Bovine tuberculosis (Mycobacterium bovis) remains a concern in Spain; although herd prevalence has been considerably reduced since the systematic implementation of the national eradication program in the 1990s, infection still affected 1.33% of herds in 2011 (per the Ministry of the Environment and Rural and Marine Affairs). In consideration of the economic and public health implications of this zoonosis, tracing the source of infection by genotyping has become a key element within the program. The most widely used technique to type M. bovis isolates is direct variable repeat (DVR) spoligotyping (1). However, in some geographical settings, this cost-effective high-throughput technique does not offer adequate discrimination (2, 3, 4) and is thus complemented with mycobacterial interspersed repetitive-unit-variablenumber tandem-repeat (MIRU-VNTR) typing (5, 6). The MIRU-VNTR technique has been evaluated in many countries using different sets of makers (2, 4, 7-15).In recent surveys in the Iberian Peninsula, SB0121 was found to be the most frequent spoligotype of M. bovis (26.3% in Portugal [16] and 27.9% in Spain [17]). These studies showed that high levels of discrimination could be achieved by spoligotyping (discriminatory index [D] ϭ 0.89 and 0.87 in Portugal and Spain, respectively), although subsequent studies from Portugal found that discrimination actually increased with MIRU-VNTR typing (12,13).In this study, we used nine MIRU-VNTR markers to type a panel of randomly selected Spanish M. bovis isolates with the spoligotype SB0121. The allelic diversity and overall discriminatory power of the technique were evaluated, using several different formats. Additionally, MIRU-VNTR was applied to all the infected animals from farms infected with M. bovis SB0121 to evaluate if heterogeneity was underdetected.
MATERIALS AND METHODSMycobacterial strains. The study was carried out in M. bovis isolates with spoligotype SB0121 that were sampled from across Spain between 1997 and 2010. To evaluate the discriminatory power of MIRU-VNTR, we used a panel (n ϭ 115) (see Table S1 in the supplemental material) comprising 111 isolates selected randomly from 3,863 M. bovis SB0121 isolates from the Spanish strain collection at the VISAVET Health Surveillance Centre affecting animals plus four additionally selected M. bovis isolates from less frequently affected species. The isolates originated from cattle (Bos taurus) (n ϭ 98), goats (Capra aegagrus hircus) (n ϭ 2), wild boars (Sus scrofa) (n ϭ 7), red deer (Cervus elaphus) (n ϭ 2), fallow deer (Dama dama) (n ϭ 2), badgers (Meles meles) (n ϭ 2 [18]), and domestic pigs (Sus scrofa domestica) (n ϭ 2).To evaluate potential heterogeneity within the farms infected by M. bovis SB0121, we randomly selected 15 farms from 1,342 cattle farms from which isolates were included in the strain collection to design a second panel that comprised all the SB0121 isolates obtained from infected animals from these farms (total n ϭ 176; number of isolates per farm, 2 to 31).