key: cord-0008943-46f7pb5w
authors: Ijaz, M.K.; Sabara, M.I.; Alkarmi, T.; Frenchick, P.J.; Ready, K.F.; Dar, F.K.; Babiuk, L.A.
title: Molecular determinants of rotavirus virulence: Localization of a potential virulence site in a murine rotavirus VP4
date: 2002-11-13
journal: Comp Immunol Microbiol Infect Dis
DOI: 10.1016/0147-9571(94)90035-3
sha: 96616d340ea0026e1fde9d7f840c1a9bc229a227
doc_id: 8943
cord_uid: 46f7pb5w

The molecular basis of pathogenesis in vivo for a virulent mouse rotavirus (MRV) and a less virulent bovine rotavirus (BRV) were compared under in vitro and in vivo conditions. Obvious differences in the mobility of several genomic RNA segments were observed in one-dimensional gels. Under in vitro conditions, partial proteolytic peptide mapping identified differences between the two outer capsid proteins of these virus and no difference in inner capsid protein was observed. Since it has been observed by us and others that the gene coding for VP4 protein plays a significant role in determining virulence, the variability observed in the present study between the 84 k proteins (VP4) provided a basis for further investigations in order to locate a potential virulence determinant. A comparison of the carboxypeptidase digests of the MRV- and BRV-VP4 revealed an area of variability between amino acids 307 and 407, which may represent a site of virulence determinant. Under in vivo conditions the virulence of both parenteral BRV and MRV isolates and their corresponding reassortants (with replaced gene 4) were studied in murine and bovine hosts. Like their parents, BRV and MRV isolates, reassortants obtained by replacement of gene 4 in BRV with MRV gene 4 indicated that the dose of the virus isolate used and the clinical outcome in vivo was determined by gene segment 4. The implications of these findings to elucidate the molecular basis of pathogenesis of rotaviruses are discussed.

Abstract--The molecular basis of pathogenesis in vivo for a virulent mouse rotavirus (MRV) and a less virulent bovine rotavirus (BRV) were compared under in vitro and in vivo conditions. Obvious differences in the mobility of several genomic RNA segments were observed in one-dimensional gels. Under in vitro conditions, partial proteolytic peptide mapping identified differences between the two outer capsid proteins of these virus and no difference in inner capsid protein was observed.

Since it has been observed by us and others that the gene coding for VP4 protein plays a significant role in determining virulence, the variability observed in the present study between the 84 k proteins (VP4) provided a basis for further investigations in order to locate a potential virulence determinant. A comparison of the carboxypeptidase digests of the MRV-and BRV-VP4 revealed an area of variability between amino acids 307 and 407, which may represent a site of virulence determinant. Under in vivo conditions the virulence of both parenteral BRV and MRV isolates and their corresponding reassortants (with replaced gene 4) were studied in routine and bovine hosts. Like their parents, BRV and MRV isolates, reassortants obtained by replacement ofgene 4 in BRV with MRV gene 4 indicated that the dose of the virus isolate used and the clinical outcome in vivo was determined by gene segment 4. The implications of these findings to elucidate the molecular basis of pathogenesis of rotaviruses are discussed.

Key words: Rotaviruses, reassortants, virulence determinant, pathogenesis, animal models.

R~sum~--Les bases mol~culaires de la pathog6n~se in vh~o d'un rotavirus murin virulent (MRV) et d'un rotavirus bovin moins virulent (BRV) ont 6t~ compar6es en conditions in vitro et in vivo.

Des differ6nces manifestes ont 6t6 observ6es en colloide unidimensionnelle quant ~i la mobilit+ de plusieurs segments g6nomiques d'ARN. En condition in vitro, un peptide marqueur partiellement prot6olytique a permis d'identifier des differences entre les deux prot6ines de la capside externe de ces virus mais aucune difference n'a ~t6 observ~e dans la prot6ine de la capside interne. Depuis que nous avions observ6 nous-m~mes et d'autres, que le g6ne codant pour la prot~ine VP4 jouait un r61e significatif dans la d~termination de la virulence, la variabilit~ observ6e dans la pr6sente 6tude entre les prot6ines 84 k (VP4) constituait une base de recherche pur l'avenir dans le but de localiser un d+terminant potentiel de la virulence. 

Rotaviruses are the major etiologic agents of acute gastroenteritis in a variety of mammalian neonates [1, 2] . Rotavirus is comprised of a number of structural and non-structural proteins. The major structural proteins includes VP4, VP7 and VP6. The outer capsid of rotaviruses is composed of two proteins, VP7 (38 k) and VP4 (84 k). Both of these proteins have been shown to induce neutralizing antibodies independently [1 and 12a] . Antibodies directed towards the VP7 or VP4 neutralize the virus, inhibit hemaglutination (HA), and passively protect neonatal murine and bovine against rotavirus challenge [12, lZa] .

To understand the pathogenesis of the virus, investigations have been underway in order to uncover the molecular mechanisms involved. Basically, all rotaviruses cause the destruction of columnar absorptive cells on the villus tips of the small intestine leading to the development of profuse diarrhoea. This observation helped to formulate the hypothesis that rotaviruses require trypsin for enhancement of infectivit~y [4] [5] [6] . In addition, it pointed to the fact that rotaviruses have a relatively specific cell tropism. The molecular mechanism involved in the latter function has been identified as the cleavage of VP4, at a specific site, to produce two polypeptides fragments VP5* (approximately 60 k) and VP8* (approximately 28 k). Recently it has been demonstrated that monoclonal antibody directed against VP8* can inhibit binding of rotavirus particles to cells in vitro [7] , indicating that VP8 cleavage product contains an important epitope(s) involved in virus-cell interaction. In addition VP8* portion of VP4 has been found to contain hemagglutinin [8] . We have demonstrated that a peptide corresponding to the trypsin cleavage site (amino acid 232 255) can induce heterotypic protection in vivo [9] . Furthermore, pre-treatment of cells with this peptide before infection with rotavirus prevented both the attachment of 35S-labelled virus and plaque formation and intragastric administration of the peptide protected neonatal mice from challenge with virulent rotavirus [10] .

In this study, we have attempted to identify virulence determinants in a virulent mouse rotavirus (MRV) and a less virulent bovine rotavirus (BRV) by determining the genetic basis of pathogenicity of these viruses in vivo and by comparing the genomic RNA, polypeptides and disassembly conditions in vitro.

Bovine rotavirus isolate C486 and murine rotavirus used in this study were propagated in confluent African green monkey kidney cells (MA-104) as described previously [4] . For radiolabelling the virus was adsorbed to confluent MA-104 monolayers for 1 hr after which time the virus inoculum was replaced with methionine-free MEM. After 3 hr of incubation at 37c'C, 25-50/~Ci of L-[35S]methionine (Amersham, Oakville, Ontario) per ml was added to the overlay as has been described [i 1].

Virus purification was achieved using the methods described previously [12] . Virus preparations were initially treated with 1% sodium dodecyl sulfate (SDS) and 100/tg of protease K per ml for 30 rain at 3TC. The RNA was then extracted once with 1 vol of phenol and twice with 1 vol of ether and precipitated by the addition of NaC1 to a final concentration of 0.15 M and 2vol in cold 95% ethanol. The RNA was precipitated at 

Virus proteins and RNA were examined by polyacrylamide gel electrophoresis as has been described by Laemmli [13] . Protein samples for electrophoresis were processed as has been described earlier [14] .

Partial proteolytic digestion of the virus protein was performed according to the procedure of Cleveland et al. [15] . A 5%-5 cm stacking gel and a 12.5% or 15%-11 cm resolving gel were employed for the digestion and resolution of digest. The resulting Cleveland digests were then electroblotted to nitrocellulose paper as described below and reacted with antibodies.

Proteins were transferred from polyacrylamide gels to 0.45 #m nitrocellulose paper as has been described by Towbin et al. [16] . After transfer, either the nitrocellulose strips were stained if non-radiolabeled samples were used or exposed to 3 M film (Picker International, Saskatoon, Saskatchewan, Canada) if radiolabeled samples were used. Nitrocellulose paper was also subjected to the immunoblot-ELISA.

Genomic RNA was transferred from polyacrylamide gels to gene screen (New England Nuclear, Lachine, Quebec, Canada) via electroblotting for 4 hr at 4°C in phosphate buffer at 1 mA. After transfer, the gene screen was subjected to hybridization.

Rotavirus-specific polypeptides were detected by the immunoblotting technique as described by Towbin et al. [16] . The procedure for reacting antisera with polypeptides transferred to nitrocellulose as has been described by Braun et al. [17] . The primary antibody used was either a monoclonal antibody directed against VP7 (amino acid 275-295) [1 i] or to the trypsin-cleavage site on VP4 (amino acid 232-255) [12b, 18] .

The gene screen filters carrying genomic RNA were prehybridized in 10-25 ml of 4 × SSC, 10 × Denhart, 0.1% SDS in a sealed bag at 65°C. Prehybridization mix was discarded after 1 hr and 10-25 ml of hybridization mix was added (4 × SSC, 1 × Denhart, 0.1% SDS, 50-100 pg/ml salmon sperm DNA) with 32P-labelled cDNA (about 1 x 10 6 cpm 102 M.K. |JAZ et al.

Cerenkov/filter). Hybridizations were done at 65°C until the background of radioactivity was reached in the wash. The filters were then dried and autoradiographed with intensifying screens at -70°C. The degree of hybridization of total genomic MRV-cDNA and total genomic BRV-cDNA to the genomic profile of BRV was determined by scanning the autoradiographs with a densitometer and then comparing the areas under corresponding peaks representing individual RNA segments.

Reassortant viruses were derived by co-infection of MA-104 cells with BRV and MRV isolates of rotaviruses at a multiplicity of 5.0 as previously described [19] . After 48 hr of incubation at 37°C, cultures infected with BRV and MRV were harvested by freezing and thawing. Progeny virus was titrated by plaque assay as previously described [20] . Individual progeny plaques were picked, passaged twice in MA-104 cells, and characterized by analysis of reassortant viruses by SDS-PAGE and RNA-cDNA hybridization as described above.

In case of murine model, neonates, 7-days old, were used in all experiments. Each group of five neonatal mice were inoculated intragastrally with either rotavirus isolates or their reassortants (2 x 103). The virus preparations were administered orally as described by [12] . Virus infection was evaluated clinically 24 hr postinoculation (HPI) and scored as follows: no sign of the disease (-) and maximum clinical score (+ + + + ).

In case of bovine, a gut loop technique was used. Basically, this approach involved surgical removal of the jejunum from a gnotobiotic calf under anaesthesia, and then segmentation by tying off every 8-10 cm. Virus was then inoculated into sections of the jejunum along with appropriate controls in adjacent sections. The volume of resting fluid in each section of the jejunum, after a 24-hr period, indicated the severity of the infection [21, 22] .

As shown in Table 1 , MRV was found to be more virulent than BRV. Table 2 also illustrates that even in the bovine system, the MRV (cell-culture adapted murine rotavirus isolate) is more virulent. It induces 1 1.7 ml of fluid as compared to undetectable fluid levels in segments inoculated with the BRV (cell-culture adapted bovine rotavirus isolate).

Comparison of the kinetics of replication of MRV and BRV in vitro supports the in t, ivo findings. Such that no BRV proteins are identifiable at 4 hr post-infection, however, in the Each group of five neonatal mice were inoculated intragastrally with either rotavirus isolates or their reassortants (2 x 103) generated and characterized as described in the Materials and Methods section. 24 HPI diarrhea was scored as reported earlier [12] .

No sign of the disease ( -) and maximum clinical score ( + + + + ). [21, 22] . These loops were inoculated with 5.0 ml of either MRV or BRV suspension (2 x 106) or mock-infected with MEM only. Intestinal loops were returned back into abdominal cavity and the cavity closed by sutures and was re-opened 24 HPI to assess the affect produced by the isolates as described in Materials and Methods section. No fluid was obtained from the mock-infected loops.

case of MRV several virus-specific proteins are already present at this time (data not shown).

The proteins and genomic profile of BRV and MRV were compared using PAGE, for both viruses, gene segments 1, 3, 4, 5, 10 and 11 were migrating at different rates. When protein profiles of MRV and BRV were examined we could only observe the difference is in the mobility of VP4. When the protein profile were compared following their digestion using different enzymes, both BRV and MRV digest patterns of the 1 l0 k (VP1), 92 k (VP2) and 45 k (VP6) inner proteins indicate that they are identical (data not shown). The outer shell proteins, however, appeared to be different when their digest patterns were compared (Fig. 1) . Specifically, the reduced forms of MRV-and BRV-VP7 glycoproteins exhibit a different chymotrypsin (Panel 1, lanes A and B) S. aureus V8 protease (Fig. 1 , Panel 1, lanes D and E) and papain (Fig. 1, Panel 1, lanes G and H) digest patterns. The reduced forms of VP4 also exhibit different papain and S. aureus V8 protease digest patterns (Fig. 1, Panel 2) .

As illustrated in Fig. 2 , the monoclonal antibodies directed against conserved sites in two outer capsid proteins i.e. VP7 (amino acid 275-295) and VP4 (amino acid 232-255) reacted equally well with both corresponding MRV-glycoproteins, suggesting that this site was not significantly altered. Lanes C, D, and corresponding immunoblots C' and D' illustrate that this antiserum also detects this site in the MRV-VP4. Figure 3A (lanes 1, 2) shows that 50 mM EDTA-0.01 M Tris-HC1 pH 7.4 can be used to remove most of the VP4 glycoprotein from the outer capsid of BRV. Under these conditions, most of the virus is converted to the single-shelled form. When MRV was incubated under the same conditions, very little glycoprotein was removed from the particles. However, if MRV was incubated in 50 mM EDTA-0.01 M Tris-HC1, pH 8.0, instead of pH 7.4, most of the glycoproteins were removed (Fig. 3A, lane 3-4) . CaC12 treatment of EDTA-treated BRV solubilized nucleocapsid protein at both pH levels (Fig. 3B, lane 2) . These preparations contain little or no contaminating capsid proteins. In contrast, CaC12 treatment of MRV at either pH (7.4 or 8.0) level solubilized very little nucleocapsid protein and these samples generally contained mostly glycoprotein (Fig. 3B,  lanes 1, 3-6) . Comparison of supernatants from EDTA and CaClz treatments (lanes 5 and 6) of MRV also confirms that glycoprotein is solubilized by both reagents.

Reassortants between MRV and BRV, where only gene 4 was exchanged, exhibited a pattern of virulence corresponding to the parent from which the gene 4 was derived ( Table 1 ). The digestive profiles of BRV-and MRV-VP4 (digested with carboxypeptidase A) were compared by scanning the autoradiographs and then matching the peaks which represent peptide bands (Fig. 4) . A comparison of peptide fragments indicates that there is a variation in the banding pattern between MRV and BRV peptides in the molecular weight range 43-53 k, a difference located between amino acid 307 and 407 of MRV-VP4.

VP4 was chosen as the most likely candidate protein for carrying out a virulence determinant since it was reported by Offit et al. [19] that the gene coding for this protein plays a major role in conferring virulence on the virus. In addition in vivo studies using BRV with replaced gene 4 from MRV confirmed this [12, 23] (Tables 1 and 2 ). Analysis of BRV-and MRV-VP4 by carboxypeptidase digestion revealed only one region of heterogeneity between these two viruses in the area spanning amino acids 307 and 407. However, it should be pointed out that the difference between the two isolates can not only be in the area spanning amino acid 307 and 407 and that the analysis of sequence data on gene 4 of both isolates might reveal other site(s) as well. Therefore, confirmation of the differences in potential virulence sites located on gene 4 of both isolates will have to wait till MRV-gene 4 sequence data becomes available. For now, it appears that there are at least two factors which may account for the increased virulence of MRV. The first one is the increased stability of MRV, which may be a consequence of the observed differences in the outer shell proteins of MRV when compared to BRV [33, 34, 37, 38] . The second factor is a specific difference in the primary structure of VP4 which may provide the molecular basis for the increased stability and consequently the increased virulence. Alternatively, the stability and virulence determinant may define two different strategies for conferring virulence on rotaviruses. The MRV has also been shown to be extremely stable in the environment [24] as compared to other human and animal rotavirus isolates explaining the difficulty in keeping rodent's colonies rotavirus-free and preventing nosocomial rotavirus infection [20, [25] [26] [27] [28] [29] .

Several in vitro and in vivo experiments carried out on different rotavirus isolates have shown that gene 4 is an important virulence factor in the pathogenesis of rotavirus [1, 19, 30] . In spite of these studies, the subject is still open for controversy. Recently, Broome et al. [31] have reported that VP4 as well as the other two major rotavirus proteins, VP6 and VP7 are not associated with virulence. However, they did not provide sufficient experimental evidence to demonstrate which other gene(s) product(s) are involved in virulence determination. Based on their limited experimental data they have speculated that either gene 5 and/or 2 may be involved in rotavirus virulence determination. It will be interesting to see what data they obtain in their proposed follow up studies.

In support of our findings, virulence studies carried out by Offit et al. [19] using heterologous rotavirus in homologous host (murine model) observed that virulence is determined by gene 4. Furthermore during the comparative analysis of the VP4 gene of Lanes A~C represent digestion patterns generated after exposure to 5.0/~g chymotrypsin, lanes D-F represent digestion patterns generated after exposure to 20 ~g S. aureus V8 protease, lanes G I represent digestion patterns generated after exposure to 50 #g papain. Panel 2 illustrates the digest pattern of BRV-VP4 in its reduced form (lanes A and C) and MRV-VP4 in its reduced form (lane B and D). Lanes A and B represent digestion patterns generated after exposure to 10/~g papain, lanes C and D represent digestion patterns generated after exposure to 5.0 l~g S. aureus V8 protease. different rotaviruses Nishikawa et al. [32] observed that the VP4 of standard strain of Nebraska calf diarrhoea virus (NCDV) which is virulent for cows differed in only one amino acid (position 23, Gln to Lys) from the VP4 of an NCDV mutant which was attenuated both for cows and for children which further underline the role played by gene 4 in determining rotavirus virulence. In our studies, however, both of the rotavirus isolates used, replicate very well in the murine system and virus isolated from neonatal mice can be assayed in vitro in cell culture [12] . Furthermore, our reassortants where only gene 4 was replaced, exhibited a pattern of virulence corresponding to the parents from which the gene 4 was derived (Table 1) .

Further studies are needed in order to identify more precisely the region responsible for conferring virulence. One of the approaches being used is to synthesize peptides corresponding to regions between amino acids 307-407 and then using them to raise antibodies. Antibodies produced in this manner will then be tested for their ability to reduce the virulence of a particular rotavirus. In addition, the genes coding VP4 from virulent and less virulent rotaviruses are being sequenced to more specifically localize the region(s) of variability.

Rotavirus gene structure and function

Virology

Molecular determinants of rotavirus neutralization and protection

Rotavirus isolation and cultivation in the presence of trypsin

Trypsin enhancement of rotavirus infectivity: mechanism of enhancement

Proteolytic enhancement of rotavirus infectivity: molecular mechanism

Antibodies to the trypsin cleavage peptide VPS* neutralize rotavirus by inhibiting binding of virions to target cells in culture

The VP8 fragment of VP4 is the rhesus rotavirus hemagglutinin

Heterotypic passive protection induced by synthetic peptides corresponding to VP7 and VP4 of bovine rotavirus

A synthetic peptide from rotavirus VP4 blocks virus attachment and prevents rotavirus infection

Preliminary characterization of an epitope involved in neutralization and cell attachment that is localized on the major bovine rotavirus glycoprotein

Development ofa murine model to study the pathogenesis of rotavirus infection

Monoclonal antibody to the immunodominant region of the major outer capsid glycoprotein (VP7-275-295 and VP4-232-255) of bovine rotavirus: passive protection in a neonatal murine model

Characterization of a synthetic peptide mimicking trypsin-cleavage site of rotavirus VP4

Cleavage of structural proteins during the assembly of the head of bacteriophage T 4

Genetic heterogeneity within individual bovine rotavirus isolates

Peptide mapping by limited proteolysis in SDS and analysis by gel electrophoresis

Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: Procedure and some applications

Application of denatured, electrophoretically separated, and immobilized lysates of herpes simplex virus-infected cells for the detection of monoclonal antibodies and for studies of properties of viral proteins

Primary structure of the cleavage site associated with trypsin enhancement of rotavirus Sal I infectivity

Molecular basis of rotavirus virulence: role of gene segment 4

Comparison of the airborne survival of calf rotavirus and poliovirus type l (Sabin) aerosolized as a mixture

Comparative virulence of different bovine rotavirus isolates

Monoclonal antibodies to bovine coronavirus glycoproteins E2 and E3: demonstration of in vivo virus-neutralizing activity

Assessment of intestinal damage in rotavirus infected neonatal mice by a D-xylose absorption test

Studies on the survival of aerosolized bovine rotavirus (UK) and a murine rotavirus

Nosocomial rotavirus infections in neonates: means of prevention and control

Airborne survival of rotaviruses and coronaviruses: implications for disease transmission

Effect of relative humidity, atmospheric temperature and suspending medium on the airborne survival of human rotavirus

Development of methods to study the survival of airborne viruses

Transmission of viral infections through airborne route

Binding to sialic acid is not an essential step for the entry of animal rotaviruses to epithelial cells in culture

Murine rotavirus genes encoding outer capsid proteins VP4 and VP7 are not major determinants of host range restriction and virulence

Comparative analysis of the VP3 gene of divergent strains of the rotaviruses simian and bovine Nebraska calf diarrhea virus

Purification and characterization of bovine rotavirus cores

Activators of rotavirus RNA polymerase by calcium chelation

Gene coding assignments for growth restriction, neutralization and subgroup specificities of the W and DS-1 strains of human rotavirus

Production and preliminary characterization of monoclonal antibodies directed at two surface proteins of rhesus rotavirus

In vitro assembly of bovine rotavirus nucleocapsid protein

In vitro assembly of the outer capsid of bovine rotavirus is calcium-dependent