Geographical distribution of Brucella melitensis inferred from rpoB gene variation

Introduction: Currently available tests have limitations for the identification of Brucella species and strains, and their genetic lineage. The genome sequence of the rpoB gene encoding the β-subunit of DNA-dependent RNA polymerase was investigated for its use in genotyping Brucella melitensis. Methodology: Complete rpoB gene sequences of globally distributed Brucella melitensis strains were analyzed. Single nucleotides polymorphisms (SNPs) of the rpoB gene sequences were identified and used to type Brucella melitensis strains. Results: Six DNA polymorphisms were identified, of which two (nucleotides 3201 and 558) were novel. Analysis of the geographical distribution of the strains revealed a spatial clustering pattern with rpoB type 1 representing European and American strains, rpoB type 2 representing European, African, and Asian strains, rpoB type 3 representing Mediterranean strains, and rpoB type 4 representing African (C3201T) and European (C3201T/T558A) strains. Conclusions: We report the discovery of two novel SNPs of rpoB gene that can serve as useful markers for epidemiology and geographical tracking of B. melitensis.


Introduction
Brucella spp. is small Gram-negative coccobacillus that lives as a facultative intracellular pathogen in the host.It is the causative agent of brucellosis, which primarily affects livestock and wildlife [1].Infection in animals causes abortion, low milk production, and stillbirth, affecting the output of animal husbandry and leading to economic losses [2].Humans usually acquire the infection through contact with infected animals or by ingestion of unpasteurized dairy products.Human infection can result in a chronic debilitating disease with non-specific symptoms affecting multiple organs [3,4].
Brucella spp. is a category B potential biological warfare agent that can be transmitted through aerosols [5].Rapid detection and precise identification of Brucella spp.is, therefore, essential to determine the possible native geographical and host origins.Diagnostics and identification of brucellosis based on biochemical tests are inadequate for the differentiation of Brucella spp.In a clinical setting, general identification of Brucella spp.can be performed using 16s rRNA gene sequencing.The method however, is still inadequate for differential identification of Brucella species and strains important for the determination of the origin of the infection.
Owing to the high genomic similarity among the Brucella spp., identification methods that target multiple genes, including multi-locus variable-number tandem-repeat analysis and multilocus sequencing, have been developed [6][7][8].In the meantime, use of the rpoB gene sequence for bacteria identification, especially between closely related species, has been reported [9].Thus far, this approach demonstrated high sensitivity in differentiating Brucella spp.down to their genotype through single-gene sequencing prodecure [10,11].In the present study, we explore the use of the rpoB gene to study intraspecies variation on a large set of diverse B. melitensis strains collected globally.

Methodology
Two Brucella melitensis strains isolated in Malaysia, namely MY/2009/1483 and BM/Phil/2012/1136, were previously recovered from brucellosis patients who had consumed unpasteurized cow milk in Malaysia and goat milk in the Philippines, respectively [12].Complete sequences of the rpoB gene were retrieved from draft genomes of previously reported isolates [12].
The rpoB gene sequences of other global B. melitensis strains were retrieved from the GenBank and the PATRIC databases [13].Complete rpoB gene sequences from 56 globally distributed B. melitensis strains, including two human B. melitensis MY1483/09 (accession number: PRJEB7499) [14] and Phil1136/12 (accession number: PRJEB7504), were aligned.SNPs from the sequence alignment were identified and compared.For phylogenetic analysis, multiple sequence alignments of the 56 rpoB gene sequences from B. melitensis and two from B. abortus (as outgroups) were carried out using ClustalW 2.0.12 [15].The phylogenetic tree of the rpoB genes was constructed using MrBayes v3.2.1 [16].Bayesian MCMC analysis was conducted by sampling across the entire general time reversible (GTR) model space.One million generations were run with a sampling frequency of 500, and diagnostics were calculated for every 5,000 generations.A burn-in setting of 25% was used to discard the first 500 trees.Convergence was assessed manually, with the standard deviation of split frequencies falling below 0.01.There was no obvious trend for the plot of the generation versus the log probability of the data (the log likelihood values), and the potential scale reduction factor (PSRF) was reasonably close to 1.0 for all parameters.The discriminatory power of the rpoB typing method was calculated as described previously [17] and implemented in insilico.ehu.eus(http://insilico.ehu.es/mini_tools/discriminatory_power/index.php)[18].

Results
Nucleotide differences in the B. melitensis rpoB gene variants are shown in Table 1.Using B. melitensis 16M as reference, six DNA polymorphisms were observed, two of which were novel (Table 1).The first novel DNA polymorphism, located at residue 3201, involved C to T substitution.It was present in most of the African isolates in the study and in one B. melitensis strain (UK22_04.201744)collected in the UK.This polymorphism enabled the segregation of 3201T-B.melitensis strains from other B. melitensis strains (rpoB types 1, 2 and 3).Isolates with the 3201T SNP are designated as rpoB type 4 in the present study.Within the rpoB type 4 isolates, an additional SNP at residue 558, T to A was observed exclusively in the UK22_04.201744strain.By incorporating geographical data (Figure 1), rpoB type 1 corresponded to strains collected from European and American countries, while rpoB type 2 comprised strains collected from Europe, Africa, and Asia.Both the Malaysian isolates used in this study were included in this group.Within the rpoB type 2 group, a few variants possessed only one or two unique type-2 SNPs out of the three (at codon 629, 985, and 1309).The UK3_06.201748 strain isolated in Cyprus was the only isolate to have a single SNP of rpoB type 2 at codon 1309.The strain was designated subtype a.The second rpoB type 2 variant possessing two SNPs located at codons 629 and 1309 was designated subtype b.These variants comprised strains from diverse geographical locations including Kosovo, Turkey, Somalia, and the United Kingdom.The remaining isolates, designated subtype c, possessed all three SNPs unique to rpoB type 2. The rpoB type 3 strain corresponded to isolates collected from Mediterranean region, Egypt, and Italy.Finally, rpoB type 4, which was newly coined in the present study, encompassed isolates collected from Africa (subtype a) and the United Kingdom (subtype b).
The phylogenetic analysis of the rpoB gene revealed a grouping similar to that proposed above.All the B. melitensis strains were segregated into four major groups (Figure 2) that corresponded well with the rpoB types assigned in the current study.The phylogenetic tree was first delineated into two major groups that corresponded to rpoB type 2 and rpoB types 1/3/4.The rpoB type 2 was then further delineated into three subgroups that corresponded well with rpoB type 2 subtypes a, b, and c.The group that corresponded with rpoB types 1/3/4 was further delineated into 3 sub-groups: rpoB type 1 (yellow), rpoB type 3 (green), and rpoB type 4 (blue).The discriminatory power of this rpoB typing method was 0.6917.

Discussion
Analysis of the complete rpoB gene sequence of globally diverse B. melitensis isolates revealed the existence of four distinct groups within the species.Here, we report, for the first time, a new SNP at residue 3201 within the B. melitensis species which functions as a canonical SNP to allow the delineation of B. melitensis rpoB type 4 strains corresponding to African strains that have not been previously well studied.One isolate with an additional SNP at residue 558 from the United Kingdom was grouped within rpoB type 4. Our study suggests that the SNP at residue 3201 could be an adaptation of B. melitensis after its introduction into Africa.It accumulated in the ancestral genome of the African lineage before subsequently spreading within the African continent.The nucleotide change at residue 558 might possibly represent a canonical SNP that defines the European B. melitensis of African origin.Nevertheless, it remains inconclusive whether the SNP at residue 558 is geographically informative, as more isolates are required to enable a better segregation analysis.
A previous study assigned the strains of rpoB type 2 according to three SNPs [10].With the current analysis, we note that the acquisition of the three nucleotide polymorphisms in the rpoB gene by type 2 strains reflects the evolution of this strain.The UK3_06 isolated from Cyprus was the only strain from this study that displayed one out of three SNPs specific for rpoB type 2 B. melitensis, whereas the strains characterized by two of the three SNPs had diverse geographical distributions, including Kosovo, Turkey, Somalia, and the United Kingdom.The co-circulation of the rpoB type 2 strains with two and three SNPs in Turkey [11] suggests that the B. melitensis strains (with two SNPs) could have persisted for a period of time in Turkey, thus allowing the accumulation of the third SNP (2954T) before spreading into other Asian countries.

Conclusions
In summary, our findings show a high degree of rpoB gene similarity among B. melitensis isolated from geographically closely related regions.Identification of SNPs in the rpoB gene allows rapid monitoring of the transcontinental spread of the B. melitensis strains.The rpoB gene could, therefore, serve as an alternative marker for differentiation and geographical identification of B. melitensis strains.

Figure 1 .
Figure 1.Geographical distribution of B. melitensis strains according to rpoB typing.

Figure 2 .
Figure 2. Phylogenetic tree of rpoB gene sequences from B. melitensis isolates.

Table 1 .
Nucleotide in the rpoB of 56 B. melitensis strains used in this study.