Bovine herpesviruses 1 (BoHV-1) and 5 (BoHV-5) are closely related alphaherpesviruses infecting cattle [1, 2]. The major difference between BoHV-5 and BoHV-1 is the distinct ability of the former to cause neurological disease in cattle [3, 4]. In a natural infection via the respiratory tract, BoHV-1 and BoHV-5 replicate similarly in the nasal mucosa but they differ in their neuroinvasiveness [5, 6]. BoHV-1 neuroinvasion usually does not go further than the first order neuron located in the trigeminal ganglia, where the latent infection is established, whereas BoHV-5 is able to infect different regions of the brain [7–9].
The BoHV-5 genome is 137,821 base pairs (bp) long and is approximately 2,000 bp longer than the BoHV-1 genome, with a G + C base composition of 75% [1, 6]. These viruses exhibit an average of 82% amino acid identity amongst different proteins. The most similar proteins (≥ 95% amino acid identity) are those involved in viral DNA replication and processing of virion proteins [1, 6].
Because of the great similarity between BoHV-1 and BoHV-5, the differentiation by conventional techniques such as viral isolation, immunofluorescence and neutralization test is difficult and has implications for the diagnosis as well as for BoHV-1 vaccination and eradication programs [10, 11]. However, these viruses exhibit some important genetic and immunogenic differences, which would explain their pathogenic and epidemiological characteristics [4, 12, 13]. BoHV-5 is the causal agent of meningo-encephalitis; a condition of low morbidity and high lethality. It usually affects cattle up to 8 months old  although older animals can also be affected. This neurological disease is associated with a variety of clinical signs, including tremors, circling, bruxism, incoordination and recumbency, followed by convulsions, padding and inevitable death [15, 3, 16, 17, 5]. Although sporadic cases of meningo-encephalitis by BoHV-5 have been reported in Australia [3, 18], Italy  and Hungary , BoHV-5 infection and disease are more frequent in Argentina and Brazil; where numerous outbreaks have been described in the last decades [16, 21–24]. The rare occurrence of BoHV-5 neurological disease in areas where BoHV-1 infection is endemic could be explained by cross-protection induced by BoHV-1 natural infection or vaccination [25, 2, 26, 27].
Restriction endonuclease analysis (REA) has been widely used to compare BoHV-1 isolates [28, 29]. It proved to be particularly useful for the differentiation among various ruminant alphaherpesviruses antigenically related to BoHV-1, which show distinct DNA fingerprints [30, 31]. Additionally, this technique allows the subtyping of viruses that apparently belong to the same type [28, 32].
Traditionally, the differentiation between BoHV-1 and BoHV-5 is based on the clinico- epidemiological characteristics of the outbreaks followed by REA of viral DNA after virus isolation. However, it is not highly precise because BoHV-1 can also be responsible for neurological disease [33, 34]. Furthermore, immunoassays using monoclonal antibodies [4, 35, 36], PCR followed by REA , nested PCR  and multiplex PCR [38–40] are also available.
Different strains of BoHV-5 were previously classified and designed as BoHV-5 "a"; BoHV-5 "b" and BoHV-5 "non-a, non-b" according to REA [4, 41]. A subtle antigenic difference was shown between subtypes "a" and "b" .
The first outbreak of BoHV-5 in Argentina was described in 1982; when the reference strain A663 was isolated. This viral strain was classified by REA as subtype "b" which has only been described in Argentina so far. Therefore, the aims of this study are to compare and characterize by REA the genomes of Argentinean BoHV-5 isolates obtained from cattle with neurological disease since 1984 until now. Furthermore, we identified the mutation that leads to the change in the restriction pattern of the strain A663. We describe a simple diagnostic method based on PCR-REA system that allows an easy discrimination between BoHV-5 "a" and "b" subtypes.