Multi-locus variable-number tandem repeat analysis of Chinese Brucella strains isolated from 1953 to 2013

Background Brucellosis was a common human and livestock disease caused by Brucella strains, the category B priority pathogens by the US Center for Disease Control (CDC). Identified as a priority disease in human and livestock populations, the increasing incidence in recent years in China needs urgent control measures for this disease but the molecular background important for monitoring the epidemiology of Brucella strains at the national level is still lacking. Methods A total of 600 Brucella isolates collected during 60 years (from 1953 to 2013) in China were genotyped by multiple locus variable-number tandem repeat analysis (MLVA) and the variation degree of MLVA11 loci was calculated by the Hunter Gaston Diversity Index (HGDI) values. The charts and map were processed by Excel 2013, and cluster analysis and epidemiological distribution was performed using BioNumerics (version 5.1). Results The 600 representative Brucella isolates fell into 104 genotypes with 58 singleton genotypes by the MLVA11 assay, including B. melitensis biovars 2 and 3 (five main genotypes), B. abortus biovars 1 and 3 (two main genotypes), B. suis biovars 1 and 3 (three main genotypes), and B. canis (two main genotypes) respectively. While most B. suis biovar 1 and biovar 3 were respectively found in northern provinces and southern provinces, B. melitensis and B. abortus strains were dominant in China. Canine Brucellosis was only found in animals without any human cases reported. Eight Brucellosis epidemic peaks emerged during the 60 years between 1953 and 2013: 1955 – 1959, 1962 – 1969, 1971 – 1975, 1977 – 1983, 1985 – 1989, 1992 – 1997, 2000 – 2008 and 2010 – 2013 in China. Conclusions Brucellosis has its unique molecular epidemiological patterns with specific spatial and temporal distribution according to MLVA. Trial registration IDOP-D-16-00101. Electronic supplementary material The online version of this article (doi:10.1186/s40249-017-0296-0) contains supplementary material, which is available to authorized users.


Background
Brucellosis is one of the world's most widespread zoonotic diseases. It is also seen as a reemerging infectious disease because of increasing incidence in recent years. It also is identified as a priority disease needing urgent control measures in human and livestock populations in China [1][2][3][4][5][6][7]. A total of 43 486 confirmed human cases were reported in 2013 (annual incidence of 3.3 cases/100 000 population) in China, and 57 222 cases were reported in 2014. Brucellosis has been found in all China.
Previous studies have confirmed that the multiple locus variable-number tandem repeat analysis (MLVA) is a useful tool for identifying and genotyping of strains and for epidemiological trace-back investigations [8]. It was also applied in Brucella [5][6][7]. The objectives of this study was to analyze the genotyping of 600 Brucella isolates, which were isolated from 1953 to 2013 in China by multiple locus variable-number tandem repeat analysis (MLVA) in order to more comprehensively understand the patterns of molecular epidemiology of Brucellosis in China, to trace the flow direction and variation of Brucella, and to provide reference information for the prevention and control of Brucellosis in China.

Brucella isolates data
There were 2 620 human and animal field Brucella isolates, which were collected from 29 provinces in China from 1953 to 2013. These strains came from the provincial CDC which were isolated and cultured from blood and tissues (such as liver and spleen). These isolates were identified as Brucella based on distinct host specificity, pathogenicity, and minor phenotypic differences based on serotyping, phage typing, fuchsin and thionin dye sensitivity, CO 2 requirement, H 2 S production and metabolic properties [9]. They were dried and preserved in the −80°C refrigerator in our laboratory. The stratified sampling method was used to select the number of Brucella isolates for this study. The definition of strata was Brucella bioytping. The selected proportion was 20.0% in each strata. 22.9% (600/2 620) of Brucella isolates were selected for this study.

Molecular typing methods
Total genomic DNA was individually extracted from the 600 representative strains and MLVA was performed as described previously [10,11]. The MLVA11 assay involved 11 loci including Bruce06, Bruce08, Bruce11, Bruce12, Bruce18, Bruce19, Bruce21, Bruce42, Bruce43, Bruce45, and Bruce55. A total of 11 PCR primers were labeled with 5′-fluorescent 6-FAM and PCR products were resolved by capillary electrophoresis on an ABI Prism 3130 automated fluorescent capillary DNA sequencer (Applied Biosystems). The PCR fragment sizes were converted to repeat units by following the published allele numbering system (http://mlva.u-psud.fr, Brucella support website for MLVA typing). All data were analyzed by using BioNumerics (version 5.1, Applied Maths, Belgium). Cluster analysis was based on the categorical coefficient and unweighted pair group method using arithmetic averages (UPGMA) method. The variation degree of MLVA11 loci was calculated by Hunter Gaston Diversity Index (HGDI) values. The charts and map were produced by Excel 2013.
The MLVA11 assay profiled the 600 Brucella isolates into 104 genotypes with 58 singleton genotypes (Fig. 1). The number of strains for each genotype was shown in Fig. 2. The largest group (59 genotypes) belonged to the 398 B. melitensis isolates, with 23, 17 and 5 genotypes representing B. abortus, B. suis and B. canis strains respectively (Fig. 1). These strains formed two branches (1 and 2) with 12 subclusters in total. Branch 1 was composed of B. abortus and B. melitensis whereas B. suis (biovars 1, 2 and 3) and B. canis belonged to branch 2. Cluster A, the largest with 394 B. melitensis biovars 2 and 3 isolates, were further divided into 4 subclusters (A1, A2, A3 and C) with a percentage of genetic similarity coefficient of 56.8%. Cluster B, including 97 B. abortus isolates, were further divided into four subclusters (B1, B2, B3 and B4) with a percentage of genetic similarity coefficient of 75.0%. Cluster C was small,  (Fig. 2).

Discussion
For many years, phenotypic characterization for Brucella strains was the only feasible way to provide epidemiological data with the disadvantage of lacking information related to their molecular background [8,12]. Now with molecular tools, more options are available.
Although having higher discriminatory power than the MLVA11 assay, the MLVA16 assay with a set of 16 repeat loci is better for teasing out the difference of closely related isolates at fine-scale resolution whereas the MLVA11 assay is more suitable to analyze the phylogenetic and epidemiological relationships of Brucella strains because of its lesser degree of variation degree (HGDI values) [11,[13][14][15][16][17][18]. In accord with this notion, the genotyping results of MLVA11 assay have been consistent with those of MLVA71 assay genotyping [19], Brucella-specific IS711 element analysis [13], and wholegenome-based phylogeny studies [15,16]. Therefore, we chose the MLVA11 assay to study the molecular epidemiology of Brucella isolates from China.
The results of our cluster analysis using the MLVA11 assay were meaningful in terms of the isolation time frame, host and region of these strains. For the past 60 years the dominant Brucella species and biovars were B. abortus biovars 1 and 3, B. melitensis biovars 2 and 3, B. suis biovar 1 and 3, and B. canis. The cluster analysis results of the 600 Brucella isolates correlated well with their genetic variation patterns and the Brucellosis epidemics. For instances, before 1980 most B. abortus biovars 1 and 3 isolates were seen in the northern region such as Inner Mongolia, Xinjiang, Gansu and Ningxia provinces. However, after 2000 B. abortus strains have emerged in the central and southern regions like Hebei, Beijing, Chongqing and Zhejiang provinces. According to results of our MLVA analysis, there were eight epidemic periods across years of 1955-1959, 1962-1969, 1971-1975, 1977-1983, 1985-1989, 1992-1997, 2000-2008 with each peak at years of 1958, 1966, 1973, 1980, 1987, 1996, 2007 and 2013. Most of the involved livestock were either raised or processed in the northern and northwestern regions in China. Results of the MLVA analysis (Figs. 1, 5 and 6) suggested that there were three predominant but  (Fig. 6). As to the species and biovar, most B. suis biovar 1 isolates were from the northern region including Inner Mongolia and Ningxia provinces while the majority of the B. suis biovar 3 isolates were from the southern region, such as Guangdong and Guangxi provinces.
Considering the finding that most of the cluster A1 strains (93.5% of the B. melitensis biovars 2 and 3, with a percentage of genetic similarity coefficient of 62.4%) have been existing in different provinces for a long time (from 1953 to 2013) whereas clusters A2 and A3 isolates started to appear in 2013 in Inner Mongolia, Xinjiang and Shanxi provinces (Fig. 1), questions arise such as whether these cluster A2 and A3 strains are variants of other biovars, or have they been causing diseases all these years but without being isolated, what is its impact on trend of Brucellosis in China, and so on. All such questions remain to be answered in the future.