In this study, 96 isolates belonging to the species B. cenocepacia IIIB and the BCC6 group, recovered from maize rhizosphere in Italy and Mexico, were characterized by using MLRT, in order to investigate the genetic diversity and relationships of bacteria associated with maize cultivated in geographically distant locations. Despite the clear relationship found between the geographic origin of isolates and grouping, identical RTs and closely related isolates were observed in geographically distant regions (Mexico and Italy). Two main complexes were identified following eBURST analysis, namely RT-4 for B. cenocepacia IIIB and RT-104 for BCC6. These two main clonal complexes included RTs shared by both Italian and Mexican maize rhizospheres, suggesting some mixing of the genotypes between the two continental regions and excluding the possibility of any kind of geographic subspeciation in the formation of these two complexes. At the genus and species level, many prokaryotes have a cosmopolitan distribution in their respective habitats and the same genotypes have often been identified in similar habitats in different geographic areas . The wide geographic distribution and substantial capability of Burkholderia spp. to colonize diverse host plants was observed in distantly separated environments [21, 24], as well as genetic identity between BCC isolates of clinical and environmental origins recovered from different countries has been proved . Grouping isolates by eBURST analysis is useful to better evaluate the RTs distribution in natural population where highly similar RTs are found, i.e. to elucidate the meaning of the presence of closely related strains in geographically separated maize rhizospheres in respect to niche specificity and adaptation. Our finding that the majority of BCC6 isolates are part of RT-104-complex suggests that large networks of closely related BCC6 isolates colonize the maize rhizosphere of plants cultivated in Italy and Mexico. On the other hand, the presence of four clonal complexes and 12 singletons within the B. cenocepacia IIIB population suggests that maize rhizosphere is commonly colonized by well adapted B. cenocepacia IIIB clones rather than large networks of closely related isolates.
In spite of its lower discriminatory power in respect to MLST (restriction fragments vs sequences), MLRT provides useful data for typing and structure population investigations [26, 28, 32, 35]. Previous MLST analyses performed on 26 Italian BCC isolates examined in the present work indicate a good correspondence between RTs and sequence types (STs) for certain isolates: i.e., three BCC6 isolates, typed by RT 34, had ST 127, and four isolates, typed by RT 81, had ST132 . Conversely, MLST and MLRT data do not always match and the same ST for different RTs as well as different STs for the same RT were occasionally found . Considering that MLRT and MLST do not rely on the same loci, we cannot strictly correlate our MLRT results with the MLST sequence database. Indeed, a previous study on S. aureus isolates  revealed that MLRT performed on the same seven loci used in MLST captures about 95% of the discrimination power of MLST, and demonstrated that MLRT approach represents a convenient alternative to MLST. The analyses of MLRT data using tools developed for MLST permit to assess clonality/recombination in our maize-rhizosphere populations. This is an important feature when assessing the risks for human health posed by opportunistic pathogens present in the natural environment. Bacterial population structures can vary from the extremes of strictly clonal to panmictic, with most populations occupying a middle ground where recombination is significant in the evolution but the emergence of epidemic clonal lineages can also occur [41–44]. The difference in the
values between complete and corrected data sets (when the RTs are taken as units) suggests that both B. cenocepacia and BCC6 group have an epidemic population structure in which occasional clones emerge and spread. Both populations are recombining in the long term but a few RTs have recently become abundant and widespread [20, 42]. Similar "epidemic" population structure has been observed in global collections of B. cenocepacia , and may occur continuously in microbial populations not affected by the severe selective constraints imposed by human activity . The
values calculated on a subset of isolates chosen on the basis of geographical origin evidenced a population structure different from that obtained considering the entire dataset. Concerning the BCC6 group, the Italian population behaved like the whole BCC6 population, showing linkage equilibrium only when RTs were taken as units (epidemic structure), while the Mexican population showed linkage equilibrium at all levels (freely recombining population structure). Regarding the B. cenocepacia IIIB populations, the Italian one was freely recombining, while the Mexican one had a clonal structure. Nevertheless, the
values of the Mexican population are quite low, which may indicate that some recombination occurs. Recombination has had an important role in the long-term evolution of B. cenocepacia and it was also found among strains from different locations [20, 32]. Most likely, the efficiency of genetic exchange mechanisms, due to BCC inherent genomic plasticity, together with ecological factors, play a crucial role.
The use of a common MLRT scheme for both B. cenocepacia IIIB and BCC6 group allowed to compare their genetic variability, relatedness, and population structure also at interspecific level. B. cenocepacia IIIB and BCC6 populations shared identical alleles but not the same RTs. In the UPGMA tree, where the genetic similarities between the restriction profiles of both B. cenocepacia IIIB and BCC6 group were represented, the isolates were grouped into two main clusters (clusters I and II) corresponding to their taxonomic status and eBURST clonal complexes; i.e., cluster I for B. cenocepacia IIIB and RT-4-complex, and cluster II for BCC6 group and RT-104-complex. Within each cluster, the occasional presence of few isolates belonging to the other BCC species is not surprising since BCC6 and B. cenocepacia IIIB are closely related, and indeed BCC6 was previously included in the B. cenocepacia species. UPGMA performed with only the isolates included in the RT-4 and RT-104 clonal complexes gave rise to a dendrogram showing two clusters exactly corresponding to them (data not shown), confirming the correspondence between eBURST and UPGMA grouping. Finally, the finding of a clear relationship between grouping and maize cultivar suggests that maize cultivars could influence rhizosphere bacterial diversity probably due to the different chemical composition of root exudates. In fact, it is well known that plant root bacterial communities are very sensitive to environmental conditions and are more strongly influenced by plant species and different cultivars rather than by other environmental factors such as soil type and agricultural practices [46–49].