Sequences of the 3′ Halves of the Genomes of Barley Yellow Dwarf Virus-PAV cpA Isolates that Vary in Symptom Severity

Barley yellow dwarf virus (BYDV)-PAV isolates from USA have been separated into two distinct clusters (Chay et al. (1996) Virology 219: 57–65; Chay et al. (1996) Phytopathology 86: 370–377). Following this finding we have shown that BYDV-PAV is divided into two groups cpA and cpB based on their coat protein gene sequence, and distinct host preferences (Mastari et al. (1998) Phytopathology 88: 818–821). We have sequenced the complete 3′ half of the genomes of two lethal and two mild cpA isolates and compared them with those of several known PAV cpA isolates to assess variability and locate potential determinants of severity. Open reading frames (ORFs) 3, 4, 5, 6 and the 3′ untranslated regions had different percent homologies between isolates: ORF5 (92–97%), ORF3 (88–98%) 3′-translational enhancer (87–100%) ORF4 (85–99%), 3′ untranslated region (72–97%) and ORF6 (61–99%). In contrast to the mild isolates, the field-lethal isolates (FHv1 and FHv2) fell into the same cluster, regardless of the genomic region analysed. The isolates FHv1 and FHv2 differed from mild isolates by eight amino acid substitutions in ORFs 3 and 4, and insertions in ORF5. Four amino acid substitutions in the 17-kDa protein encoded by ORF4 caused a change in local net charge in the field-lethal isolates. Two insertions of four amino acids were identified in the C-terminal half of ORF5 of the field-lethal isolates, but were not present systematically in all lethal isolates analysed. The potential relationships of these differences in predicted amino acid sequences to disease severity are discussed.