Genetic Polymorphism Characteristics of Brucella canis Isolated in China

In China, brucellosis is an endemic disease typically caused by Brucella melitensis infection (biovars 1 and 3). Brucella canis infection in dogs has not traditionally recognized as a major problem. In recent years however, brucellosis resulting from Brucella canis infection has also been reported, suggesting that infections from this species may be increasing. Data concerning the epidemiology of brucellosis resulting from Brucella canis infection is limited. Therefore, the purpose of this study was to assess the diversity among Chinese Brucella canis strains for epidemiological purposes. First, we employed a 16-marker VNTR assay (Brucella MLVA-16) to assess the diversity and epidemiological relationship of 29 Brucella canis isolates from diverse locations throughout China with 38 isolates from other countries. MLVA-16 analysis separated the 67 Brucella canis isolates into 57 genotypes that grouped into five clusters with genetic similarity coefficients ranging from 67.73 to 100%. Moreover, this analysis revealed a new genotype (2-3-9-11-3-1-5-1:118), which was present in two isolates recovered from Guangxi in 1986 and 1987. Second, multiplex PCR and sequencing analysis were used to determine whether the 29 Chinese Brucella canis isolates had the characteristic BMEI1435 gene deletion. Only two isolates had this deletion. Third, amplification of the omp25 gene revealed that 26 isolates from China had a T545C mutation. Collectively, this study reveals that considerable diversity exists among Brucella canis isolates in China and provides resources for studying the genetic variation and microevolution of Brucella.


Introduction
Brucellosis, caused by various species of the Gram-negative bacterium Brucella, continues to be one of the most serious zoonotic diseases for humans and animals throughout the world [1]. In recent years, the number of published reports describing brucellosis resulting from B. canis infection has increased [2][3][4][5][6]. Although, B. canis infection has not been the major cause of brucellosis in China [7][8][9][10], recent outbreaks in Beijing and other provinces suggest that this infection may be on the rise [8]. In order to effectively prevent this disease, it is important to identify what strain of Brucella has caused the infection. However, most molecular subtyping tools and ''classical biotyping'' methods lack sufficient discriminatory power for epidemiological investigations. Thus, effective molecular subtyping tools capable of reproducibly distinguishing differences between strains must be implemented.
The Brucella genus is genetically conserved [11][12][13], making rapid and accurate identification of species in the genus difficult. For instance, distinguishing between the species B. suis and B. canis has been particularly challenging [14][15][16][17][18][19][20]. This difficulty is due to the fact that few genetic polymorphisms discriminate between these two species at the molecular level as well as B. canis being a clonal lineage that arose from B. suis [21]. Several studies have optimized PCR based tests to address this issue [22]. For example, three multiplex PCR assays, a Bruce-ladder, a 19-primer multiplex PCR, and a new Bruce-ladder multiplex PCR assay called ''v2.0 multiplex PCR'' have been developed to characterize Brucella isolates at the species level [21,23,24]. More recently, multi-locus variable-number tandem-repeat analysis (MLVA) has been confirmed as a useful tool for identifying and genotyping Brucella isolates, and the data have been used for epidemiological investigations [25][26][27][28].
Bruce-ladder analysis generally performs well and has been recommended by the World Organization for Animal Health (OIE) as a rapid and simple one-step molecular test for identification and typing of Brucella species [24]. The 19-primer multiplex PCR assay was used to differentiate isolates of the genus Brucella at the species and biovar levels [21]. However, routine and widespread use of these two multiplex PCR typing tools has revealed that they misidentify a substantial proportion of B. canis isolates as B. suis [21,29,30]. The primary reason for this misclassification is the selection of the BMEI1435 gene as a marker for B. canis. This gene was originally chosen because it was thought to be naturally deleted in isolates of B. canis and would thus yield a smaller PCR product than other Brucella species containing this gene [31]; however, the BMEI1435 gene is not deleted in all B. canis isolates. Therefore, a substantial proportion of B. canis isolates cannot be correctly identified using this approach.
In this report, we present both Brucella MLVA-16 and PCR data from Chinese B. canis isolates recovered from brucellosis outbreaks from different geographical origins. The aims of this study were to twofold. Firstly, we sought to assess the performance of MLVA-16 molecular typing assay applied to Chinese B. canis isolates and assess the diversity among B. canis strains for epidemiological purposes. Secondly, we wanted to evaluate the genetic relationship between Chinese B. canis strains and exogenous strains. We compared the results of Chinese B. canis isolates identified by deletion of the BMEI1435 gene, 19-primer multiplex PCR, Bruceladder PCR, and Bruce-ladder v2.0. In addition, because the omp25 gene can be used to distinguish B. canis from other Brucella species [11,[32][33][34], we employed SNP-based typing using the omp25 gene as an alternative target.

Strains
A collection of B. canis isolates from infected dogs was obtained from various locations in China (including ten outbreak isolates from Beijing in 2011) and characterized using classical biotyping [35] (Table 1). Growth and harvesting of Brucella cells and bacterial DNA extraction were performed as previously described [23,36]. All isolates were further examined using SNP typing, Bruce-ladder v2.0, 19-primer multiplex PCR (B. canis specific primers only) and Bruce-ladder. SNP-based typing to identify Brucella isolates at the species level was carried out as described by Gopaul et al. [32]. Bruce-ladder v2.0 was performed according to the method of Lopez-Goni et al. [24]. The 19-primer multiplex PCR was performed as described by Huber et al. [21]. Bruceladder was performed as described by Garcia-Yoldi et al. [23].

Characterization of B. canis BMEI1435 gene deletion
In order to characterize deletion of the BMEI1435 gene, primers BMEI 1434F (59-GCCAGCCACAGGATCAGGTGAT-39) and BMEI 1436R (59-GGATCCGTTCGTTTCGCTCG-39) [31] were used to amplify the BMEI1435 gene. The length of the product was 1674 bp (containing the BMEI1435 gene) or 607 bp (the BMEI1435 gene deletion). Products were then separated by agarose gel electrophoresis and sequenced.

Characterization of B. canis in omp25 gene
Primers were designed based on the sequence of the omp25 gene from B. melitensis 16M (omp25-F: CATGGGCGGTTTACTC; omp25-R: CGGCCAGATCATAGTTC). The omp25 gene of 29 Chinese B. canis isolates was amplified using primers omp25-F and omp25-R. The length of the product was 652 bp, and the products were amplified and sequenced by the Sanger method. These omp25 gene sequences were aligned with B. canis RM6/66 and 13 other B. canis omp25 gene sequences obtained from GenBank.
Characterization of B. canis by MLVA genotyping MLVA was performed as described previously with the following modifications [37,38]. The 16 primer pairs were divided into three groups: MLVA-8 (eight loci including bruce06, bruce08, bruce11, bruce12, bruce42, bruce43, bruce45 and bruce55), panel 2A (three loci including bruce18, bruce19 and bruce21), and panel 2B (five loci including bruce04, bruce07, bruce09, bruce16 and bruce30). Forward primers of the panel 2A and 2B loci were labeled with one of four 59-fluorescent labels (6-FAM, ROX, HEX, or TAMRA). These primers were obtained from Shenggong Biosciences, Inc., (Shanghai, China). After an initial denaturation at 94uC for 3 min, the PCR conditions were as follows: 30 cycles of denaturation at 94uC for 30 seconds, annealing at 60uC for 30 seconds, and extension at 72uC for 50 seconds. Five microliters of the panel one loci amplification products were loaded in to 2% agarose gels containing ethidium bromide (0.5 mg/ml), visualized under UV light, and photographed. To determine the number of repeats from the sample products, PCR products were purified and directly sequenced using an ABI Prism Big Dye Terminator (v3.1) cycle sequencing ready reaction kit (v5.0; Applied Biosystems, Foster City, CA, USA). The PCR products of these samples were then sequenced and compared to the sequence of B. melitensis 16M. PCR products of panel 2A and 2B loci were denatured and resolved by capillary electrophoresis on an ABI Prism 3130 automated fluorescent capillary DNA sequencer (Applied Biosystems). Fragments were sized by comparison to a ROX (carboxy-X-rhodamine)-labeled molecular ladder (MapMaker 1000; BioVentures Inc., Murfreesboro, TN, USA) with GeneMapper software version 4.0 (Applied Biosystems). Appropriate VNTR designations of the fragments were assigned based on size calling through internal software binning capabilities and the corresponding repeat copy numbers of B. melitensis 16M.

Analysis of MLVA data
All data were analyzed using BioNumerics software version 5.1 (Applied Maths, Sint-Martins-Latem, Belgium). Cluster analysis was based on the categorical coefficient and the unweighted pair group method using arithmetic averages (UPGMA). The genotypes were then compared using the web-based Brucella 2012 MLVA database (http://mlva.u-psud.fr/). The genotyping data can be found in the supplementary data (Table S1).

Molecular typing
Given that routine multiplex PCR typing tools often misidentify a substantial proportion of B. canis isolates as B. suis, 29 Chinese B. canis isolates were identified using classical biotyping method, and then further examined them by SNP typing, Bruce-ladder v2.0, 19-primer multiplex PCR (B. canis specific primers only) and Bruce-ladder. Consistent with phenotypic typing, SNP genotyping and Bruce-ladder v2.0 identified all 29 isolates as B. canis. However, only two of the 29 B. canis isolates (1124 and 233) were identified as B. canis by Bruce-ladder and 19-primer multiplex PCR. The other 27 isolates were identified as B. suis (Table 1). In addition, using the B. canis specific primers from the 19-primer multiplex PCR assay reported by Huber et al. (2009) [21], we amplified products of 836 and 1903 base pairs. These two products represented B. canis isolates lacking and containing the BMEI1435 gene, respectively. To assess the correlation between B. canis and BMEI1435 gene deletion, these 29 isolates were then analyzed for the presence of the BMEI1435 gene by PCR amplification. The BMEI1435 gene was only deleted in isolates 1124 and 233, the same two isolates identified as B. canis by Bruce ladder and 19-primer multiplex (Table 1). These data suggest that SNP typing and Bruce-ladder v2.0 assay are more effective methods for identifying B. canis and that not all B. canis strains have a deletion of the BMEI1435 gene.

Characterization of the omp25 gene in B. canis
In addition to the BMEI1435 marker, the omp25 gene has also been used to distinguish between species of Brucella. Therefore, the omp25 gene of the 29 Chinese B. canis isolates was amplified a 625 bp product. Twenty-six of 29 B. canis isolates had an omp25 T545C mutation, while the other three had no mutations (Table 1).

Discussion
B. canis was first isolated in China from domestic and imported beagles in 1984 [7]; however, the epidemiological characteristics of B. canis infection are limited. Although B. canis infection has not been a major national concern, recent reports of B. canis infections in Beijing and other provinces highlight the significance of defining the epidemiological characteristics of this organism [8]. Currently, the MLVA-16 assay is being used to analyze epidemiological correlations between various strains. This method can also be used to track the geographic origin by comparing genetic patterns of endogenous strains with foreign isolates [2,27,28]. In this study, 29 Chinese B. canis isolates from different locations were analyzed by MLVA-16 assay, and compared with 38 B. canis isolates from other countries.
This study revealed that MLVA-16 genotype 31 (249 from Liaoning in 1988 and XUE1 from Inner Mongolia in 2010) was detected in different provinces with over twenty years between outbreaks. In addition, five of ten Beijing B. canis outbreak isolates (in 2011) and the Jiangxi isolate (in 1987) belonged to MLVA-16 genotype 47 [8]. These data suggest that B. canis strains might spread throughout China. Similarly, when compared with foreign strains, the MLVA-16 pattern of the domestic B. canis isolate 251 was identical with a Korean strain (IC-1). This suggests that poor importation quarantine may account for a subset of B. canis infections. Together, these data suggest that the MLVA-16 assay can be applied to long-term surveillance, and investigation of B. canis origins and epidemiological relatedness. This information will be valuable to establish strategies for a nationwide survey of B. canis infections and potential human exposure risks.
Another layer of B. canis diversity is the presence or absence of the BMEI1435 gene. Garcia-Yoldi et al. reported that 11 B. canis isolates (from USA, Mexico, Argentina, Germany, South Africa and Japan) contained the BMEI1435 gene; however this gene was not present in 13 different B. canis isolates (from USA, Germany, Peru and United Kingdom) [29]. Similarly, Huber et al. reported that nine B. canis isolates contained the BMEI1435 gene, but this gene was absent in six B. canis isolates (isolated from dogs and one unknown source) [21]. In the present study, sequence analysis of the BMEI1435 gene region was performed after amplification using the primers BMEI1434F and BMEI1436R. Compared to the reference train RM6/66, the BMEI1435 gene was absent in only two of the 29 B. canis isolates (1124 and 233) analyzed. The remaining 27 isolates contained the BMEI1435 gene, which was identical to B. canis HSK A52141 (CP003174). These data reveal that B. canis isolates with and without the BMEI1435 gene exist in China.
Huber et al. also reported that a PCR product of 887 bp (B. canis isolates lacking the BMEI1435 gene, including B. canis reference strain RM6/66) or a faint amplicon of 1863 bp (B. canis isolates containing the BMEI1435 gene) were amplified using B. canis specific primers [21]. In this study, using the B. canis specific primers reported by Huber et al. [21], PCR products of 836 bp We also sought to identify various B. canis strains by sequencing the omp25 gene. Gene sequences from the 29 Chinese B. canis isolates were aligned with omp25 gene sequences from B. canis RM6/66 and other strains obtained from GenBank. This analysis revealed that the omp25 gene sequences of 26 Chinese B. canis isolates have the omp25 T545C mutation, and this mutation is consistent with three B. canis isolates isolated from Germany and South Africa (AM695188, AM695179 and AM695170). Like B. canis RM6/66, the omp25 gene position T545 of the remaining three isolates (1124, 233 and 231) was not mutated. The omp25 gene sequence also correlated to the presence of the BMEI1435 gene. Two of the B. canis isolates without the omp25 T545C mutation (1124 and 233) lacked the BMEI1435 gene like B. canis RM6/66. Twenty-six of the 27 remaining isolates were strains with the omp25 T545C mutation also had the BMEI1435 gene. These data revealed that the base at position 545 of the omp25 gene correlated well with the presence of the BMEI1435 gene. One exception was the isolate 231. This isolate did not have the typical B. canis deletion of BMEI1435 gene, but had the omp25 T545C mutation. Further experimentation is required to fully elucidate this relationship.

Conclusions
SNP genotyping and Bruce-ladder 2.0 assay were able to sufficiently resolve B. canis and B. suis species. B. canis isolates with and without the BMEI1435 gene were present in China. In addition, a point mutation in the omp25 gene position 545 correlated with the presence of the BMEI1435 gene. This study reveals that the MLVA-16 assay is an effective molecular tool for detecting genotype distribution of Brucella isolates from endemic and non-endemic areas, and that considerable diversity among B. canis isolates in China.