The infectious bursal disease virus (IBDV; Birnaviridae family) constitutes one of the main threats to the poultry industry worldwide. Most of the progress in the molecular epidemiology of this virus has been achieved through the study of the coding region of the capsid protein VP2. Little research has been done regarding the molecular evolution and the epidemiological implications of genetic variability of other IBDV genome regions. In this article, the gene that codes the non-structural protein VP5 was analyzed. Although this protein is not essential for the virus replication, recent evidence indicates that it could be related to the virulent phenotype and the adaptive capacity of the virus. The VP5 gene is also of evolutionary interest because it has an open reading frame that terminally overlaps with the pVP2-VP4-VP3 polyprotein coding region. In the first part of this study, the full VP5 gene of a South American strain was characterized. The results revealed that the VP5 gene of Uruguayan hypervirulent IBDV strains (vvIBDV) lacks the alternative AUG start codon characteristic of the vvIBDV strains that have been described to date. Instead, as occurs in classic and variant strains, this VP5 gene has an AUG start site located four codons downstream and, consequently, it codes for a 145 amino acid long protein rather than the putative 149 amino acid long protein of other vvIBDV. In spite of this, these viruses conserved the VP5 and VP2 amino acid signature of the hypervirulent strains and clustered with reference vvIBDV sequences. This finding may represent evidence that the VP5 gene could be evolving by changing the translation initiation site. In the second part of this study, an evolutionary analysis including the sequences reported in this study together with most of VP5 sequences available in the GenBank, showed the existence of a complex system of selective pressures controlling the evolution of the VP5 gene. Using the dN/dS index, we found a strong purifying selection exerted on the 5' terminal overlapping region of VP2 that would be constraining the evolution of VP5. These results reinforce the hypothesis that the VP5 gene was originated late in the IBDV evolution by a mechanism of genetic overprinting. The results described in this study provided new information about the dynamics of the IBDV genome and revealed some of the mechanisms at play in the evolution of this virus. Since VP5 seems to be related to viral pathogenicity, this evolutionary information might be useful to highlight the impact of the genetic variation of this protein on the epidemiology of IBDV.