from Thailand and Korea. X can be any amino acid other than proline) were predicted The VP1, VP2, VP3 and VP6 gene sequences of strains M21, using the NetNGlyc 1.0 server (http://www.cbs.dtu.dk/ M13 and M23 had the highest nucleotide sequence identities services/NetNGlyc/). with either Myanmar G12P[6] strain A23 (99.4%–99.5 %) or Nepali G12P[6] strain 10N4594 (99.4%–99.7 %). Phylogeneti- RESULTS cally, VP1, VP2, VP3 and VP6 of these strains clustered with G12P[6] strains from Nepal and Myanmar. The VP6 gene Clinical features of detected Lebanese G12P[6] clustered additionally with two USA strains VU12-13-103 strains and VU12-13-76. The demographic distribution and clinical characteristics of H270, H280, R80 and R99 possessed VP1 genes sharing the children detected with G12P[6] strains are displayed in highest similarity to G1P[8] strains CK00084 from Australia Table 1. In total, seven G12P[6] specimens (out of 428 RVA- (99.4%–99.9 %), 2008747323 from the USA (99.3%–99.4 %) positive stool specimens) were collected from hospitalized and clustered together with G1P[8] and G12P[8] from children with severe gastroenteritis: M13, M21 and M23 were Australia, Italy, Malawi and other countries. Their VP2 gene detected in 2011, R80 in 2012, and R99, H270 and H280 in sequences were closely related to the Canadian G1P[8] strain 2013. None of the patients had received rotavirus vaccines. RT070-09 with which they showed the highest nucleotide The genomic constellation of the Lebanese G12P[6] similarity (97.6 %–97.9 %). Similarly, the VP3 gene sequences strains of H270, H280 and R80 exhibited the highest nucleotide similarity with Canadian G1P[8] strain RT070-09 and Australian A comparison of the complete genotype constellations of G1P[8] strain CK00084 (99.4%). These strains clustered in strains M13, M21, M23, H280, H270, R80 and R99 with those a common branch with Belgian G12P[8] strain B4633. The of other G12 and non-G12 strains is shown in Table 2. Whole- remaining Lebanese strain R99, which showed the highest genome analysis of the seven Lebanese strains revealed a pure nucleotide similarity with USA strain 2008747323 and Korean Wa-like genomic constellation: G12-P[6]-I1-R1-C1-M1- CAU214 strain (98.6%) clustered in a common branch with A1-N1-T1-E1-H1, that represents the predominant genetic other globally circulating G12P[6] (Korea), G1P[6] (Congo) backbone found in circulating G12 strains worldwide. and G1P[8] (USA and Japan) strains. The VP6 genes of H270, The M13, M21 and M23 exhibited high nucleotide sequence H280, R80 and R99 clustered and showed the highest nucleo- identities (99.8–100 %) among each other for the 11 gene tide similarity (99.2%–99.7 %) with Canadian and Australian segments. Similarly, H280, H270, R80 and R99 were very G1P[8] strains (RT070-09) and (CK00084), respectively. similar to each other (96.8–100 %) (Table S1, available in the The NSP [1–4, 8, 9] genes of M21, M13 and M23 showed the online version of this article). The average similarity between highest nucleotide similarity (99.2%–99.8 %) with the Nepal both groups was 95.9% and 94.8 % for the VPs and NSPs, G12P[6] strain 10N4594 and the Myanmar G12[P6] strain respectively. A23. Phylogenetically, NSP1, NSP3 and NSP4 of the Leba- nese strains clustered in a common branch with G12P[6] and Phylogenetic analyses G12P[8] strain circulating in countries like Nepal, Myanmar, To further investigate the genetic relatedness among Lebanese Italy, Kenya, USA and Zimbabwe. NSP2 and NSP5/6 clustered and global G12 strains, we conducted nucleotide similarity with Nepal and Myanmar G12P[6] strains (10N4594 and A23, and phylogenetic analyses using the full-genome sequences respectively) as well as G9P[8] strain MRC-DPRU1723 from for each of the 11 gene segments (Fig. 1, Table S1). Zimbabwe. NSP2 additionally clustered with G12P[8] strains from the USA (VU12-13-76 and VU12-13-103), G12P[11] The VP7 and VP4 genes of M21, M13 and M23 strains exhib- strain from India (MCS-KOL-16) and G11P[8] from Bang- ited the maximum nucleotide sequence identities (98.9% and ladesh. The NSP1-4 gene clusters harbouring the Lebanese 99.4 %, respectively) with Myanmar G12P[6] strain A23. The M21, M13 and M23 strains also accommodated the bovine VP7 gene of H280, H270, R80 and R99 strains showed the G12P[8] strain BUW-14-A035 from Uganda. highest nucleotide sequence similarities (97.3%–99.3 %) with Thailand strain CMHN49-12 (G12P[6]); whereas their VP4 The NSP1 genes of H270, H280, R80 and R99 strains showed genes showed a maximum similarity (99.2 %–99.4 %) with the highest similarity (98.6%–99.2 %) with Korean G12P[6] the Korean strain CAU 214 (G12P[6]). Phylogenetically, strain CAU214 and clustered together with G1P[8] strains the VP7 genes of M21, M13 and M23 were closely related to from the USA (2008747323) and Canada (RT070-09). NSP2 G12P[6] strains from Bangladesh, Myanmar, Nepal and South and NSP3 exhibited the highest similarity with Canadian Africa, and G12P[8] strains from the USA. These samples G1P[8] strain RT070-09 (99.2%–99.7 %). Phylogenetically, were co-clustered with Uganda bovine G12P[8] strain (BUW- NSP2 clustered with the Canadian and American G1P[8] 14-A035) as well. The VP4 genes of M21, M13 and M23 strains (RT070-09 and 2008747323) and Argentinian G4P[8] tightly clustered with G12P[6] strains from different countries strain Tuc1650. NSP3 clustered together with the Canadian such as Myanmar, Nepal, South Africa and the Democratic and Australian G1P[8] strains (RT070-09 and CK00084). Republic of the Congo. The VP7 and VP4 genes of the H280, NSP4 exhibited the highest nucleotide similarity with

3 Reslan et al., Journal of General Virology 2020

Fig. 1. Phylogenetic analysis of the 11 gene segments of the Lebanese G12P[6] RVA strains and representative RVA strains. The phylogenetic trees were inferred using the maximum-likelihood analysis based on the best-fit nucleotide substitution model for each individual gene. The Lebanese strains are in bold, whereas those of the animal and vaccine strains are underlined and italicized, respectively. Bootstrap values equal or greater than 70 % are shown. Scale bars indicate the substitutions per nucleotide.

American G1P[8] strain 2 008 747 323 (99.2 %) and clustered Korea (CAU 214), G1P[8] strain from Japan (MU14-19), and with G1P[8] strains from Australia and North America. G12P[8] strain from Belgium (B4633). Finally, the NSP5/6 gene sequences of H270, H280, R80 and It is interesting to note that the Lebanese G12P[6] strains, R99 clustered with Bangladeshi G11P[8] strain Matlab36 M21, M13 and M23 showed high similarity with the bovine (nucleotide similarity 98.4–98.8 %), G12P[6] strain from

4 Reslan et al., Journal of General Virology 2020

Table 2. Comparison of the genotype constellations of the Lebanese G12 with a globally representative sample of RVA

Strain names VP7 VP4 VP6 VP1 VP2 VP3 NSP1 NSP2 NSP3 NSP4 NSP5/6 Type

RVA/Human-wt/LBN/M21/2011/G12P[6] G12 P[6] I1 R1 C1 M1 A1 N1 T1 E1 H1 Wa-Like

RVA/Human-wt/LBN/M23/2011/G12P[6] G12 P[6] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/LBN/M13/2011/G12P[6] G12 P[6] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/LBN/H280/2013/G12P[6] G12 P[6] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/LBN/R80/2012/G12P[6] G12 P[6] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/LBN/H270/2013/G12P[6] G12 P[6] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/LBN/RO99/2013/G12P[6] G12 P[6] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/ZAF/3133WC/2009/G12P[4] G12 P[4] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/NPL/5N0014/2005/G12P[6] G12 P[6] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/ZAF/MRC-DPRU2130-05/2005/ G12 P[6] I1 R1 C1 M1 A1 N1 T1 E1 H1 G12P[6]

RVA/Human-wt/BGD/Dhaka12/2003/G12P[6] G12 P[6] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/ZAF/MRC-DPRU1911/2007/ G12 P[6] I1 R1 C1 M1 A1 N1 T1 E1 H1 G12P[6]

RVA/Human-wt/NPL/5N0109/2005/G12P[6] G12 P[6] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-tc/MMR/A23/2011/G12P[6] G12 P[6] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/NPL/10N4594/2010/G12P[6] G12 P[6] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/ARG/SF931/1998/G4P[8] G4 P[8] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/ARG/Tuc1650/1998/G4P[8] G4 P[8] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/BGD/Matlab36/2002/G11P[8] G11 P[8] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/BEL/B4633/2003/G12P[8] G12 P[8] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/JPN/MU14-19/2014/G1P[8] G1 P[8] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/BGD/Bang-130/2008/G1P[8] G1 P[8] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/KOR/CAU 214/2006/G12P[6] G12 P[6] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/COD/KisB504/2009/G1P[6] G1 P[6] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/ZWE/MRC-DPRU1723/2009/ G9 P[8] I1 R1 C1 M1 A1 N1 T1 E1 H1 G9P[8]

RVA/Cow-wt/UGA/BUW-14-A035/2014/G12P[8] G12 P[8] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/USA/VU12-13-76/2012/G12P[8] G12 P[8] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/USA/VU12-13-103/2013/G12P[8] G12 P[8] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-tc/KEN/KDH651/2010/G12P[8] G12 P[8] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/ZWE/MRC-DPRU1858/2011/ G12 P[8] I1 R1 C1 M1 A1 N1 T1 E1 H1 G12P[8]

RVA/Human-wt/AUS/CK00084/2008/G1P[8] G1 P[8] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/CAN/RT070-09/2009/G1P[8] G1 P[8] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/ITA/RG179/2013/G12P[8] G12 P[8] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Human-wt/USA/2008747323/2008/G1P[8] G1 P[8] I1 R1 C1 M1 A1 N1 T1 E1 H1

RVA/pig-wt/IND/RU172/2005/G12P[7] G12 P[7] I5 R1 C1 M1 A1 N1 T1 E1 H1

RVA/Pig-wt/UGA/BUW-14-A003/2014/G3P[13] G3 P[13] I1 R1 C1 M1 A8 N1 T7 E1 H1

RVA/Pig-wt/UGA/KYE-14-A047/2014/G3P[13] G3 P[13] I1 R1 C1 M1 A8 N1 T1 E1 H1

RVA/Human-wt/IND/MCS-KOL-16/2014/G12P[11] G12 P[11] I1 R1 C1 M1 A5 N1 T1 E1 H1 Continued

5 Reslan et al., Journal of General Virology 2020

Table 2. Continued

Strain names VP7 VP4 VP6 VP1 VP2 VP3 NSP1 NSP2 NSP3 NSP4 NSP5/6 Type

RVA/Human-wt/ZAF/MRC-DPRU9317/1999/G9P[6] G9 P[6] I2 R2 C2 M2 A2 N1 T2 E2 H2 DS1-like

RVA/Human-wt/GER/GR10924/1999/G9P[6] G9 P[6] I2 R2 C2 M2 A2 N2 T2 E2 H2

RVA/Human-wt/BGD/RV161/2000/G12P[6] G12 P[6] I2 R2 C2 M2 A2 N2 T2 E1 H2

RVA/Human-wt/MWI/BID1LN/2013/G1P[8] G1 P[8] I2 R1 C2 M2 A2 N2 T2 E2 H2

RVA/Human-wt/THA/CMHN49-12/2012/G12P[6] G12 P[6] I2 R2 C2 M2 A2 N2 T2 E2 H2

RVA/Vaccine/USA/Rotarix-A41CB052A/1988/ G1 P1A[8] I1 R1 C1 M1 A1 N1 T1 E1 H1 Vaccine G1P1A[8]

RVA/Vaccine/USA/RotaTeq-SC2-9/1992/G2P7[5] G2 P7[5] I2 R2 C2 M1 A3 N2 T6 E2 H3

RVA/Vaccine/USA/RotaTeq-WI78-8/1992/G3P7[5] G3 P7[5] I2 R2 C2 M2 A3 N2 T6 E2 H3

RVA/Vaccine/USA/RotaTeq-WI79-4/1992/G6P1A[8] G6 P1A[8] I2 R2 C2 M2 A3 N2 T6 E2 H3

Grey color indicates the Lebanese strains (LBN) characterized in this study. Genes with Wa-like genotype are in bold. Light green indicates genome segments with the highest nucleotide sequence identity to M21, M13 and M23; red indicates genome segments with the highest nucleotide sequence identity to H270, H280 R80 and R99; cyan indicates the highest nucleotide sequence identity to R99, orange indicates the highest nucleotide sequence identity to H270, H280, R80.

G12P[8] strain BUW-14A035 in VP3 and all NSP genes arginine residues are highly conserved in RVA strains inde- except for NSP4 with nucleotide identities ≥98 %. pendent of genotype; cleavage at arginine 246 correlates with enhanced infectivity [29–31]. In addition, another potential Taken together, these observations suggest that the genomic trypsin cleavage site (lysine) was observed at residue 257 [31] constellations of the Wa-like Lebanese G12P[6] M13, M21, as well as a proline at positions 68, 71, 224 and 225, and a M23 strains largely resemble that of Southeast Asian Wa-like cysteine at position 215 (data not shown). These residues have strains. While, the Lebanese H270, H280, R80 and R99 been described as highly conserved in the VP8* gene portion G12P[6] strains seem to have a mixed backbone between of human RV strains [32]. They were also conserved in all G12 Southeast Asian and North American strains, with the NSP2, strains included in this study. 3 and 4 segments originating from North American strains. Twelve potential N-glycosylation sites were identified Comparison of VP4 and VP7 antigenic epitopes of among the Lebanese VP4 components where three of them G12P[6] Lebanese rotavirus strains and rotavirus were located at positions 114, 132 and 192 within the VP8* vaccines epitopes, and three others are located in proximity to anti- We next performed amino acid sequence alignements to genic sites at positions 85, 97 and 111 in both Lebanese and compare the composition of the VP7 and VP4 antigenic global G12 strains. epitopes of the Lebanese G12P[6] strains with those of the The VP7 protein consists of two antigenic epitopes: 7–1 (7-1a vaccine strains and G12P[6] circulating globally (Fig. 2). The and 7-1b) and 7–2 encompassing 29 amino acid residues [1]. amino acid residues constituting the antigenic epitopes were Eleven of the 29 epitope residues were conserved among Leba- previously determined by mapping neutralization escape nese and global G12 strains. Six amino acids were consistently mutants and identifying surface-exposed amino acids that divergent from the VP7 components of both the Rotarix and differ among prevalent human P and G types [25–27]. RotaTeq vaccines. These substitutions were dispersed across The activation of the VP4 spike protein requires its proteolytic all the VP7 epitopes and are therefore expected to affect the cleavage into VP5* and VP8*. The VP8* fragment forms a antigenic properties of the viruses [33]. globular head and theVP5* forms the stalk domain. VP8* All the VP7 proteins of the study strains contained the and VP5* consist of four (8–1 to 8–4) and five (5–1 to 5–5) conserved proline, cysteine and glutamine residues identi- surface-exposed antigenic epitopes, respectively, comprised fied previously [34, 35]. The VP7 proteins of our G12P6 of 37 amino acids (Fig. 2). We identified 19 residues (51.3 %) strains possessed two potential N-linked glycosylation sites that were consistently divergent from the VP4 components at positions 69 and 238 [5], required for disulfide bond forma- of both the Rotarix and RotaTeq vaccines. The differences tion and proper folding of the protein in the endoplasmic were mainly concentrated in all VP8* epitopes and in 5–1 reticulum and, ultimately, viral assembly [36]. and 5–3 VP5* epitopes. Interestingly, 11 amino acids were consistently conserved among the Lebanese and global G12 strains in all VP5* and VP8* epitopes except the 5–3 and 8–4 DISCUSSION ones compared to both Rotarix and RotaTeq components. In this study, we characterized the full genomes of seven All of the Lebanese G12P[6] rotaviruses analysed in this study Lebanese G12P[6] strains that were sporadically detected in contained potential trypsin cleavage sites (arginine) at posi- Lebanon over the interval of 2011 through 2013 [21]. The tions 230, 240, 246, 466 and 581 [28]. The aforementioned Lebanese G12P[6] strains exhibited typical Wa-like genotype

6 Reslan et al., Journal of General Virology 2020

Fig. 2. Alignment of the amino acid residues in VP4 and VP7 antigenic epitopes of the Lebanese G12P[6] RVA strains and those of vaccine strains. Blue coloured residues are residues that are different from Rotarix, green coloured residues are residues that are different from RotaTeq, and orange coloured residues are those different from both vaccines. Amino acid residues known to mediate escape from neutralization are indicated by triangle (∆). Brackets indicate the number of specimens having the indicated amino acid substitution.

constellation, commonly detected worldwide [9, 37]. Another, reported in Saudi Arabia (7 %) [10], Yemen (2 %) [11], Iraq less prevalent G12P[6] lineage within the DS-1-like geno- (7 %) [12], Iran (9.4 %) [13], Egypt (9.2 %) [16], Tunisia group has also been reported in many countries including (5.2 %) [14, 15], Pakistan (12.85 %) [17, 54] and Turkey Bangladesh, Malawi and Thailand [8, 37, 38]. G12 strains (15.6 %) [18]. In Lebanon, G12 accounted for 1.6 % of all has spread globally showing a steady mutation rate with the rotavirus infections [21], where the sole genotype detected potential to diversify rapidly [39]. was G12P[6]. G12 is an emerging rotavirus genotype usually detected as The Lebanese G12P[6] strains, despite having similar G12P[6] or G12P[8], and less frequently with G12P[4] or genotype constellations, consistently fell into two distinct G12P[9] [7, 8, 39]. It is subdivided into four lineages where clusters suggesting that each cluster was independently the majority of the currently known G12 strains (more introduced into Lebanon. The first cluster, M13, M21 and than 90 % and in combination with G12P[8] or G12P[6]) M23 was isolated at the same hospital within a 1 month cluster in a single lineage (lineage III). This lineage was interval between M13 and both M21 and M23. All of their able to spread across the globe and cause disease in little genes were closely related to G12P[6] strains A23 and more than a decade [8]. G12 rotaviruses were first detected 10N4594 detected in Myanmar and Nepal, respectively. in the Philippines [40]. They were subsequently isolated Nepal has contributed to the United Nations Interim Force in the USA [41], Argentina and Brazil [9, 42], as well as in Lebanon since 2000 (www.unifil.unmissions.org). This in several Asian countries including Nepal [43, 44], Thai- might partially explain the origin of the G12 isolated and land [45], India [46, 47], Bangladesh [48], Vietnam [49], their association with strains detected in Southeast Asia. Korea [50] and Myanmar [51]. In Europe, G12 strains were The second cluster including H280, H270, R80 and R99 identified in the UK, Belgium and Spain [52]. Moreover, was detected in 2012 and 2013 in different cities. Our data G12P[8] and G12P[6] were the most prevalent genotypes suggest that the Lebanese G12P[6] strains might have being responsible for over 43 % of all rotavirus infections a mixed origin with both a Southeast Asian and North in Ivorian children [38, 53]. In the MENA region, G12 was American background.

7 Reslan et al., Journal of General Virology 2020

The emergence of a genotype different from those found in Conflicts of interest licensed rotavirus vaccines raises the concern of decreased The authors declare that there are no conflicts of interest. heterotypic protection. Both vaccines, Rotarix and RotaTeq, Ethical statement have been reported to protect against the common VP7 Patient recruitment and stool specimen collection were approved by the American University of Beirut Institutional Review Board. Unique (G1, G2, G3, G4 and G6) and VP4 genotypes (P[4], P[6] study identification codes were used in order to ensure confidentiality. and P) [6, 55]. Vaccine use can impose selective pressure on A written consent was obtained from the patients’ parents prior to circulating strains that results in a shift toward less common sample collection. genotypes or selection of mutant viruses that are not References effectively neutralized [56]. Convalescent sera of G12P[6]- 1. Troeger C, Khalil IA, Rao PC, Cao S, Blacker BF et al. Rotavirus infected children demonstrated heterotypic response vaccination and the global burden of rotavirus diarrhea among against 116E (G9P[11]) and Wa-like (G1P[8]) strains (50% children younger than 5 years. JAMA Pediatr 2018;172:958–965. 2. Dbaibo G, Rajab M, Inati A, Mikhael R, Choueiry E et al. Hospital- and 33.3 %, respectively). The mean neutralizing antibody Based surveillance study of rotavirus gastroenteritis in children titre in convalescent sera of G12P[6]-infected children was under 5 years of age in Lebanon. Trials in Vaccinology2013;2:Pages eightfold higher against strains belonging to the Wa-like 25–30. genogroup (i.e. G1, G9 and G12 rotavirus) than against 3. Crawford SE, Ramani S, Tate JE, Parashar UD, Svensson L et al. strains belonging to the DS-1-like genogroup (G2 rota- Rotavirus infection. Nat Rev Dis Primers 2017;3:17083. virus). The same study demonstrated that neutralization 4. Matthijnssens J, Bilcke J, Ciarlet M, Martella V, Bányai K et al. Rotavirus disease and vaccination: impact on genotype diversity. in part may be genogroup specific, and thus a monovalent Future Microbiol 2009;4:1303–1316. vaccine based on the Wa-like genogroup is likely to confer 5. Jere KC, Mlera L, O'Neill HG, Potgieter AC, Page NA et al. Whole protection against G12 RVs [57]. genome analyses of African G2, G8, G9, and G12 rotavirus strains using sequence-independent amplification and 454® pyrose- Numerous changes were detected in the antigenic epitopes quencing. J Med Virol 2011;83:2018–2042. of the VP4 and VP7 proteins of the G12 strains with respect 6. Bwogi J, Jere KC, Karamagi C, Byarugaba DK, Namuwulya P et al. to the vaccine strains. The extent to which these differences Whole genome analysis of selected human and animal rotaviruses result in the reduction of heterotypic vaccine efficacy is not identified in Uganda from 2012 to 2014 reveals complex genome reassortment events between human, bovine, caprine and porcine known. A previous study in the USA estimated the vaccine strains. PLoS One 2017;12:e0178855. efficacy of Rotarix and RotaTeq against G12P[8] strains 7. Gentsch JR, Laird AR, Bielfelt B, Griffin DD, Banyai K et al. Serotype to be 82% and 78%, respectively [58]. This high vaccine diversity and reassortment between human and animal rotavirus efficacy could be partially attributed to the homotypic VP4 strains: implications for rotavirus vaccine programs. J Infect Dis component [33, 58]. The VP4 P[6] differs substantially from 2005;192 Suppl 1:S146–159. 8. Matthijnssens J, Heylen E, Zeller M, Rahman M, Lemey P et al. that of the P[8] component of the vaccines, which could Phylodynamic analyses of rotavirus genotypes G9 and G12 compromise vaccine efficacy. However, internal segments underscore their potential for swift global spread. Mol Biol Evol of the vaccine strains might provide heterotypic immunity. 2010;27:2431–2436. VP6 induced non-neutralizing antibodies were shown to 9. Rahman M, Matthijnssens J, Yang X, Delbeke T, Arijs I et al. Evolu- prevent rotavirus infection in mice [59]. Other structural tionary history and global spread of the emerging G12 human rotaviruses. J Virol 2007;81:2382–2390. and non-structural proteins might also contribute to T-cell 10. Aly M, Al Khairy A, Al Johani S, Balkhy H. Unusual rotavirus geno- mediated protection [60, 61]. Therefore, immunological types among children with acute diarrhea in Saudi Arabia. BMC pressure alone might not explain the emergence and in Infect Dis 2015;15:192. some cases predominance of G12 strains in different parts 11. Kirby A, Al-Eryani A, Al-Sonboli N, Hafiz T, Beyer M et al. Rotavirus of the world. The changing dynamics of rotavirus genotypes and norovirus infections in children in Sana'a, Yemen. Trop Med Int globally might be affected by the natural evolution and Health 2011;16:680–684. 12. Ahmed S, Klena J, Albana A, Alhamdani F, Oskoff J et al. Char- selection of certain genomic constellations with enhanced acterization of human rotaviruses circulating in Iraq in 2008: or weakened fitness that could allow one genotype to prevail atypical G8 and high prevalence of P[6] strains. Infect Genet Evol over others. The effect of various genomic constellations on 2013;16:212–217. the fitness of rotavirus needs further research. 13. Azaran A, Makvandi M, Teimoori A, Ebrahimi S, Heydari F et al. Distribution of rotavirus genotypes Ccirculating in Ahvaz, Iran in Continued molecular surveillance and simultaneous moni- 2016. Iran Biomed J 2018;22:107–116. toring of rotavirus emerging rotavirus strains such as G12 are 14. Moussa A, Ben Hadj Fredj M, Fodha I, BenHamida-Rebaï M, essential for a better understanding of their origins and evolu- Kacem S et al. Distribution of rotavirus VP7 and VP4 genotypes circulating in Tunisia from 2009 to 2014: emergence of the geno- tion. Consequently, studies examining the efficacy of Rotarix type G12. J Med Microbiol 2016;65:1028–1037. and RotaTeq against these emerging strains are warranted to 15. Moussa A, Fredj MBH, BenHamida-Rebaï M, Fodha I, Boujaafar N maintain the effectiveness of vaccination. et al. Phylogenetic analysis of partial VP7 gene of the emerging human group A rotavirus G12 strains circulating in Tunisia. J Med Microbiol 2017;66:112–118. Funding information 16. Kamel AH, Ali MA, El-Nady HG, de Rougemont A, Pothier P et al. Funding support to CII team (N.M., R.T., C.G., W.I.L.) and financial support Predominance and circulation of enteric viruses in the region of for VirCapSeq-VERT analysis was provided by the National Institutes of greater Cairo, Egypt. J Clin Microbiol 2009;47:1037–1045. Health award U19AI109761 [CETR: Centre for Research in Diagnostics and Discovery].

8 Reslan et al., Journal of General Virology 2020

17. Alam MM, Malik SA, Shaukat S, Naeem A, Sharif S et al. Genetic 38. Motayo BO, Oluwasemowo OO, Olusola BA, Opayele AV, Faneye AO. characterization of rotavirus subtypes in Pakistan-first report of Phylogeography and evolutionary analysis of African rotavirus a G12 genotype from Pakistan under WHO-Eastern Mediterranean genotype G12 reveals district genetic diversification within lineage region. Virus Res 2009;144:280–284. III. Heliyon 2019;5:e02680. 18. Aydin H, Aktaş O, Aktas O. Rotavirus genotypes in children with 39. Bányai K, László B, Duque J, Steele AD, Nelson EAS et al. System- gastroenteritis in Erzurum: first detection of G12P[6] and G12P[8] atic review of regional and temporal trends in global rotavirus genotypes in Turkey. Prz Gastroenterol 2017;12:122–127. strain diversity in the pre rotavirus vaccine era: insights for under- 19. Matthijnssens J, Van Ranst M. Genotype constellation and evolu- standing the impact of rotavirus vaccination programs. Vaccine tion of group A rotaviruses infecting humans. Curr Opin Virol 2012;30 Suppl 1:A122–130. 2012;2:426–433. 40. Taniguchi K, Urasawa T, Kobayashi N, Gorziglia M, Urasawa S. 20. Jere KC, Chaguza C, Bar-Zeev N, Lowe J, Peno C et al. Emer- Nucleotide sequence of VP4 and VP7 genes of human rotavi- gence of Double- and Triple-Gene Reassortant G1P[8] Rotaviruses ruses with subgroup I specificity and long RNA pattern: implica- Possessing a DS-1-Like Backbone after Rotavirus Vaccine Intro- tion for new G serotype specificity. J Virol 1990;64:5640–5644. duction in Malawi. J Virol 2018;92 [Epub ahead of print 01 02 2018]. 41. Wylie KM, Stanley KM, TeKippe EM, Mihindukulasuriya K, 21. Ali Z, Harastani H, Hammadi M, Reslan L, Ghanem S et al. Rotavirus Storch GA. Resurgence of rotavirus genotype G12 in St. Louis genotypes and vaccine effectiveness from a sentinel, hospital- during the 2014-2015 rotavirus season. J Pediatric Infect Dis Soc based, surveillance study for three consecutive rotavirus seasons 2017;6:346–351. in Lebanon. PLoS One 2016;11:e0161345. 42. Castello AA, Argüelles MH, Rota RP, Olthoff A, Jiang B et al. 22. Lewis K. Vesikari clinical severity scoring manual. PATH 2011. Molecular epidemiology of group A rotavirus diarrhea among children in Buenos Aires, Argentina, from 1999 to 2003 and 23. Mishra N, Reslan L, El-Husseini M, Raoof H, Finianos M et al. Full emergence of the infrequent genotype G12. J Clin Microbiol genome characterization of human G3P[6] and G3P[9] rotavirus 2006;44:2046–2050. strains in Lebanon. Infect Genet Evol 2020;78:104133. 43. Pun SB, Nakagomi T, Sherchand JB, Pandey BD, Cuevas LE et al. 24. Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. MEGA6: Detection of G12 human rotaviruses in Nepal. Emerg Infect Dis molecular evolutionary genetics analysis version 6.0. Mol Biol Evol 2007;13:482–484. 2013;30:2725–2729. 44. Sherchand JB, Nakagomi O, Dove W, Nakagomi T, Yokoo M et al. 25. Aoki ST, Settembre EC, Trask SD, Greenberg HB, Harrison SC Molecular epidemiology of rotavirus diarrhea among children et al. Structure of rotavirus outer-layer protein VP7 bound with a aged <5 years in nepal: predominance of emergent G12 strains neutralizing Fab. Science 2009;324:1444–1447. during 2 years. J Infect Dis 2009;200 Suppl 1:S182–187. 26. Dormitzer PR, Nason EB, Prasad BVV, Harrison SC. Structural 45. Khananurak K, Vutithanachot V, Simakachorn N, Theamboon- rearrangements in the membrane penetration protein of a non- lers A, Chongsrisawat V et al. Prevalence and phylogenetic anal- enveloped virus. Nature 2004;430:1053–1058. ysis of rotavirus genotypes in Thailand between 2007 and 2009. 27. Dormitzer PR, Sun Z-YJ , Wagner G, Harrison SC. The rhesus rota- Infect Genet Evol 2010;10:537–545. virus VP4 sialic acid binding domain has a galectin fold with a 46. Kang G, Desai R, Arora R, Chitamabar S, Naik TN et al. Diversity novel carbohydrate binding site. Embo J 2002;21:885–897. of circulating rotavirus strains in children hospitalized with diar- 28. Estes M, Kapikian AZR, Knipe IDM, Howley PM, Griffin RA et al. rhea in India, 2005-2009. Vaccine 2013;31:2879–2883. Fields virology, 2, 5th ed. Philadelphia: Lippincott Williams Wolters 47. Mukherjee A, Chattopadhyay S, Bagchi P, Dutta D, Singh NB et al. Kluwer; 2007. pp. 1917–1974. Surveillance and molecular characterization of rotavirus strains 29. Gilbert JM, Greenberg HB. Cleavage of rhesus rotavirus VP4 after circulating in Manipur, north-eastern India: increasing preva- arginine 247 is essential for rotavirus-like particle-induced fusion lence of emerging G12 strains. Infect Genet Evol 2010;10:311–320. from without. J Virol 1998;72:5323–5327. 48. Rahman M, Sultana R, Ahmed G, Nahar S, Hassan ZM et al. Preva- 30. Arias CF, Romero P, Alvarez V, López S. Trypsin activation pathway lence of G2P[4] and G12P[6] rotavirus, Bangladesh. Emerg Infect of rotavirus infectivity. J Virol 1996;70:5832–5839. Dis 2007;13:18–24. 31. Crawford SE, Mukherjee SK, Estes MK, Lawton JA, Shaw AL 49. Tra My PV, Rabaa MA, Vinh H, Holmes EC, Hoang NVM et al. The et al. Trypsin cleavage stabilizes the rotavirus VP4 spike. J Virol emergence of rotavirus G12 and the prevalence of enteric viruses 2001;75:6052–6061. in hospitalized pediatric diarrheal patients in southern Vietnam. Am J Trop Med Hyg 2011;85:768–775. 32. Vizzi E, Piñeros OA, Oropeza MD, Naranjo L, Suárez JA et al. Human rotavirus strains circulating in Venezuela after vaccine introduc- 50. Le VP, Kim J-Y , Cho S-L, Nam S-W, Lim I et al. Detection of unusual tion: predominance of G2P[4] and reemergence of G1P[8]. Virol J rotavirus genotypes G8P[8] and G12P[6] in South Korea. J Med Virol 2017;14:58. 2008;80:175–182. 33. Ogden KM, Tan Y, Akopov A, Stewart LS, McHenry R et al. 51. Ide T, Komoto S, Higo-Moriguchi K, Htun KW, Myint YY et al. Whole Multiple Introductions and Antigenic Mismatch with Vaccines May Genomic Analysis of Human G12P[6] and G12P[8] Rotavirus Strains Contribute to Increased Predominance of G12P[8] Rotaviruses in that Have Emerged in Myanmar. PLoS One 2015;10:e0124965. the United States. J Virol 2019;93 [Epub ahead of print 01 01 2019]. 52. Cilla G, Montes M, Gomariz M, Alkorta M, Iturzaeta A et al. Rota- 34. Ciarlet M, Hyser JM, Estes MK. Sequence analysis of the VP4, VP6, virus genotypes in children in the Basque Country (North of Spain): VP7, and NSP4 gene products of the bovine rotavirus WC3. Virus rapid and intense emergence of the G12[P8] genotype. Epidemiol Genes 2002;24:107–118. Infect 2013;141:868–874. 35. Stirzaker SC, Whitfeld PL, Christie DL, Bellamy AR, Both GW. 53. Boni-Cisse C, Meite S, Mlan AB, Zaba F, N'Guessan R et al. Geno- Processing of rotavirus glycoprotein VP7: implications for the typic characterization of rotavirus in children under 5 years circu- retention of the protein in the endoplasmic reticulum. J Cell Biol lating in Côte d'Ivoire from 2010 to 2013. Virol J 2018;15:78. 1987;105:2897–2903. 54. Umair M, Salman M, Alam MM, Rana MS, Zaidi SSZ et al. Rotavirus 36. Mirazimi A, Svensson L. Carbohydrates facilitate correct surveillance in Pakistan during 2015-2016 reveals high preva- disulfide bond formation and folding of rotavirus VP7. J Virol lence of G12P[6]. J Med Virol 2018;90:1272–1276. 1998;72:3887–3892. 55. Clarke E, Desselberger U. Correlates of protection against human 37. Nakagomi T, Do LP, Agbemabiese CA, Kaneko M, Gauchan P et al. rotavirus disease and the factors influencing protection in low- Whole-genome characterisation of G12P[6] rotavirus strains income settings. Mucosal Immunol 2015;8:1–17. possessing two distinct genotype constellations co-circulating in 56. Greenberg HB, Estes MK. Rotaviruses: from pathogenesis to vacci- Blantyre, Malawi, 2008. Arch Virol 2017;162:213–226. nation. Gastroenterology 2009;136:1939–1951.

9 Reslan et al., Journal of General Virology 2020

57. Sharma S, Nakagomi T, Nakagomi O, Paul VK, Bhan MK et al. 59. Burns JW, Siadat-Pajouh M, Krishnaney AA, Greenberg HB. Protec- Convalescent phase sera from children infected with G12 rotavirus tive effect of rotavirus VP6-specific IgA monoclonal antibodies that cross-neutralize rotavirus strains belonging to the WA genogroup. lack neutralizing activity. Science 1996;272:104–107. J Gen Virol 2010;91:1794–1799. 60. Ward RL. Rotavirus vaccines: how they work or don't work. Expert 58. Payne DC, Baggs J, Klein NP, Parashar UD. Does preventing Rev Mol Med 2008;10:e5. rotavirus infections through vaccination also protect against 61. Parra M, Herrera D, Jácome MF, Mesa MC, Rodríguez L-S et al. naturally occurring intussusception over time? Clin Infect Dis Circulating rotavirus-specific T cells have a poor functional profile. 2015;60:163–164. Virology 2014;468-470:340–350.

Five reasons to publish your next article with a Microbiology Society journal 1. The Microbiology Society is a not-for-profit organization. 2. We offer fast and rigorous peer review – average time to first decision is 4–6 weeks. 3. Our journals have a global readership with subscriptions held in research institutions around the world. 4. 80% of our authors rate our submission process as ‘excellent’ or ‘very good’. 5. Your article will be published on an interactive journal platform with advanced metrics.

Find out more and submit your article at microbiologyresearch.org.