ABSTRACT
The molecular basis for the aberrant migration pattern of the gene 11 equivalent in rotaviruses with "short" (human DS-1) and "super short" (human 69M and bovine VMRI) electropherotypes was investigated. The mRNAs of these viruses were synthesized in vitro, and the entire gene 11 equivalent of each of these viruses was sequenced with specific synthetic oligonucleotide primers. These sequences were compared with previously published sequences of "long" pattern rotavirus gene 11 segments. The increased lengths of the gene 11 equivalents of DS-1, 69M, and VMRI are due to a prolonged, 3' untranslated region in this gene segment. The 3' untranslated region of the VMRI gene 11 equivalent contains a clear duplication of a portion of its coding sequence. A stretch of 18 consecutive nucleotides within the 330-nucleotide, 3' untranslated region of 69M is identical to a section of UK coding sequence. The DS-1 and the remainder of the 69M 3'-end additional sequences are similar to each other, but neither is similar to any other currently available rotavirus gene sequence. This finding suggests that a process other than homologous duplication is involved in the evolution of these sequences. The widespread occurrence of human and animal rotaviruses with short and super short electropherotypes provides evidence that intragenic and possibly intergenic recombinational events associated with an error-prone viral RNA polymerase may play a role in increasing the genetic repertoire of rotaviruses.
