Comparison of human, simian, and bovine rotaviruses for requirement of sialic acid in hemagglutination and cell adsorption

Biochemical characterization of rotavirus receptors in MA104 cells

We have tested the effect of metabolic inhibitors, membrane cholesterol depletion, and detergent extraction of cell surface molecules on the susceptibility of MA104 cells to infection by rotaviruses. Treatment of cells with tunicamycin, an inhibitor of protein N glycosylation, blocked the infectivity of the SA-dependent rotavirus RRV and its SA-independent variant nar3 by about 50%, while the inhibition of O glycosylation had no effect. The inhibitor of glycolipid biosynthesis d, l-threo-1-phenyl-2-decanoylamino-3-morpholino-1-propanol (PDMP) blocked the infectivity of RRV, nar3, and the human rotavirus strain Wa by about 70%. Sequestration of cholesterol from the cell membrane with beta-cyclodextrin reduced the infectivity of the three viruses by more than 90%. The involvement of N-glycoproteins, glycolipids, and cholesterol in rotavirus infection suggests that the virus receptor(s) might be forming part of lipid microdomains in the cell membrane. MA104 cells incubated with the nonionic detergent octyl-beta-glucoside (OG) showed a ca. 60% reduction in their ability to bind rotaviruses, the same degree to which they became refractory to infection, suggesting that OG extracts the potential virus receptor(s) from the cell surface. Accordingly, when preincubated with the viruses, the OG extract inhibited the virus infectivity by more than 95%. This inhibition was abolished when the extract was treated with either proteases or heat but not when it was treated with neuraminidase, indicating the protein nature of the inhibitor. Two protein fractions of around 57 and 75 kDa were isolated from the extract, and these fractions were shown to have rotavirus-blocking activity. Also, antibodies to these fractions efficiently inhibited the infectivity of the viruses in untreated as well as in neuraminidase-treated cells. Five individual protein bands of 30, 45, 57, 75, and 110 kDa, which exhibited virus-blocking activity, were finally isolated from the OG extract. These proteins are good candidates to function as rotavirus receptors.

Rotavirus infection of MA104 cells is inhibited by Ricinus lectin and separately expressed single binding domains

Various lectins were tested for blocking rotavirus infection of MA104 cells and it was observed that galactose-specific lectins were the most inhibitory. Of these Ricinus agglutinin was able to inhibit infection (by human and animal strains) at concentrations as low as 10(-9) M. In addition, in a virus overlay protein blot assay Ricinus agglutinin competed with simian rotavirus SA11 for binding to solubilized MA104 proteins. Amino acid sequence comparisons revealed similarity between the ricin toxin B subunit (which contains two separate carbohydrate-binding motifs: single binding domains (SBD) 1 and 2) and rotavirus spike protein VP4. A filamentous phage display system was used to independently express the two binding domains and while SBD1 inhibited infection of MA104 cells by CRW8, NCDV, and to a lesser extent Wa, SBD2 blocked only CRW8 and NCDV infection. Furthermore inhibition of CRW8 infection was a direct result of phage inhibiting virus attachment to cells. When amino acid 248 within SBD2 was mutated from the ricin toxin to the Ricinus agglutinin sequence this phage clone showed reduced binding to galactose and was no longer able to inhibit virus infection. Thus, rotavirus recognizes galactose as an important component of the receptor on MA104 cells.

Growth of rotaviruses in continuous human and monkey cell lines that vary in their expression of integrins

Rotavirus replication occurs in vivo in intestinal epithelial cells. Cell lines fully permissive to rotavirus include kidney epithelial (MA104), colonic (Caco-2) and hepatic (HepG2) types. Previously, it has been shown that cellular integrins alpha 2 beta 1, alpha 4 beta 1 and alpha X beta 2 are involved in rotavirus cell entry. As receptor usage is a major determinant of virus tropism, the levels of cell surface expression of these integrins have now been investigated by flow cytometry on cell lines of human (Caco-2, HepG2, RD, K562) and monkey (MA104, COS-7) origin in relation to cellular susceptibility to infection with monkey and human rotaviruses. Cells supporting any replication of human rotaviruses (RD, HepG2, Caco-2, COS-7 and MA104) expressed alpha 2 beta 1 and (when tested) alpha X beta 2, whereas the non-permissive K562 cells did not express alpha 2 beta 1, alpha 4 beta 1 or alpha X beta 2. Only RD cells expressed alpha 4 beta 1. Although SA11 grew to higher titres in RD, HepG2, Caco-2, COS-7 and MA104 cells, this virus still replicated at a low level in K562 cells. In all cell lines tested, SA11 replicated to higher titres than did human strains, consistent with the ability of SA11 to use sialic acids as alternative receptors. Levels of cell surface alpha 2 integrin correlated with levels of rotavirus growth. The alpha 2 integrin relative linear median fluorescence intensity on K562, RD, COS-7, MA104 and Caco-2 cells correlated linearly with the titre of SA11 produced in these cells at 20 h after infection at a multiplicity of 0.1, and the data best fitted a sigmoidal dose-response curve (r(2)=1.00, P=0.005). Thus, growth of rotaviruses in these cell lines correlates with their surface expression of alpha 2 beta 1 integrin and is consistent with their expression of alpha X beta 2 and alpha 4 beta 1 integrins.

Characterization of a monoclonal antibody directed to the surface of MA104 cells that blocks the infectivity of rotaviruses

Rhesus rotavirus (RRV) binds to sialic acid residues on the surface of target cells, and treatment of these cells with neuraminidase greatly reduces virus binding with the consequent reduction of infectivity. Variants that can efficiently infect neuraminidase-treated cells have been isolated, indicating that attachment to sialic acid is not an essential step for animal rotaviruses to infect cells. To identify and characterize the neuraminidase-resistant receptor for rotaviruses, we have isolated a hybridoma that secrets a monoclonal antibody (MAb) (2D9) that specifically blocks the infectivity of wild-type (wt) RRV and of its sialic acid-independent variant nar3, in untreated as well as in neuraminidase-treated cells. The infectivity of a human rotavirus was also inhibited, although to a lesser extent. MAb 2D9 blocks the binding of the variant to MA104 cells, while not affecting the binding of wt RRV; in addition, this MAb blocked the attachment of a recombinant glutathione S-transferase (GST)-VP5 fusion protein, but did not affect the binding of GST-VP8. Altogether these results suggest that MAb 2D9 is directed to the neuraminidase-resistant receptor. This receptor seems to mediate the direct attachment of the variant to the cell, through VP5, while the receptor is used by wt RRV for a secondary interaction, after its initial binding to sialic acid, through VP8. MAb 2D9 interacts specifically with the cell surface by indirect immunofluorescence, immunoelectron microscopy, and FACS. By a solid-phase immunoisolation technique, MAb 2D9 was found to react with three proteins of ca. 47, 55, and 220 kDa, which might form a complex.

Sialic acid dependence and independence of group A rotaviruses

We have found (1), in contrast to previous reports, the human rotavirus Wa strain is sialic acid-dependent for binding to and infectivity of MA-104 cells and (2), a dual carbohydrate binding specificity is associated with both human Wa and Porcine OSU rotaviruses. One carbohydrate binding activity is associated with triple-layered virus particles (TLP) and the other with double-layered virus particles (DLP). In binding and infectivity studies, we found that gangliosides were the most potent inhibitors of both the human and procine rotavirus TLP. Furthermore, glycosylation mutant cells deficient in sialylation or neuraminidase-treated MA104 cells, did not bind rotavirus TLP from either strain. Our results show that human Wa binding and infectivity cannot be distinguished from the porcine OSU strain and appears to be sialic acid-dependent. Direct binding of human or porcine TLP to a variety of intact gangliosides was demonstrated in an thin-layer chromatographic (TLC) overlay assay. Human or porcine rotavirus DLP did not bind to any of the intact gangliosides but surprisingly bound asialogangliosides. This binding was abolished by prior treatment of the glycolipids with ceramide glycanase suggesting the intact asialoglycolipid was required for DLP binding. After treatment of either human or porcine TLP with EDTA to remove the outer shell, virus particles bound only to the immobilized asialogangliosides. These results suggest that rotavirus sugar binding specificity can be interpreted either as sialic acid-dependent or independent based on whether the virus preparation consists primarily of triple-layered or double-layered particles. Of perhaps greater interest is the possibility that sialic acid-independent carbohydrate binding activity plays a role in virus maturation or assembly.

The VP5 domain of VP4 can mediate attachment of rotaviruses to cells

Some animal rotaviruses require the presence of sialic acid (SA) on the cell surface to infect the cell. We have isolated variants of rhesus rotavirus (RRV) whose infectivity no longer depends on SA. Both the SA-dependent and -independent interactions of these viruses with the cell are mediated by the virus spike protein VP4, which is cleaved by trypsin into two domains, VP5 and VP8. In this work we have compared the binding characteristics of wild-type RRV and its variant nar3 to MA104 cells. In a direct nonradioactive binding assay, both viruses bound to the cells in a saturable and specific manner. When neutralizing monoclonal antibodies directed to both the VP8 and VP5 domains of VP4 were used to block virus binding, antibodies to VP8 blocked the cell attachment of wild-type RRV but not that of the variant nar3. Conversely, an antibody to VP5 inhibited the binding of nar3 but not that of RRV. These results suggest that while RRV binds to the cell through VP8, the variant does so through the VP5 domain of VP4. This observation was further sustained by the fact that recombinant VP8 and VP5 proteins, produced in bacteria as fusion products with glutathione S-transferase, were found to bind to MA104 cells in a specific and saturable manner and, when preincubated with the cell, were capable of inhibiting the binding of wild-type and variant viruses, respectively. In addition, the VP5 and VP8 recombinant proteins inhibited the infectivity of nar3 and RRV, respectively, confirming the results obtained in the binding assays. Interestingly, when the infectivity assay was performed on neuraminidase-treated cells, the VP5 fusion protein was also found to inhibit the infectivity of RRV, suggesting that RRV could bind to the cell through two sequential steps mediated by the interaction of VP8 and VP5 with SA-containing and SA-independent cell surface receptors, respectively.

Integrins alpha2beta1 and alpha4beta1 can mediate SA11 rotavirus attachment and entry into cells

Most mammalian rotaviruses contain tripeptide amino acid sequences in outer capsid proteins VP4 and VP7 which have been shown to act as ligands for integrins alpha2beta1 and alpha4beta1. Peptides containing these sequences and monoclonal antibodies directed to these integrins block rotavirus infection of cells. Here we report that SA11 rotavirus binding to and infection of K562 cells expressing alpha2beta1 or alpha4beta1 integrins via transfection is increased over virus binding to and infection of cells transfected with alpha3 integrin or parent cells. The increased binding and growth were specifically blocked by a monoclonal antibody to the transfected integrin subunit but not by irrelevant antibodies. In our experiments, integrin activation with phorbol ester did not affect virus binding to cells. However, phorbol ester treatment of K562 parent and transfected cells induced endogenous gene expression of alpha2beta1 integrin, which was detectable by flow cytometry 16 h after treatment and quantitatively correlated with the increased level of SA11 virus growth observed after this time. Virus binding to K562 cells treated with phorbol ester 24 h previously and expressing alpha2beta1 was elevated over binding to control cells and was specifically blocked by the anti-alpha2 monoclonal antibody AK7. Virus growth in alpha4-transfected K562 cells which had also been induced to express alpha2beta1 integrin with phorbol ester occurred at a level approaching that in the permissive MA104 cell line. We therefore have demonstrated that two integrins, alpha2beta1 and alpha4beta1, are capable of acting as cellular receptors for SA11 rotavirus.

Entry of rotaviruses is a multistep process

The infection of epithelial cells by some animal rotavirus strains requires the presence of sialic acid (SA) on the cell surface. Recently, we isolated rhesus rotavirus variants, named nar, whose infectivity, like that of human rotaviruses, is not dependent on SA. In this work, we have determined the binding properties of these SA-dependent and -independent rotavirus strains to MA104 cells. The half-time of attachment of the SA-dependent porcine rotavirus YM and reassortant virus DS1xRRV was found to be about 10 times longer in neuraminidase-treated cells than in untreated cells. On the other hand, human rotaviruses Wa and DS1, and the variant nar3, bound to cells two to three times more rapidly in the absence of SA. To investigate whether the SA-independent cellular structure recognized by the variant and human rotaviruses was the same, we used an infection assay designed to detect competition for cell surface molecules at both attachment and post-attachment steps. In this assay, human rotavirus Wa efficiently competed the infectivity of YM in untreated cells and that of the variant nar3 in untreated, as well as neuraminidase-treated, cells. This competition was nonreciprocal, since YM and nar3 did not compete, but rather increased three- to fivefold the infectivity of Wa. In contrast, a two-direction competition between the variant nar3 and DS1xRRV was found. Similar results were obtained when psoralen-inactivated viruses were used as competitors, indicating that the competition observed was during the early stages of infection. Altogether, these results suggest the existence of multiple interactions between rotaviruses and the cell surface and revealed the existence of common steps during the entry of human and animal rotavirus strains.

Animal virus receptors

The term 'receptor' is generally accepted as the cell-surface component that participates in virus binding and facilitates subsequent viral infection. Recent advances in technology have permitted the identification of several virus receptors, increasing our understanding of the significance of this initial virus-cell and virus-host interaction. Virus binding was previously considered to involve simple recognition and attachment to a single cell surface molecule by virus attachment proteins. The classical concept of these as single entities that participate in a lock-and-key-type process has been superseded by new data indicating that binding can be a multistep process, often involving different virus-attachment proteins and more than one host-cell receptor.

Attachment and growth of human rotaviruses RV-3 and S12/85 in Caco-2 cells depend on VP4

Studies with human neonatal rotaviruses RV-3 and S12/85 and their reassortants showed that VP4 is a determinant of rotavirus attachment to and growth in Caco-2 cells. The binding of these viruses to MA104 and Caco-2 cells correlated with their growth ability. Virus sensitivity to trypsin and the VP4 fusion region may be implicated in these processes.

Structure and function of a ganglioside receptor for porcine rotavirus

A ganglioside fraction isolated from pooled intestines from newborn to 4-week-old piglets, which we previously partially characterized and showed to specifically inhibit the binding of porcine rotavirus (OSU strain) to host cells (M. D. Rolsma, H. B. Gelberg, and M. S. Kuhlenschmidt, J. Virol. 68:258-268, 1994), was further purified and found to contain two major monosialogangliosides. Each ganglioside was purified to apparent homogeneity, and their carbohydrate structure was examined by high-pH anion-exchange chromatography coupled with pulsed amperometric detection and fast atom bombardment mass spectroscopy. Both gangliosides possessed a sialyllactose oligosaccharide moiety characteristic of GM3 gangliosides. Compositional analyses indicated that each ganglioside was composed of sialic acid, galactose, glucose, and sphingosine in approximately a 1:1:1:1 molar ratio. Each ganglioside differed, however, in the type of sialic acid residue it contained. An N-glycolylneuraminic acid (NeuGc) moiety was found in the more polar porcine GM3, whereas the less polar GM3 species contained N-acetylneuraminic acid (NeuAc). Both NeuGcGM3 and NeuAcGM3 displayed dose-dependent inhibition of virus binding to host cells. NeuGcGM3 was approximately two to three times more effective than NeuAcGM3 in blocking virus binding. Inhibition of binding occurred with as little as 400 pmol of NeuGcGM3/50 ng of virus (approximately 2 x 10(7) virions) and 2 x 10(6) cells/ml. Fifty percent inhibition of binding was achieved with 0.64 and 1.5 microM NeuGcGM3 and NeuAcGM3, respectively. The free oligosaccharides 3'- and 6'-sialyllactose inhibited binding 50% at millimolar concentrations, which were nearly 1,000 times the concentration of intact gangliosides required for the same degree of inhibition. Direct binding of infectious, triple-layer rotavirus particles, but not noninfectious, double-layered rotavirus particles, to NeuGcGM3 and NeuAcGM3 was demonstrated by using a thin-layer chromatographic overlay assay. NeuGcGM3 and NeuAcGM3 inhibited virus infectivity of MA-104 cells by 50% at concentrations of 3.97 and 9. 84 microM, respectively. NeuGcGM3 (700 nmol/g [dry weight] of intestine) was found to be the predominant enterocyte ganglioside (comprising 75% of the total lipid-bound sialic acid) in neonatal piglets, followed by NeuAcGM3 (200 nmol/g [dry weight] of intestine). NeuGcGM3 and NeuAcGM3 together comprised nearly 100% of the lipid-bound sialic acid in the neonatal intestine, but their quantities rapidly diminished during the first 5 weeks of life. These data support the hypothesis that porcine NeuGcGM3 and NeuAcGM3 are physiologically relevant receptors for porcine rotavirus (OSU strain). Further support for this hypothesis was obtained from virus binding studies using mutant or neuraminidase-treated cell lines. Lec-2 cells, a mutant clone of CHO cells characterized by a 90% reduction in sialyllation of its glycoconjugates, bound less than 5% of the virus compared to control cell binding. In contrast, Lec-1 cells, a mutant CHO clone characterized by a deficiency in glycosylation of N-linked oligosaccharides, still bound rotavirus. Furthermore, exogenous addition of NeuGcGM3 to the Lec-2 mutant cells restored their ability to bind rotavirus in amounts equivalent to that of their parent (CHO) cell line. In the virus-permissive MA-104 cell line, NeuGcGM3 was also able to partially restore rotavirus infectivity in neuraminidase-treated cells. These data suggest that gangliosides play a major role in recognition of host cells by porcine rotavirus (OSU strain).

Rotaviruses induce an early membrane permeabilization of MA104 cells and do not require a low intracellular Ca2+ concentration to initiate their replication cycle

In this work, we found that rotavirus infection induces an early membrane permeabilization of MA104 cells and promotes the coentry of toxins, such as alpha-sarcin, into the cell. This cell permeability was shown to depend on infectious virus and was also shown to be virus dose dependent, with 10 infectious particles per cell being sufficient to achieve maximum permeability; transient, lasting no more than 15 min after virus entry and probably occurring concomitantly with virus penetration; and specific, since cells that are poorly permissive for rotavirus were not permeabilized. The rotavirus-mediated coentry of toxins was not blocked by the endocytosis inhibitors dansylcadaverine and cytochalasin D or by the vacuolar proton-ATPase inhibitor bafilomycin A1, suggesting that neither endocytocis nor an intraendosomal acidic pH or a proton gradient is required for permeabilization of the cells. Compounds that raise the intracellular concentration of calcium ([Ca2+]i) by different mechanisms, such as the calcium ionophores A23187 and ionomycin and the endoplasmic reticulum calcium-ATPase inhibitor thapsigargin, did not block the coentry of alpha-sarcin or affect the onset of viral protein synthesis, suggesting that a low [Ca2+]i is not essential for the initial steps of the virus life cycle. Since the entry of alpha-sarcin correlates with virus penetration in all parameters tested, the assay for permeabilization to toxins might be a useful tool for studying and characterizing the route of entry and the mechanism used by rotaviruses to traverse the cell membrane and initiate a productive replication cycle.

Productive penetration of rotavirus in cultured cells induces coentry of the translation inhibitor alpha-sarcin

Internalization of rotavirus in MA104 cells was found to induce coentry of alpha-sarcin, a toxin that inhibits translation in cell-free systems and to which cells are normally impermeable. Entry of the toxin, measured by inhibition of protein synthesis at early times after infection, correlated with virus penetration leading to expression of infectivity, since toxin entry (1) was induced only by trypsin-treated triple-layered virions, to a degree dependent on the toxin and the virus concentration; (2) correlated with the degree of permissivity of different cell lines to rotavirus infection; (3) was inhibited to a similar extent as infectivity by treatment of cells with neuraminidase; and (4) was inhibited by pre- or postadsorption incubation of the virus with neutralizing monoclonal antibodies to VP7 and VP4 (VP8*). Neither the virus infectivity nor the toxin coentry was significantly affected by treatment of cells with bafilomycin A1, an inhibitor of the vacuolar proton ATPase, indicating that both events are independent of the endosomal acid pH. Virus-like particles (VLP), composed of rotavirus proteins 2/6/7/4, but not 2/6/7 or 2/6, were able to induce toxin entry as efficiently as virions. Use of genetically modified VLP in combination with the toxin coentry assay, which measures entry through a productive pathway, should allow identification of the regions of the outer capsid proteins essential for rotavirus penetration.

Synthesis of a novel sialic acid derivative (sialylphospholipid) as an antirotaviral agent

A novel sialylphospholipid (SPL) was synthesized from N-acetylneuraminic acid (NeuAc) and phosphatidylcholine (PC) by a chemical and enzymatic method and evaluated as an inhibitor of rotavirus. PC and 1,8-octanediol were conjugated by transesterification reaction of Streptomyces phospholipase D (PLD) under a water-chloroform biphasic system to afford phosphatidyloctanol, which was condensed with a protected 2-chloro-2-deoxyneuraminic acid derivative by using silver trifluoromethanesulfonate as an activator in chloroform and converted, after deprotection, to SPL. Rhesus monkey kidney cells (MA-104) were incubated with simian (SA-11 strain) and human (MO strain) rotaviruses in the presence of SPL, and the cells infected were detected indirectly with anti-rotavirus antibody. SPL showed dose dependent inhibition against both virus strains. The concentrations required for 50% inhibition (IC50) against SA-11 and MO were 4.35 and 16.1 microM, respectively, corresponding to 10(3)- and 10(4)-fold increases in inhibition as compared to monomeric NeuAc.

MUC1 and other sialoglycoconjugates inhibit adenovirus-mediated gene transfer to epithelial cells

Recombinant adenoviruses are currently being evaluated as gene transfer vectors for the treatment of airway diseases. Recent evidence indicates that gene transfer to differentiated airway epithelial cells is inefficient. We hypothesized that apical membrane glycoconjugates, such as the transmembrane mucin MUC1, reduce the efficiency of adenovirus-mediated gene transfer. To address this, studies were performed in primary bronchial epithelial and Madin Darby canine kidney (MDCK) cells transduced to express human MUC1. Colocalization of MUC1 and an adenoviral lacZ transgene in the bronchial epithelial cells revealed that at several multiplicities of infection, the percentage of cells expressing lacZ was five-fold less in MUC1-expressing cells. Moreover, lacZ expression was three- to eight-fold lower in MUC1-expressing than in control MDCK cells, demonstrating that MUC1 interferes with gene transfer and is not merely a phenotypic marker of a cell that is refractory to adenovirus infection. Neuraminidase pretreatment of cells to remove sialic acid residues prior to viral adsorption increased the efficiency of gene transfer two- to five-fold in human airway and MDCK cells, and in a xenograft model of human airway. This effect was also observed in cultured cells that do not express MUC1, suggesting that other sialylated glycoconjugates impact on the efficiency of gene transfer. An inhibitory effect of negatively charged glycoconjugates on adenovirus binding was further supported by the finding that adsorption of adenovirus with a polycation significantly increased gene transfer efficiency. These data demonstrate for the first time that sialoglycoconjugates on epithelial cells reduce the efficiency of adenovirus-mediated gene transfer.

Functional and structural analysis of the sialic acid-binding domain of rotaviruses

The infectivity of most animal rotaviruses is dependent on the interaction of the virus spike protein VP4 with a sialic acid (SA)-containing cell receptor, and the SA-binding domain of this protein has been mapped between amino acids 93 and 208 of its trypsin cleavage fragment VP8. To identify which residues in this region are essential for the SA-binding activity, we performed alanine mutagenesis of the rotavirus RRV VP8 expressed in bacteria as a fusion polypeptide with glutathione S-transferase. Tyrosines were primarily targeted since tyrosine has been involved in the interaction of other viral hemagglutinins with SA. Of the 15 substitutions carried out, 10 abolished the SA-dependent hemagglutination activity of the protein, as well as its ability to bind to glycophorin A in a solid-phase assay. However, only alanine substitutions for tyrosines 155 and 188 and for serine 190 did not affect the overall conformation of the protein, as judged by their interaction with a panel of conformationally sensitive neutralizing VP8 monoclonal antibodies (MAbs). These findings suggest that these three amino acids play an essential role in the SA-binding activity of the protein, presumably by interacting directly with the SA molecule. The predicted secondary structure of VP8 suggests that it is organized as 11 beta-strands separated by loops; in this model, Tyr-155 maps to loop 7 while Tyr-188 and Ser-190 map to loop 9. The close proximity of these two loops is also supported by previous results from competition experiments with neutralizing MAbs directed at RRV VP8.

Virus-like particle-induced fusion from without in tissue culture cells: role of outer-layer proteins VP4 and VP7

We recently described an assay that measures fusion from without induced in tissue culture cells by rotavirus, a nonenveloped, triple-protein-layered member of the Reoviridae family (M. M. Falconer, J. M. Gilbert, A. M. Roper, H. B. Greenberg, and J. S. Gavora, J. Virol. 69:5582-5591, 1995). The conditions required for syncytium formation are similar to those for viral penetration of the plasma membrane during the course of viral infection of host cells, as the presence of the outer-layer proteins VP4 and VP7 and the cleavage of VP4 are required. Here we present evidence that virus-like particles (VLPs) produced in Spodoptera frugiperda Sf-9 cells from recombinant baculoviruses expressing the four structural proteins of rotavirus can induce cell-cell fusion to the same extent as native rotavirus. This VLP-mediated fusion activity was dependent on trypsinization of VP4, and the strain-specific phenotype of individual VP4 molecules was retained in the syncytium assay similar to what has been seen with reassortant rotaviruses. We show that intact rotavirus and VLPs induce syncytia with cells that are permissive to rotavirus infection whereas nonpermissive cells are refractory to syncytium formation. This finding further supports our hypothesis that the syncytium assay accurately reflects very early events involved in viral infection and specifically the events related to viral entry into the cell. Our results also demonstrate that neither viral replication nor rotavirus proteins other than VP2, VP6, VP4, and VP7 are required for fusion and that both VP4 and VP7 are essential. The combination of a cell-cell fusion assay and the availability of recombinant VLPs will permit us to dissect the mechanisms of rotavirus penetration into host cells.

Synthesis and biological evaluation of N-acetylneuraminic acid-based rotavirus inhibitors

Rotavirus can cause severe gastrointestinal disease, especially in infants and young children, and is particularly prevalent in Third-World countries. Therefore, the development of potential inhibitors of this virus is of great interest. The present study describes the synthesis and in vitro biological evaluation of a number of N-acetylneuraminic acid-based compounds as potential rotavirus inhibitors. Our data suggests that it is indeed possible to inhibit adhesion of the virus, and hence in vitro replication, with carbohydrate-based molecules, although this inhibition does appear to be strain dependent.

Isolation of a human rotavirus containing a bovine rotavirus VP4 gene that suppresses replication of other rotaviruses in coinfected cells

Bovine-human reassortant strains containing ten human rotavirus gene segments and segment 4, encoding VP4, of a bovine rotavirus were isolated from the stool of an infected Bangladeshi infant during cell culture adaptation. Two plaque purified variants of this reassortant, one making very large (429-L4) and the other tiny (429-S4) plaques, were further analyzed. The electropherotypes of these variants were identical except for slight mobility differences in segment 4. The predicted sequence of amino acids (aa) 16-280 in VP4 proteins revealed four differences between variants even in this limited region, so no single difference could be linked to plaque size. The small plaque variant S4 was phenotypically unstable and mutated to a large plaque-former within a single cell culture passage. The predicted sequence of aa 16-280 of a large plaque variant derived from S4 revealed six changes, only one of which was common to that of the L4 strain, thus suggesting that multiple amino acid changes in VP4 may affect plaque size. Although the large plaque variant L4 grew faster and was released from cells more rapidly than S4, its replication and that of other rotaviruses tested (i.e. RRV, NCDV and Wa) was suppressed by S4 in coinfected cells. Using an RRV x S4 reassortant containing only RRV segment 4, it was established that suppression was linked to the S4 VP4 protein. This suppression could not be associated with inhibition of viral adsorption and, therefore, appeared to occur following internalization. Thus, a new property of the rotavirus VP4 protein has been identified in a bovine-human rotavirus reas-sortant.

Interactions between the two surface proteins of rotavirus may alter the receptor-binding specificity of the virus

The infection of target cells by most animal rotavirus strains requires the presence of sialic acids (SAs) on the cell surface. We recently isolated variants from simian rotavirus RRV whose infectivity is no longer dependent on SAs and showed that the mutant phenotype segregates with the gene coding for VP4, one of the two surface proteins of rotaviruses (the other one being VP7). The nucleotide sequence of the VP4 gene of four independently isolated variants showed three amino acid changes, at positions 37 (Leu to Pro), 187 (Lys to Arg), and 267 (Tyr to Cys), in all mutant VP4 proteins compared with RRV VP4. The characterization of revertant viruses from two independent mutants showed that the arginine residue at position 187 changed back to lysine, indicating that this amino acid is involved in the determination of the mutant phenotype. Surprisingly, sequence analysis of reassortant virus DS1XRRV, which depends on SAs to infect the cell, showed that its VP4 gene is identical to the VP4 gene of the variants. Since the only difference between DS1XRRV and the RRV variants is the parental origin of the VP7 gene (human rotavirus DS1 in the reassortant), these findings suggest that the receptor-binding specificity of rotaviruses, via VP4, may be influenced by the associated VP7 protein.

Nucleotide sequence and expression in E. coli of the complete P4 type VP4 from a G2 serotype human rotavirus

The complete sequence of a P4 type VP4 gene from a G2 serotype human rotavirus, IS2, isolated in India has been determined. Although the IS2 VP4 is highly homologous to the other P4 type alleles, it contained acidic amino acid substitutions at several positions that make it acidic among the P4 type alleles that are basic. Moreover, comparative sequence analysis revealed unusual polymorphism in members of the P4 type at amino acid position 393 which is highly conserved in members of other VP4 types. To date, expression of complete VP4 in E. colic has not been achieved. In this study we present successful expression in E. coli of the complete VP4 as well as VP8* and VP5* cleavage subunits in soluble form as fusion proteins of the maltose-binding protein (MBP) and their purification by single-step affinity chromatography. The hemagglutinating activity exhibited by the recombinant protein was specifically inhibited by the antiserum raised against it. Availability of pure VP4 proteins should facilitate development of polyclonal and monoclonal antibodies (MAbs) for P serotyping of rotaviruses.

Genetic mapping indicates that VP4 is the rotavirus cell attachment protein in vitro and in vivo

To identify the rotavirus protein which mediates attachment to cells in culture, viral reassortants between the simian rotavirus strain RRV and the murine strains EHP and EW or between the simian strain SA-11 and the human strain DS-1 were isolated. These parental strains differ in the requirement for sialic acid to bind and infect cells in culture. Infectivity and binding assays with the parental and reassortant rotaviruses indicate that gene 4 encodes the rotavirus protein which mediates attachment to cells in culture for both sialic acid-dependent and -independent strains. Using ligated intestinal segments of newborn mice and reassortants obtained between the murine strain EW and RRV, we developed an in vivo infectivity assay. In this system, the infectivity of EW was not affected by prior treatment of the enterocytes with neuraminidase, while neuraminidase treatment reduced the infectivity of a reassortant carrying gene 4 from RRV on an EW background more than 80% relative to the controls. Thus, VP4 appears to function as the cell attachment protein in vivo as well as in vitro.

Characterization of SA-11 rotavirus receptorial structures on human colon carcinoma cell line HT-29

The involvement of different cell membrane components in the receptor structures for SA-11 rotavirus was investigated. As experimental model, the human enterocyte-like HT-29 cell line, was used because of its closer resemblance to the in vivo viral cellular target as compared to other in vitro systems. Rotavirus was incubated with whole membranes or their separated protein and lipid fractions before infection. Either isolated cell membranes or lipid components were capable of binding to the virus and to prevent infection, whereas proteins did not show any inhibitory activity. Among lipids, the glycolipid fraction was shown to impede rotaviral antigen synthesis with a dose-dependent relationship, whereas phospholipids failed to prevent viral infection. To confirm these findings, membranes and target cells were subjected to different enzymatic treatments prior to infection. In addition, HT-29 cells were also incubated with different lectins before infection. The blocking activity of membranes was inhibited by treatment with ceramide glycanase, neuraminidase, and beta-galactosidase but not by treatment with proteases or heat (100 degrees C). Viral infection was prevented by preincubation of target cells with lectins specific for sialic acid and galactose or with ceramide glycanase, neuraminidase, and beta-galactosidase, whereas protease treatments were not active. The results of these experimental procedures indicate that glycolipids containing specific carbohydrate moieties, such as sialic acid and galactose, contribute to the SA-11 rotavirus receptor structure on HT-29 cells.

Mapping the hemagglutination domain of rotaviruses

Most strains of animal rotaviruses are able to agglutinate erythrocytes, and the surface protein VP4 is the virus hemagglutinin. To map the hemagglutination domain on VP4 while preserving the conformation of the protein, we constructed full-length chimeras between the VP4 genes of hemagglutinating (YM) and nonhemagglutinating (KU) rotavirus strains. The parental and chimeric genes were expressed in insect cells, and the recombinant VP4 proteins were evaluated for their capacity to agglutinate human type O erythrocytes. Three chimeric genes, encoding amino acids 1 to 208 (QKU), 93 to 208 (QC), and 93 to 776 (QYM) of the YM VP4 protein in a KU VP4 background, were constructed. YM VP4 and chimeras QKU and QC were shown to specifically hemagglutinate, indicating that the region between amino acids 93 and 208 of YM VP4 is sufficient to determine the hemagglutination activity of the protein.

Identification of two independent neutralization domains on the VP4 trypsin cleavage products VP5* and VP8* of human rotavirus ST3

The antigenic structure of the VP4 protein of human rotavirus (HRV) strains Wa and ST3 was studied by using a panel of Wa- and ST3-derived VP4-specific neutralizing monoclonal antibodies (NMAbs) and NMAb-resistant variants. The VP4-coding genes from three Wa and three ST3 variants were sequenced. For Wa VP4, one homotypic and one heterotypic neutralization site, at amino acids 458 and 392, respectively, were identified. For ST3 VP4, three neutralization sites were found at amino acids 72, 217, and 385 that are either homotypic or associated with limited cross-reactivity. Cross-neutralization assays using several pairs of NMAbs and resistant variants showed that Wa VP4 has at least one large neutralization domain on its larger trypsin cleavage product, VP5*, consisting of several operationally related epitopes. VP4 of ST3 has at least two neutralization domains, one located on VP5* that is operationally related to the large neutralization domains on VP5* from HRVs Wa and KU, as well as an independent neutralization domain on VP8*, the smaller trypsin cleavage product of VP4.

Rotavirus interaction with isolated membrane vesicles

To gain information about the mechanism of epithelial cell infection by rotavirus, we studied the interaction of bovine rotavirus, RF strain, with isolated membrane vesicles from apical membrane of pig enterocytes. Vesicles were charged with high (quenching) concentrations of either carboxyfluorescein or calcein, and the rate of fluorophore release (dequenching) was monitored as a function of time after mixing with purified virus particles. Purified single-shelled particles and untrypsinized double-shelled ones had no effect. Trypsinized double-shelled virions induced carboxyfluorescein release according to sigmoid curves whose lag period and amplitude were a function of virus concentration and depended on both temperature and pH. The presence of 100 mM salts (Tris Cl, NaCl, or KCl) was required, since there was no reaction in isoosmotic salt-free sorbitol media. Other membrane vesicle preparations such as apical membranes of piglet enterocyte and rat placenta syncytiotrophoblasts, basolateral membranes of pig enterocytes, and the undifferentiated plasma membrane of cultured MA104 cells all gave qualitatively similar responses. Inhibition by a specific monoclonal antibody suggests that the active species causing carboxyfluorescein release is VP5*. Ca2+ (1 mM), but not Mg2+, inhibited the reaction. In situ solubilization of the outer capsid of trypsinized double-shelled particles changed release kinetics from sigmoidal to hyperbolic and was not inhibited by Ca2+. Our results indicate that membrane destabilization caused by trypsinized outer capsid proteins of rotavirus leads to fluorophore release. From the data presented here, a hypothetical model of the interaction of the various states of the viral particles with the membrane lipid phase is proposed. Membrane permeabilization induced by rotavirus may be related to the mechanism of entry of the virus into the host cell.

Assay for evaluation of rotavirus-cell interactions: identification of an enterocyte ganglioside fraction that mediates group A porcine rotavirus recognition

A virus-host cell-binding assay was developed and used to investigate specific binding between group A porcine rotavirus and MA-104 cells or porcine enterocytes. A variety of glycoconjugates and cellular components were screened for their ability to block rotavirus binding to cells. During these experiments a crude ganglioside mixture was observed to specifically block rotavirus binding. On the basis of these results, enterocytes were harvested from susceptible piglets and a polar lipid fraction was isolated by solvent extraction and partitioning. Throughout subsequent purification of this fraction by Sephadex partition, ion-exchange, silicic acid, and thin-layer chromatography, blocking activity behaved as a monosialoganglioside (GMX) that displayed a thin-layer chromatographic mobility between those of GM2 and GM3. The blocking activity of GMX was inhibited by treatment with neuraminidase and ceramide glycanase but not by treatment with protease or heat (100 degrees C). Further purification of GMX by high-pressure liquid chromatography resulted in the resolution of two monosialogangliosides, GMX and a band which comigrated with GM1 on thin-layer chromatography. These data suggest that a cell surface monosialoganglioside or family of monosialogangliosides may function as an in vivo relevant receptor for group A porcine rotavirus and that sialic acid is a required epitope for virus-binding activity.

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

The infection of target cells by animal rotaviruses requires the presence of sialic acids on the cell surface. Treatment of the cells with neuraminidases or incubation of the viruses with some sialoglycoproteins, such as glycophorin A, greatly reduces virus binding, with the consequent reduction of viral infectivity. In this work, we report the isolation of animal rotavirus variants whose infectivity is no longer dependent on the presence of sialic acids on the cell surface. In addition, although these variants bind to glycophorin A as efficiently as the wild-type virus, this interaction no longer inhibit viral infectivity. These observations indicate that the initial interaction of the mutants with the cell occurs at a site different from the sialic acid-binding site located on VP8, the smaller trypsin cleavage product of VP4. Reassortant analysis showed that the mutant phenotype segregates with the VP4 gene. Neutralizing monoclonal antibodies directed to VP4 and VP7 were tested for their ability to neutralize the variants. Antibodies to VP7 and VP5, the larger trypsin cleavage product of VP4, neutralized the mutants as efficiently as the wild-type virus. In contrast, although antibodies to VP8 were able to bind to the mutants, they showed little or no neutralizing activity. The implications of these findings in rotavirus attachment to and penetration of epithelial cells in culture are discussed.

Rotavirus-inhibitory activity in serial milk samples from Mexican women and rotavirus infections in their children during their first year of life

A total of 75 children born in rural Mexico were followed for diarrheal diseases and rotavirus (RV) excretion during the first year of life. For 18 children, an average of 14 serial breast milk samples were obtained between days 2 and 360 after delivery and were tested for RV-inhibitory activity. Of these samples, 70, 62, and 85% showed inhibitory activity against serotype (ST) 1 human RV, ST4 human RV, and ST3 simian RV, respectively; the median titers were 10, 10, and 20, respectively. Some 89% of the milk samples showed RV-specific antibodies in an enzyme-linked immunosorbent assay (median titer, 20). Surprisingly, 98% of the milk samples inhibited ST6 bovine RV. ST6, but not ST1, RV-inhibitory activity survived heat treatment (10 min at 80 degrees C). Of the 18 children tested, 13 children experienced 23 episodes of diarrhea (enterotoxigenic [n = 8] and enteropathogenic [n = 3] Escherichia coli, Campylobacter jejuni [n = 4], Shigella flexneri [n = 2], RV [n = 1]) and 5 children experienced 6 RV infections. Only one RV infection was associated with diarrhea. The five RV excretors did not differ from the nonexcretors with respect to the RV-inhibitory activity in the breast milk fed to them. The RV-inhibitory titers were too low in the majority of the studied Mexican milk samples to indicate an important effect of breast-feeding on the take rate of oral human, simian, or reassortant RV vaccines. Breast-feeding might, however, inhibit the take rate of a bovine RV vaccine.

Bidirectional entry of poliovirus into polarized epithelial cells

The interactions of viruses with polarized epithelial cells are of some significance to the pathogenesis of disease because these cell types comprise the primary barrier to many virus infections and also serve as the sites for virus release from the host. Poliovirus-epithelial cell interactions are of particular interest since this virus is an important enteric pathogen and the host cell receptor has been identified. In this study, poliovirus was observed to adsorb to both the apical and basolateral surfaces of polarized monkey kidney (Vero C1008) and human intestinal (Caco-2) epithelial cells but exhibited preferential binding to the basolateral surfaces of both cell types. Localization of the poliovirus receptor by a receptor-specific monoclonal antibody (D171) revealed a similar distribution predominantly on basolateral membranes, and treatment of cells with antibody D171 inhibited virus adsorption to both membrane surfaces. Poliovirus was able to initiate infection with similar efficiency following adsorption to either surface, and infection was blocked at both surfaces by D171, indicating that functional receptor molecules are expressed on both surfaces at sufficient density to mediate efficient infection at the apical and basolateral plasma membranes. Poliovirus infection resulted in a decrease in transepithelial resistance which was inhibited by prior treatment with monoclonal antibody D171 and occurred prior to other visible cytopathic effects. These results have interesting implications for viral pathogenesis in the human gut.

Human milk mucin inhibits rotavirus replication and prevents experimental gastroenteritis

Acute gastrointestinal infections due to rotaviruses and other enteric pathogens are major causes of morbidity and mortality in infants and young children throughout the world. Breast-feeding can reduce the rate of serious gastroenteritis in infants; however, the degrees of protection offered against rotavirus infection vary in different populations. The mechanisms associated with milk-mediated protection against viral gastroenteritis have not been fully elucidated. We have isolated a macromolecular component of human milk that inhibits the replication of rotaviruses in tissue culture and prevents the development of gastroenteritis in an animal model system. Purification of the component indicates that the antiviral activity is associated with an acidic fraction (pI = 4.0-4.6), which is free of detectable immunoglobulins. Furthermore, high levels of antiviral activity are associated with an affinity-purified complex of human milk mucin. Deglycosylation of the mucin complex results in the loss of antiviral activity. Further purification indicated that rotavirus specifically binds to the milk mucin complex as well as to the 46-kD glycoprotein component of the complex. Binding to the 46-kD component was substantially reduced after chemical hydrolysis of sialic acid. We have documented that human milk mucin can bind to rotavirus and inhibit viral replication in vitro and in vivo. Variations in milk mucin glycoproteins may be associated with different levels of protection against infection with gastrointestinal pathogens.

Cell surface ligands for rotavirus: mouse intestinal glycolipids and synthetic carbohydrate analogs

Rotaviral binding to receptors on epithelial cells in the small intestine is thought to be a key event in the infection process and may be carbohydrate-mediated. Strain SA11 of rotavirus bound in vitro both to glycolipids isolated from mouse small intestine and to authentic glycolipids using thin layer chromatography overlay and microtiter well adsorption assays. Neutral mouse intestinal glycolipids which bound rotavirus were GA1 (Gal beta 1----3GalNAc beta 1---4Glc beta 1----4Glc beta 1----1-ceramide) and pentaosylceramides with terminal N-acetylgalactosamine, while acidic lipids which bound rotavirus included cholesterol 3-sulfate and two compounds termed bands 80 and 81. Digestion with ceramide glycanase suggested that bands 80 and 81 have lactosyl ceramide cores and an unidentified acidic moiety(s). No sialic-acid-containing glycolipids tested were active in viral binding. Band 81, which may have a ganglio core, bound rotavirus with greatest avidity, followed by GA1. Of authentic glycolipids assayed, only GA1 and GA2 (GalNAc beta 1----4Gal beta 1----4Glc beta 1----1-ceramide) displayed rotaviral binding. A phosphatidylethanolamide dipalmitoyl-containing neoglycolipid analog of GA2 bound rotavirus with avidity similar to native GA2. Substitution of beta 1----4-linked GlcNAc or beta 1----3-linked GalNAc for terminal GalNAc of GA2 neoglycolipid supported rotaviral binding, while other substitutions abrogated it. These findings suggest that a carbohydrate epitope similar to that of GA2 is sufficient for in vitro rotaviral binding, although binding may be enhanced by galactose and/or an acidic moiety in a secondary epitope.

Group C rotavirus requires sialic acid for erythrocyte and cell receptor binding

Current immunological and biochemical information regarding the hemagglutinin and virus-cell interactions of rotavirus is obtained exclusively from studies with group A rotaviruses. In this study, I report that the immunologically and genetically distinct group C rotavirus also possesses a hemagglutinin. The viral hemagglutinin was identified on a cultivable porcine group C rotavirus strain (strain AmC-1) by using agglutinated human and guinea pig erythrocytes. Neuraminidase treatment of fresh human erythrocytes or blocking with glycophorin A or fetuin prevented hemagglutination. Infection of swine testicular cells with group C AmC-1 virus was also prevented by glycophorin A, fetuin, and neuraminidase treatment, suggesting that sialic acid constitutes an essential part of the cell receptor.

Hemagglutination by a human rotavirus isolate as evidence for transmission of animal rotaviruses to humans

Human rotavirus strain Ro1845, which was isolated in 1985 from an Israeli child with diarrhea, has a hemagglutinin that is capable of agglutinating erythrocytes from guinea pigs, sheep, chickens, and humans (group O). Hemagglutination was inhibited after incubation with hyperimmune sera or in the presence of glycophorin, the erythrocyte receptor for animal rotaviruses. These results suggest that Ro1845 is an animal rotavirus that infected a human child.

Symmetric infection of rotavirus on polarized human intestinal epithelial (Caco-2) cells

When rotavirus infects the mature villus tip cells of the small intestine, it encounters a highly polarized epithelium. In order to understand this virus-cell interaction more completely, we utilized a cell culture-adapted rhesus rotavirus (RRV) to infect human intestinal (Caco-2) and Madin-Darby canine kidney (MDCK-1) polarized epithelial cells grown on a permeable support. Filter-grown Caco-2 cells and MDCK-1 cells, producing a transepithelial resistance of 300 to 500 and greater than 1,000 omega . cm2, respectively, were infected from either the apical or basolateral domain with RRV or Semliki Forest virus. Whereas Semliki Forest virus infection only occurred when input virions had access to the basolateral domain of MDCK-1 or Caco-2 cells, RRV infected MDCK-1 and Caco-2 monolayers in a symmetric manner. The effect of rotavirus infection on monolayer permeability was analyzed by measuring the transepithelial electrical resistance. Rotavirus infection on filter-grown Caco-2 cells caused a transmembrane leak at 18 h postinfection, before the development of the cytopathic effect (CPE) and extensive virus release. Electrical resistance was completely abolished between 24 and 36 h postinfection. Although no CPE could be detected on RRV-infected MDCK cells, the infection caused a transmembrane leak that totally abolished the electrical resistance at 18 to 24 h postinfection. Cell viability and the CPE analysis together with immunohistochemistry and immunofluorescence data indicated that the abolishment of resistance across the monolayer was due not to an effect on the plasma membrane of the cells but to an effect on the paracellular pathway limited by tight junctions. Attachment and penetration of rotavirus onto Caco-2 cells caused no measurable transmembrane leak during the first hour of infection.

Identification and partial characterization of a rhesus rotavirus binding glycoprotein on murine enterocytes

In order to assess the possibility that rotavirus binds to a specific cellular receptor on enterocytes, we have used a viral overlay protein blot assay to study viral binding to murine intestinal brush border membranes (BBM). Infectious double-shelled particles of rhesus rotavirus bound specifically to two approximately 300- and 330-kDa glycoproteins from BBM prepared from suckling mice. Significantly less rotavirus binding was observed when adult BBM were examined. Rats have never been shown to harbor natural group A rotavirus infection and correspondingly, rat BBM showed no rotavirus binding activity. In suckling mice, rotavirus was found to bind to villus tip membranes to a much greater extent than to crypt preparations. Rotavirus binding activity was abolished by treatment of membrane preparations with protease. Analysis by glycolytic digestion of BBM with N- and O-glyconases revealed evidence for both N- and O-linked glycosylation of the rotavirus binding protein. Also neuraminidase digestion showed that O-linked sialic acid residues were required for virus binding. Monoclonal antibodies which immunoprecipitate the 300-kDa viral binding glycoprotein react with the apical surface of suckling but not adult enterocytes by Western blot. Baculovirus-expressed vp4, the rotavirus outer capsid spike protein, bound to the 300- and 330-kDa proteins and competed with rotavirus particles for binding sites. The ability of rotavirus to bind via vp4 to large BBM glycoproteins correlates with in vivo rotavirus cell tropism and host range restriction. Specific host cell receptor expression may be important in rotavirus pathogenesis.

Rotaviruses specifically bind to the neutral glycosphingolipid asialo-GM1

Rotaviruses are the major etiologic agents of severe diarrhea in children. Many rotaviruses encode a hemagglutinin which binds to sialic acids. We report that rotaviruses specifically recognize the neutral glycosphingolipid gangliotetraosylceramide (asialo-GM1 or GA1). GA1 resolved by thin-layer chromatography is bound by rotavirus, and binding is blocked by neutralizing rotavirus antiserum. Similar glycosphingolipid structures, such as globoside, gangliotriaosylceramide, and GA1 oxidized with galactose oxidase are ineffective in binding rotavirus. Other enteric viruses also specifically bind GA1. GA1 adsorbed to polystyrene beads inhibits rotavirus replication in vitro (as do anti-GA1 antibodies). The use of orally administered immobilized GA1 or anti-GA1 antibodies may prove useful in preventing or attenuating rotaviral and other enteric viral infections.

Rotavirus gene structure and function

Knowledge of the structure and function of the genes and proteins of the rotaviruses has expanded rapidly. Information obtained in the last 5 years has revealed unexpected and unique molecular properties of rotavirus proteins of general interest to virologists, biochemists, and cell biologists. Rotaviruses share some features of replication with reoviruses, yet antigenic and molecular properties of the outer capsid proteins, VP4 (a protein whose cleavage is required for infectivity, possibly by mediating fusion with the cell membrane) and VP7 (a glycoprotein), show more similarities with those of other viruses such as the orthomyxoviruses, paramyxoviruses, and alphaviruses. Rotavirus morphogenesis is a unique process, during which immature subviral particles bud through the membrane of the endoplasmic reticulum (ER). During this process, transiently enveloped particles form, the outer capsid proteins are assembled onto particles, and mature particles accumulate in the lumen of the ER. Two ER-specific viral glycoproteins are involved in virus maturation, and these glycoproteins have been shown to be useful models for studying protein targeting and retention in the ER and for studying mechanisms of virus budding. New ideas and approaches to understanding how each gene functions to replicate and assemble the segmented viral genome have emerged from knowledge of the primary structure of rotavirus genes and their proteins and from knowledge of the properties of domains on individual proteins. Localization of type-specific and cross-reactive neutralizing epitopes on the outer capsid proteins is becoming increasingly useful in dissecting the protective immune response, including evaluation of vaccine trials, with the practical possibility of enhancing the production of new, more effective vaccines. Finally, future analyses with recently characterized immunologic and gene probes and new animal models can be expected to provide a basic understanding of what regulates the primary interactions of these viruses with the gastrointestinal tract and the subsequent responses of infected hosts.