Phylogenetic analysis of a G6P[5] bovine rotavirus strain isolated in a neonatal diarrhea outbreak in a beef cattle herd vaccinated with G6P[1] and G10P[11] genotypes.

The aim of this study was to perform the molecular characterization of the eleven genes of a G6P[5] bovine group A rotavirus (RVA) strain detected in a diarrhea outbreak from a vaccinated beef cattle herd. The outbreak affected 80 % of calves between 15-30 days old. RVA was identified by RT-PCR in 12 (70.6 %) out of 17 diarrheic fecal samples evaluated. The rotavirus wild-type strain had the genotype constellation G6(IV)-P[5](IX)-I2c-R2-C2-M2-A3-N2-T6-E2e-H3a. This study confirms the importance of homotypic immunity against the bovine RVA P[5] genotype in neonatal diarrhea in cattle herds that are regularly vaccinated against rotaviruses.

Bovine group A rotavirus (RVA) is one of the main etiological agents of neonatal diarrhea in calves worldwide. Morbidity and mortality rates due to bovine RVA infections are high, which in turn cause important direct and indirect economic losses to beef and dairy production [1].

Rotavirus belongs to the family Reoviridae, and it is surrounded by a triple-layered protein capsid. The genome is formed by 11 double-stranded RNA segments, which encode six structural proteins (VP1-VP4, VP6, and VP7) and six non-structural proteins (NSP1-NSP5/6) [2]. Rotaviruses (RVs) are classified into eight distinct groups/species (A-H) [3, 4].

The VP7 and VP4 proteins of RVA are located in the outer layer of the capsid and induce neutralizing antibodies. The antigenic variation of the genome segments that code for these proteins determines the binary RVA genotype classification system [2]. Currently, 27 G (VP7) and 37 P genotypes (VP4) of RVA have been described in mammals and avian species [5, 6].

In 2008, the Rotavirus Classification Working Group (RCWG) defined the notations Gx-P[x]-Ix-Rx-Cx-Mx-Ax-Nx-Tx-Ex-Hx (x-Arabic numbers starting from 1) to VP7-VP4-VP6-VP1-VP2-VP3-NSP1-NSP2-NSP3-NSP4-NSP5/6 genes of RV strains [5]. Until now, the following genotypes have been described: VP6, I1-I17; VP1, R1-R9; VP2, C1-C9; VP3, M1-M8; NSP1, A1-A18; NSP2, N1-N10; NSP3, T1-T12; NSP4, E1-E15; NSP5/6, H1-H11 [6-10].

The most common combinations of G and P genotypes found in bovine RVA strains isolated from diarrhea episodes in calves are G6P[1] (Nebraska calf diarrhea virus [NCDV]-Lincoln), G6P[5] (UK), G8P[1] (A5), and G10P[11] (B223) [11-14]. These genotypes have been described in Brazil with frequencies of 9.7 % (3/31) to 33.3 % (12/36), 8.6 % (3/35) to 40 % (20/50), 16.7 % (6/36), and 12.9 % (4/31) to 16 % (8/50), respectively [15-17]. G6P[5] has been reported to be a common genotype in bovine RVA strains and has been found in cattle herds from Asia [18], Europe [14], Oceania [13], and America [19, 20].

The RVA G6 genotype is divided into five lineages (I–V). The bovine RVA strains can be found in the lineages G6-II, G6-III, G6-IV, and G6-V. The P[5] genotype is divided into eight (VIII) lineages. The lineages P[5]-I, P[5]-II, P[5]-III, P[5]-V, P[5]-VI, P[5]-VII, and P[5]-VIII are composed of bovine strains [21].

Diarrhea in calves can be prevented by appropriate management practices and programs for vaccination of pregnant cows [16]. However, vaccine failures may occur for many reasons, such as reassortment between strains with different genes [22], antigenic or genetic diversity [23], interaction between different or less common genotypes [16, 24, 25], insufficient heterologous immunity [26], and inappropriate vaccination management [27].

The aim of this study was to perform a molecular characterization of the eleven genes of a G6P[5] bovine RVA strain detected in a neonatal diarrhea outbreak in a beef cattle herd that had been regularly vaccinated with G6P[1] and G10P[11] strains.

The beef cattle farm was located in Mato Grosso do Sul state, Central-West Brazil. The herd was managed on extensive pasture and had good health and nutritional practices. The farm had adopted a breeding and, consequently, birthing season of 90 days. In total, 600 cows were divided in lots ranging from 100 to 120 animals. During the neonatal diarrhea outbreak, there were 170 crossbred calves (Nelore × Angus), 15 to 30 days old.

Every cow was regularly vaccinated for neonatal diarrhea control with an inactivated commercial vaccine containing the bovine RVA genotypes G6P[1] (NCDV-Lincoln strain) and G10P[11] (B223 strain), bovine coronavirus (Hanson strain), Clostridium perfringens type C (toxoid/beta toxin, NL-1003 strain), and Escherichia coli pili adherence factor (K99 strain). The vaccination schedule was according to the manufacturer’s instructions.

Diarrheic fecal samples (n = 17) were collected from calves aged less than 30 days old. The fecal samples were stored at −20 °C until analysis. The study was submitted to the Ethics Committee on Animal Experiments of the Universidade Estadual de Londrina and approved (46/09).

Nucleic acid was extracted using a combination of phenol/chloroform/isoamyl alcohol (25:24:1) and silica/guanidinium isothiocyanate [28] extraction methods.

All diarrheic fecal samples was subjected to RT-PCR assay with previously reported consensus primers to amplify VP7 (1,062 bp) and VP4 (VP8*) (876 bp) genes to determine the G and P genotype, respectively [29-31]. A diarrheic fecal sample (BRA1532) was selected according to the quality score for analysis of the eleven genes using previously described primers [32-36].

The RT-PCR products were purified using an Illustra GFX PCR DNA and Gel Band Purification Kit and sequenced using an ABI3500 Genetic Analyzer sequencer with a BigDye® Terminator v3.1 Cycle Sequencing Kit.

Nucleotide quality analysis and contig assembly of the RVA gene sequences were performed with Phred and CAP3 software, respectively. The identity matrix was constructed using BioEdit software version 7.0.8.0. Phylogenetic trees based on the nucleotides were obtained using the neighbor-joining method in MEGA v6 software. The bootstrapping probabilities were calculated using 1,000 replicates. The sequences described in this study are available in GenBank with the accession numbers JQ943546 to JQ943578.

Eighty percent (136/170) of the calves that were 15 to 30 days old had diarrhea. All diarrheic calves were treated with broad-spectrum antibiotics by the parenteral route but were unresponsive to antibiotic therapy, and 2.94 % (4/136) of the calves died.

Before analysis for bovine RVA, the diarrheic fecal samples were evaluated by the modified Ziehl-Neelsen technique [37] and SN-PCR assay [38] for Cryptosporidium spp. and bovine coronavirus detection, respectively, and the results were negative (data not shown).

RT-PCR assays using consensus primers of VP7 and VP4 (VP8*) genes showed that 12 (70.6 %) out of 17 diarrheic fecal samples analyzed were RVA positive. The nucleotide sequence analysis of the VP7 and VP4 (VP8*) genes of 12 Brazilian wild-type bovine RVA strains showed the highest nucleotide sequence identity (94.9 %) to G6 genotype lineage IV (OH-4 equine strain) (Fig. 1a), and 92 % with the bovine strain VMRI, which belongs to the P[5] genotype lineage VIII (Fig. 1b). The cutoff value to be considered of the same lineage of the P[5] genotype is 96 % nucleotide sequence identity [21]; therefore, we propose a new lineage, named P[5]-IX. In the phylogenetic tree, the BRA1532 strain formed a new branch separated from the other lineages of the P[5] genotype.

Fig. 1

a. Phylogenetic tree of the VP7 gene, reconstructed using the sequence of a 933-bp amplicon (nt 64–996) of genotype G6 of the BRA1532 strain, represented by a filled circle. The vaccine strains are indicated by filled triangles. The numbers adjacent to the nodes represent the percentage of bootstrap support (1,000 replicates) for the clusters. Bootstrap values less than 50 % are not shown. b. Phylogenetic tree of the VP4 (VP8*) gene, reconstructed using the sequence of a 644-bp amplicon (nt 61–704) of genotype P[5] of the BRA1532 strain, represented by a filled circle. The vaccine strain is indicated by a filled triangle. The numbers adjacent to the nodes represent the percentage of bootstrap support (1,000 replicates) for the clusters. Bootstrap values less than 50 % are not shown

For the VP6, VP1-VP3, and NSP1-NSP5/6 genes, the nucleotide sequence analysis of the RT-PCR products of the bovine RVA BRA1532 field strain revealed the presence of the following genotypes: G6(IV)-P[5](IX)-I2c-R2-C2-M2-A3-N2-T6-E2e-H3a. The bovine WC3 strain used in the human vaccine and the BRA1532 strain has the same constellation of genotypes. However, the Brazilian field strain showed high nucleotide sequence identity to the WC3 strain only in the VP1 and VP2 genes (Table 1).

Table 1

Previously described rotavirus strains with nucleotide sequence identity to the G6P[5] Brazilian bovine RVA field strain (BRA1532)

Gene	BRA1532 genotype	Percent nucleotide sequence identity
		NCDV	UK	WC3	Most similar strain
VP1	R2	93.3	91.3	95.7	WC3-Bo
VP2	C2	94.4	94.7	97.2	WC3-Bo
VP3	M2	94.8	94.3	93.3	NCDV-Bo
VP4	P[5]-IX	64.5	86.9	89.6	VMRI-Bo (92)
VP6	I2c	92.3	97.2	96.3	KJ9-1-Bo (97.8)
VP7	G6-IV	91.9	93.0	91.4	UK-Bo
NSP1	A3	94.8	96.9	93.3	UK-Bo
NSP2	N2	94.8	88.7	95.1	PTRV-Si (96.5)
NSP3	T6	92.6	80.5	92.6	RF-Bo (92.8)
NSP4	E2e	93.6	86.9	91.9	NCDV-Bo
NSP5/6	H3a	98.1	92.7	95.9	PTRV-Si (98.9)

The 70.6 % (12/17) of diarrheic fecal samples that were positive for bovine RVA in this study suggested that rotavirus was the etiological agent involved in this neonatal diarrhea outbreak. Additionally, all samples were negative for common enteropathogens such as bovine coronavirus and Cryptosporidium spp. (data not shown), and the diarrheic calves were unresponsive to two or three doses of broad-spectrum antibiotic therapy [39].

The nucleotide sequence analysis of the VP1–VP3, VP4, VP6–VP7, and NSP1-NSP5/6 genes of the BRA1532 strain showed the same genome constellation described in the bovine WC3 (G6P[5]) strain; however, the Brazilian bovine RVA strain belonged to the P[5] genotype lineage IX, and the WC3 strain belonged to the P[5] genotype lineage IV. The BRA1532 strain displayed the T6 (NSP3 gene) genotype, which was distinct from the gene found in the bovine UK strain (G6P[5]), which belongs to the T7 genotype. Additionally, the VP4 (VP8*) and NSP5/6 genes of the BRA1532 strain belonged to different lineages (P[5]-IX and H3a) when compared with the UK strain (P[5]-V and H3e) [6, 40]. According to the NCDV-Lincoln strain, the field Brazilian strain shared the same genotypes and lineages for all genes analyzed, which are present in the commercial vaccine used in this beef cattle herd, with the exception of the VP4 (VP8*) gene.

There is significant homotypic divergence within genotype G6 of RVA strains identified in human and bovine hosts, suggesting the existence of antigenic differences between G6 strains [41, 42]. Homologous immunity occurs within the same G6 genotype, and even a small genotypic difference is enough to result in insufficient protection [26].

Some studies have shown that vaccination with the prototype G6P[1] of bovine RVA results in poor heterologous protection against RVA G6 strains containing different P genotypes from the vaccine [27, 43]. Moreover, commercial vaccines containing the genotype G6P[1] may not be as effective, as P[1] may not be the most common genotype depending on the geographic region studied [44]. The vaccine pressure may generate competition, selection or variation between strains [45], which could be responsible for the introduction or emergence of novel adapted strains or differences in genotypes [23, 45].

Different genotypes have caused diarrhea in calves born to dams vaccinated against rotavirus due to poor heterologous protection against a reassortment between B641 (G6P[5]) and B223 (G10P[11]) bovine RVA strains [24]. However, it has been reported that immunity against the different G6 strains is due, at least in part, to heterologous protection against another lineage, even when the P genotype is different [41].

A polyvalent vaccine may not offer protection against all of the currently circulating bovine RVA genotypes, and multiple G and P genotypes can circulate simultaneously in a given population. Their impact on morbidity and mortality rates, high treatment costs, and reduced growth rates would not be minimized [45].

Monitoring the G and P genotypes of bovine RVA strains that are circulating or involved in outbreaks of neonatal calf diarrhea in a vaccinated cattle herd should be performed more often for a comprehensive analysis of the genetic diversity of RVA. It should also be emphasized that a new classification system would also help in understanding the inefficient protection.

In conclusion, the G6-IV and P[5]-IX genotypes were identified in the bovine RVA field strain as the cause of a diarrhea outbreak in a vaccinated beef cattle herd. These genotypes were distinct from the G and P genotypes of bovine RVA strains included in the commercial vaccine used in the herd. These results highlight the importance of homotypic immunity and monitoring of the bovine RVA genotypes circulating in cattle populations.