MLVA Genotyping of Brucella melitensis and Brucella abortus Isolates from Different Animal Species and Humans and Identification of Brucella suis Vaccine Strain S2 from Cattle in China

In China, brucellosis is an endemic disease and the main sources of brucellosis in animals and humans are infected sheep, cattle and swine. Brucella melitensis (biovars 1 and 3) is the predominant species, associated with sporadic cases and outbreak in humans. Isolates of B. abortus, primarily biovars 1 and 3, and B. suis biovars 1 and 3 are also associated with sporadic human brucellosis. In this study, the genetic profiles of B. melitensis and B. abortus isolates from humans and animals were analyzed and compared by multi-locus variable-number tandem-repeat analysis (MLVA). Among the B. melitensis isolates, the majority (74/82) belonged to MLVA8 genotype 42, clustering in the ‘East Mediterranean’ group. Two B. melitensis biovar 1 genotype 47 isolates, belonging to the ‘Americas’ group, were recovered; both were from the Himalayan blue sheep (Pseudois nayaur, a wild animal). The majority of B. abortus isolates (51/70) were biovar 3, genotype 36. Ten B. suis biovar 1 field isolates, including seven outbreak isolates recovered from a cattle farm in Inner Mongolia, were genetically indistinguishable from the vaccine strain S2, based on MLVA cluster analysis. MLVA analysis provided important information for epidemiological trace-back. To the best of our knowledge, this is the first report to associate Brucella cross-infection with the vaccine strain S2 based on molecular comparison of recovered isolates to the vaccine strain. MLVA typing could be an essential assay to improve brucellosis surveillance and control programs.


Introduction
Brucellosis, recognized as a zoonotic disease of global importance, is caused by bacteria of the genus Brucella, which currently encompasses ten recognized species [1][2][3]. B. melitensis predominantly infects sheep and goats, B. abortus infects cattle, and B. suis infects swine and a range of wild animals. Cross-infection of other mammalian species, including humans, may occur [4]. Brucellosis is prevalent in China, especially in northern China, where people are economically dependent on ruminant livestock. B. melitensis has been the predominant species associated with human outbreaks and sporadic cases in China; B. abortus and B. suis are also associated with sporadic epidemics [5]. Based on analysis of epidemiological data in the 1990s, the incidence of animal brucellosis is stable and relatively low, whereas the incidence of human brucellosis during this time increased. The reason was probably due to the improved surveillance or public awareness [5]. Since 2008, 21 sentinel surveillance sites for animal and human brucellosis were established in 19 provinces nationwide. During this 5 years period, a number of animal and human isolates have been collected. To compare the epidemiological relationships of Brucella isolates recovered from different sources, a multi-locus variable-number tandemrepeat analysis (MLVA) assay was used [6]. In China, brucellosis is an endemic disease where B. melitensis biovars 1 and 3 and B. abortus biovars 1 and 3 are the prevailing species. B. suis isolates are scarce [5]. MLVA genotyping of Chinese human B. melitensis isolates and B. suis biovar 3 isolates have been reported previously by our team [7]. However, none of these studies included animal isolates. Thus, the primary aim of this study was to achieve globally Brucella isolates baseline genotyping data and assess the diversity among strains for epidemiological purposes in human and animal brucellosis in China.
In animals, Brucella vaccination is widely used for the prevention and control of brucellosis. Unique live, attenuated Brucella vaccines have been used depending on the preferred host, such as B. abortus S19 for cattle and B. melitensis Rev. 1 for sheep and goats [8,9]. The two vaccines, when administered correctly, can protect live-stock from brucellosis but can still cause abortions when administered at the wrong time [10][11][12][13]. Furthermore, while the vaccines are considered sufficiently attenuated for animal use, they may still be pathogenic to humans. There are documented cases showing the pathogenic nature of strains Rev. 1 and S19 in humans [14]. Thus, for the effective monitoring of both brucellosis control programs and human disease it is important to have reliable tests to differentiate vaccine and field strains. Many molecular approaches have been developed to detect vaccine strains [15][16][17][18]. The live attenuated strain Brucella suis S2 was isolated from swine fetus by the China Institute of Veterinary Drug Control (IVDC) in 1952. It has been passaged in vitro for more than 100 generations during the last 2 decades, and is the most widely used animal vaccine against brucellosis in China. The vaccine is administered through drinking water; a dose of 10 billion bacterial gives 2-3 years of protection [5]. To date, no specific molecular markers have yet been identified for this vaccine strain.
Species identification and subtyping of Brucella isolates are very important for epidemiologic surveillance and investigation of outbreaks in brucellosis endemic regions [19,20]. Previous studies have confirmed that MLVA is a useful tool for identifying and genotyping Brucella isolates and the resultant data can be used for epidemiological trace-back investigations [21][22][23][24]. Recently, a MLVA assay was used to assess the genetic stability of the B. melitensis Rev. 1 vaccine strain [25]. The lack of specific molecular markers has hampered attempts to distinguish the S2 vaccine strain from field isolates. In this report, we also present the results of an investigation employing the Brucella MLVA assay to specifically address the identification of the S2 vaccine strain in animals in China.

Genetic characteristics of B. abortus isolates from human and animal
The majority of B. abortus isolates were recovered from cattle from Inner Mongolia, an area in China of known endemicity. The most often recovered genotype 36 was distributed in animals and humans. The human analysis was limited to four isolates. The two Sichuan isolates (bru1310 and bru1311) were collected from asymptomatic infections in geographically separate areas and they differed only by the loss of one repeat unit at locus bruce07. Two additional isolates were recovered from patients who worked on a livestock farm that had reported bovine brucellosis. Among 15 homogenous clusters, 6 clusters consisted of epidemiologically related isolates exclusively from the same province and the same time, represented an outbreak respectively. Among four human isolates, two isolates (bru1311 and bru1312) had the identical MLVA pattern with no apparent epidemiological link.

Discussion
Since the establishment of sentinel surveillance points, a collection of animal and human isolates have been collected and analyzed. Analysis of the molecular characteristics of isolates recovered from the nationwide surveillance can reflect the epidemic features of brucellosis. B. melitensis has been shown to be the leading cause of brucellosis in the different epidemic regions of China [5,7]. The Inner Mongolia Autonomous Region is the most severely affected province in China, suffering from B. melitensis as the dominant species, as well as B. abortus and B. suis as the next most prevalent species [26].
The B. melitensis isolates displayed a greater diversity than B. abortus when the MLVA11 markers are considered. It is likely that the diversity of B. melitensis isolates in China is related with the high frequency of infected sheep and their subsequent movement throughout China. MLVA8 genotypes 42, 43, 58 and 63 were detected in humans, MLVA8 genotypes 42 and 63 were detected in domestic animals. The two B. melitensis MLVA8 genotype 47 isolates recovered from Pseudois nayaur (a wild animal) in Qinghai, were exceptions to this finding. However, this is the first report of the 'Americas' group in China. Whole genome sequencing of isolates from this lineage might provide information about host adaptation and microevolution. MLVA8 genotype 42, 'East Mediterranean' group (humans and animals isolates), as we have shown, is widely distributed throughout China, and has previously been reported to be predominant in Turkey, Portugal and Spain [19,24,27,28]. In Zhejiang province, an area with low prevalence of brucellosis, applying MLVA added valuable information to the epidemiological investigation and could be used as a tool to follow the spread and control of the disease. Additionally, the cluster analysis of B. melitensis forms a heterogeneous group including all three biovars (biovars 1, 2 and 3). Neither MLVA nor multilocus sequence typing distinguished the B. melitensis biovars [19,29,30]. This indicates that variable number tandem-repeat loci and the single-nucleotide polymorphisms which provide congruent data evolve independently of the putative genetic determinants for these biovars [21].
Based upon the year of B. abortus isolation, it is evident that isolates designated as genotype 36 (by MLVA panel 1) appear to have persisted in China since the 1980s and maybe associated with spread of isolates from northern to southern China; more genotyping data will need to be collected to reenforce these observations. Although the number of human B. abortus isolates included in the study were modest, comparison of the genotypes of humans and animal isolates are very important to prevent the expansion and achieve the eradication of the disease particularly in areas with low prevalence of brucellosis.
In this study, the phenomenon of host shift is uncommon between B. suis (swine) and the accessory hosts (eg, sheep, cattle, and spotted deer). Considering MLVA results on S2 strain, we conclude that S2 is genetically very stable when the strain was passaged 40 times in vitro. We therefore believe that MLVA typing could be an essential assay to guarantee the quality and stability of the live attenuated vaccine. Detection of field isolates with the same MLVA pattern as the B. suis vaccine strain S2 raises the concern of the origin of the field isolates. The fact that strains 9, 210 and 218 were grouped together strongly suggests that the B. suis vaccine strain S2 has resulted in cross infection between these three animal species. This is not unexpected since the vaccine strain S2 was widely used in China. Results obtained by Garcia-Yoldi et al. also support that the B. melitensis vaccine strain Rev. 1 group as assayed by MLVA is genetically very homogeneous [25]. These findings suggest that the MLVA assay is useful for trace-back investigations during an epidemiological investigation.
The results presented in this study have highlighted some of the potential hazards associated with using the S2 vaccine in national control programs. Implementation of whole-flock vaccination procedures, including vaccination of pregnant animals, has led to outbreaks of abortions in several intensively managed flocks and isolation of the strain from the milk of the aborting animals [31]. Detailed analysis of relevant data must be reviewed by policy makers in order to revise current national vaccine standard guidelines/regulations. To date, human infection with the vaccine strain S2 in China has not been reported. However, in previous reports, isolates were only biotyped using conventional methods and no direct molecular linkage was shown between isolates and the commercially used S2 vaccine strain.

Over the last 20 years, the geographic distribution of brucellosis in China has been
changing from pastoral areas to periurban and urban areas, according to information extracted from the Chinese National Notifiable Disease Surveillance System. It is worth noting that brucellosis has become endemic in southern provinces in China; these provinces tend to have higher gross domestic revenue and are more developed than other areas in China. On the other hand, alterations in socioeconomic and political systems, increasing animal trade and a decreasing awareness by practitioners and public health authorities led to the reemergence of new endemic foci. Our results indicate that MLVA typing could be an essential assay to determine strain relatedness and trace-back investigation in endemic or nonendemic regions of brucellosis.

Conclusions
From the collection of isolates analyzed in this study we conclude that MLVA confirmed the epidemiological linkage in outbreak and trace-back investigations. MLVA appears to be an important molecular genotyping tool that will enable improved surveillance for brucellosis and assist investigators in assessing what approaches will inhibit or reduce transmission of the diseases among animals as well as humans [32,33].

Ethics Statement
This study is a retrospective investigation of our institute strain collection with modern typing methods. The study does not involve the collection or reporting of patient data, and no patient intervention occurred with the obtained results.
Bacterial isolates and DNA preparation 82 B. melitensis isolates from 1960s to 2010s (78 recovered from animals, 4 isolated from an outbreak in humans), 70 B. abortus isolates from 1980s to 2010s (66 recovered from animals and 4 isolated humans) and 16 B. suis biovar 1 isolates from 1960s to 2010s (12 recovered from animals and 4 from humans) were examined. The B. suis biovar 1 vaccine strain S2 and the reference strain 1330 (ATCC 23444) were included. All field isolates were stored in department of brucellosis, China CDC. Bacterial isolates were cultured on trypticase soy agar containing 5% sheep's blood (BD Diagnostic Systems, China Ltd., Beijing, China) at 37°C for 48 h. All isolates were identified as Brucella species on the basis of classical identification procedures: CO 2 requirement, H 2 S production, inhibition of growth by basic fuchsin and thionin, agglutination with monospecific antisera, and phage typing [34]. Total genomic DNA was extracted using the DNeasy Blood & Tissue Kit (Qiagen China Ltd., Beijing, China) following the manufacturer's protocol for extraction of genomic DNA from Gram-negative bacteria. Species-level identification was undertaken by the AMOS-PCR assay [35].

Stability of the MLVA loci
MLVA loci stability was investigated after in vitro passage experiments. The Brucella vaccine strain S2 was inoculated on trypticase soy agar containing 5% sheep's blood (BD Diagnostic Systems) at 37°C for 48 h. A single bacterial colony was plated to fresh medium 40 times at approximately 2-days intervals and subjected to MLVA genotyping at each interval. Isolate MLVA profiles were compared with that obtained from the original seed stock to determine the stability of tandem repeat patterns at each of the 16 loci.

Analysis of MLVA data
All data were analyzed using BioNumerics version 5.1 software (Applied Maths, Belgium). Cluster analysis was based on the categorical coefficient and unweighted pair group method using arithmetic averages (UPGMA) method. Polymorphisms at each loci were quantified using the Hunter & Gaston diversity index (HGDI) [36] via the online tool V-DICE available at the HPA website (http://www.hpabioinformatics.org.uk/cgi-bin/DICI/DICI.pl).
Resultant genotypes were compared using the web-based Brucella2012 MLVA database (http://mlva.u-psud.fr/). The genotyping data can be found in the supplementary file (Table  S1).