Our aim was to evaluate the reverse transcription and polymerase chain reaction (RT/PCR) technique for the detection of rotavirus shedding by infected children as a routine diagnostic procedure, in comparison to the enzyme-linked immunosorbent assay (ELISA), electron microscopy (EM) and polyacrylamide gel electrophoresis (PAGE) of rotavirus double-stranded RNA. Two-hundred and twenty stool specimens were collected from infants and young children with diarrhoea, and 10-20% faecal suspensions were made. Several methods of rotavirus dsRNA extraction were assayed. Electrophoretic analysis of viral RNA was carried out on 10% polyacrylamide gels followed by silver staining. RT/PCR was performed using oligonucleotide primers specific for both 3' and 5' ends of the rotavirus gene encoding VP7 which are highly conserved among group A rotaviruses. Following RNA extraction with phenol-chloroform and ethanol precipitation, RT/PCR could detect rotaviral RNA in only 11 of 25 samples known to contain rotaviruses by conventional methods. The purification of RNA extracts by CF11 cellulose and the application of the RNAID method were equally effective in extracting RNA and/or removing inhibitory substances from the faecal samples. RT/PCR led to the detection of 66 positive samples from 220 specimens tested (30%), whilst 64 specimens were positive by ELISA (29%), 59 (26.8%) by PAGE and 56 (25.4%) by EM. In our study, RT/PCR was 100 times more sensitive than the ELISA test in detecting rotaviruses serially diluted in a faecal suspension. Although RT/PCR is theoretically much more sensitive than ELISA, PAGE and EM for detection of rotaviruses, great care must be taken to remove inhibitory substances from the enzymatic reactions. We do not consider that RT/PCR should replace immunoassays with high sensitivity and specificity for rotavirus testing in faecal samples, although this technique has other applications, like the search for rotavirus in different clinical specimens (sera, cerebrospinal fluid, respiratory secretions, etc.) and in environmental samples, as well as the typing of viral strains using serotype-specific primers.

Rotavirus-specific cytotoxic T lymphocytes (CTL) play an important role in the resolution of rotavirus infection. The outer capsid glycoprotein, VP7, elicits a class I MHC-restricted CTL response. Vaccinia virus recombinants expressing the VP7 genes from simian rotavirus SA11 (serotype G3) and from the RF strain of bovine rotavirus (serotype G6) were used to analyze the CTL activity to this antigen in BALB/c (H-2(d)) and C57BL/6 (H-2(b)) mice neonatally infected with homologous and heterologous rotaviruses. A vaccinia virus recombinant expressing the first amino-terminal 88 amino acids of VP7 was constructed and used to search for cross-reactive CTL against this region of the protein. By using synthetic Kb, Db, and Kd motif-fitting peptides two overlapping CTL epitopes have been identified located in the first hydrophobic domain (H1) of VP7. Splenocytes obtained from rotavirus SA11-infected C57BL/6 mice induced the strongest CTL response against target cells sensitized with a peptide containing a Kb-restricted CTL epitope (amino acids 8-16). A second Kd-restricted epitope (residues 5-13) was recognized by splenocytes derived from rotavirus-infected BALB/c mice. These findings reveal the existence of CTL epitopes in the H1 signal sequence of the VP7 glycoprotein that coexist with a CTL epitope (residues 31-40) previously described within the H2 region.