The complete nucleotide sequence of the fourth gene of 8 human rotaviruses was determined by the dideoxy chain-termination method. Gene 4, which encodes VP3, is 2359 base pairs in length and contains a single long open reading frame of 2325 bases, capable of coding for a protein of 775 amino acids, with 5' and 3' noncoding regions of 9 and 25 nucleotides respectively. The VP3 of human rotaviruses contains identical N-terminal amino acid sequences, and there is conservation of arginine at the two trypsin cleavage sites as well as conservation of clusters of amino acids in different regions of the two VP3 cleavage products, VP8 and VP5. There are 14 regions in the VP3 protein sequence which exhibit the greatest variability. These regions may represent potential antigenic sites. Alignment of amino acid sequences of asymptomatic and virulent human rotavirus strains indicates a high degree of homology (96% or more) among the asymptomatic viruses (serotypes 1, 2, 3 and 4), while homology between asymptomatic strains and virulent viruses is considerably less (76%). A high degree of conservation of amino acid sequence (90- 93%) is also observed among 3 of the virulent strains (serotypes 1, 3 and 4). At 89 positions in the protein sequence of VP3, an amino acid is conserved among asymptomatic rotaviruses, while a different amino acid is conserved among virulent rotaviruses. Notably, three of these differences are located within the short 6 amino acid cleavage region between VP8 and VP5. It is possible that some or all of this sequence dimorphism may be responsible for the difference in virulence of these two groups of human rotaviruses.
