Background

Brucellosis is globally one of the most severe and debilitating zoonoses affecting livestock and humans [1, 2]. Ruminants are also considered the major natural hosts, maintaining the causative agent of the disease in the environment [3,4,5]. The disease is transmitted to humans either by direct contact with infected animals or by consumption of unpasteurized dairy products [6]. In this context, high prevalence of the disease is accompanied by economic collapse for the stakeholders owing to the produced milk reduction, abortions and forced slaughter. Brucellosis is endemic in countries around the Mediterranean basin, western Asia and parts of Africa as well as Latin America [7] while it constitutes an important occupational disease for breeders, employees in slaughterhouses and veterinarians as well.

The etiological agents of brucellosis are the proteobacteria belonging in the genus Brucella (family: Brucellaceae). Twelve species have been described, with B. abortus mostly infecting cattle while B. melitensis mostly infecting small ruminants. It should be also mentioned that Brucella melitensis is responsible for the vast majority of human cases worldwide [8]. In a relevant study in Greece, more than 90% of Brucella spp. strains from clinical specimens was identified as B. melitensis [9]. It should be noted that in Greece, an ongoing vaccination and monitoring program takes place during the last two decades, in an effort to eradicate the disease by slaughtering infected animals and protecting young ones [10,11,12,13].

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 [14,15,16,17].

More specifically, many PCR based attempts have been conducted to develop an efficient protocol, able to identify Brucella genus and distinguish species. At genus specific level, Brucella detection molecular tools have been developed targeting various conserved genomic regions, such as the 16S–23S intergenic transcribed [18, 19], the BCSP31 gene [20, 21], the 16S rRNA [22], the perosamine synthetase (per) gene [23], the gene encoding the Omp2a protein antigen [24], the outer membrane proteins (omp2b, omp2a and omp31) [25] and the proteins of the omp25/omp31 family of Brucella spp. [26]. Nevertheless, the above assays vary greatly in sensitivity and specificity [27].

PCR based assays targeting species specific level identification are also numerous [28,29,30,31,32,33], whereas the multi-locus variable-number tandem-repeat analysis (MLVA) assay developed by Le Flèche et al. [30] is considered the gold standard molecular methodology for Brucella typing [34]. An appropriate PCR based detection protocol should be able to discriminate vaccine Brucella strains (RB51 and Rev1) from pathogenic ones, with many available assays towards this direction [16, 29, 35,36,37].

Altogether, the main scope of the present study was the investigation of Brucella species in ruminants from Greece, at both species and subtype level. Particularly, our objectives were to identify Brucella species and subtypes of all the strains, which are very important for epidemiologic surveillance and investigation of outbreaks in brucellosis endemic regions such as Greece [38, 39]. Eventually, in an effort to evaluate the advantages and drawbacks of different molecular techniques utilized for detection and discrimination of Brucella species, an additional goal of the study was to evaluate the efficacy and reliability of different Brucella identification methodologies.

Results

16S rRNA sequencing and phylogenetic analysis

Initially all 39 examined samples were subjected to amplification, sequencing and neighbor joining phylogenetic analysis of 16S rRNA. Sequencing results revealed only one identical haplotype in all derived sequences, confirming that all samples belonged to the genus Brucella with 100% similarity with other species of Brucella (Fig. 1). This haplotype was deposited in the GenBank database and given the accession number OM570553. Those inferences demonstrate that 16S rRNA based molecular identification of Brucella bacteria may be feasible solely up to genus level.

Fig. 1
figure 1

Phylogenetic dendrogram of the examined Brucella samples in comparison with haplotypes retrieved from GenBank, based on the 16S rRNA (a) and on the BP26 gene (b). Haplotypes found in the present study are indicated with a square

Full size image

MLVA assay

The panel 1 of MLVA minisatellite markers was applied afterwards in the 39 samples with each PCR being performed at least twice. Firstly, three pairs of primers (Bruce06, Bruce11, Bruce42) were used since they provide an easy to observe differentiation method between the species. Bruce06 amplifies an expected amplicon of 408 bp in case of Brucella melitensis and 542 bp in Brucella abortus, Bruce11 a product of 257 bp and 383 bp, and Bruce42 539 bp and 289 bp respectively. Afterwards, the pairs of Bruce08 and Bruce12 were used, followed by the remaining three markers i.e. Bruce42, Bruce 43 and Bruce55, which were applied only for samples in which a positive result was observed with at least one of the first five markers, as Bruce42, Bruce 43 and Bruce55 result in an identical amplicon for Brucella melitensis and Brucella abortus.

The majority of the samples (28 out of the 39) was successfully identified with all the 8 primer pairs of the MLVA assay Panel 1 (Supplementary Material 1). The remaining 11 samples gave the expected amplicon with two or three primer pairs (from the first five markers) and failed to give an amplicon or gave non-specific bands with the rest of the first five primer pairs (Table 1). Interestingly, these samples were of lower DNA quality in terms of purity as measured by the 260/280 absorbance ratio (Table 1).

Table 1 Geographic, animal and tissue origin of the examined samples, and Brucella identification based on various molecular techniques
Full size table

BP26 and OMP31 analysis

Identification based on the BP26 marker was in complete agreement with MLVA results, as only two haplotypes were defined, one assigned to B. melitensis and one to B. abortus, clearly differentiated from each other and identical with conspecific sequences obtained from GenBank (Fig. 1). These two haplotypes were deposited in the GenBank database and given the accession numbers OM628689 and OM628690, for B. melitensis and for B. abortus respectively. Nevertheless, this was not always the case regarding the identification with the screening PCR, where there were two samples in which the two methods did not agree (Table 1). Finally, OMP31 marker failed to provide a clear band or sequence in any of the examined samples.

Brucella infection summarized results

In total, concerning billy goats all of the examined samples were identified as Brucella melitensis, whereas in rams and bulls, only B. melitensis and B. abortus were identified, respectively. Furthermore, in 3 billy goats and in 1 ram which were PCR positive for vaccine strain Rev1, it was confirmed by MLVA. Regarding females, in 2/3 goats B. melitensis was detected, with the remaining on being B. abortus, whereas 5/6 ewes hosted B. melitensis with again the remaining one being B. abortus, and all 17 examined cows hosting B. abortus.

Discussion

In Greece, brucellosis still remains an endemic zoonosis infecting a wide range of animal species with an influence both in public health and in national economy. Although in 1977, a national program against ruminants’ brucellosis was designed and initiated, the disease is not eradicated yet. The generally accepted pathogenic agent of bovine brucellosis is B. abortus, wherea B. melitensis is only occasionally detected. Respectively, the common pathogenic agent of caprine and ovine brucellosis is B. melitensis, whereas B. abortus is only rarely found in small ruminants. Thus, small ruminants are considered as the main hosts for B. melitensis [5, 40]. Nevertheless, even though Giantzis et al. [41, 42] detected B. melitensis in bovine aborted fetuses, they found no B. abortus in aborted fetuses from ewes and goats during a five-year period of 1976-1981. The clinical, pathological and epidemiological picture of caprine brucellosis due to B. melitensis is similar to B. abortus infection in cattle. The dominant strain for human brucellosis is B. melitensis [43,44,45,46]. On the other hand, Giannakopoulos et al. [47] referred human cases in Western Greece where B. abortus was identified as an equally frequent pathogenic agent. On top of that, it is of high importance to investigate whether interspecies transmission of B. melitensis, B. abortus and the vaccine strain Rev1 may occur naturally and cause clinical disease in domestic ruminants or may perplex the standard laboratory exams (RBT & CFT). It should be emphasized that control and eradication policies may have to be sometimes readapted to the new data.

In our study, the presence of B. abortus was revealed in two such cases, detected from one ewe and one goat. This could be explained by several reasons such as the existence of mixed livestock pastures, the coexistence of sheep and goats in the same shelter with cattle, or occasionally owing to herd movements with infected but non-detected animals. The results suggest cross-species infection of B. abortus from cattle to small ruminants raised in close contact [48].

The fact that in four non-vaccinated male small ruminants, the vaccine strain Rev1 was detected (Table 1), possibly indicates that during the vaccine administration in females, errors may have occurred. As a result of the accidental vaccination, the male animals may be characterized as false seropositive during the control program. Finally, due to the aforementioned vaccine administration errors, male animals end up to the slaughterhouse while the farm comes under quarantine for a period of at least 2 to 6 weeks. For these reasons, it is important to improve molecular methods to detect the exact pathogenic agent from samples from alive seropositive animals or in case of slaughter to detect it in a rapid and effective way.

In particular, development and improvement of efficient DNA-based methods is a consistent demand, as well their comparative efficiency. In this context, here we utilized and evaluated four different molecular tools towards this scope. Particularly, apart from the “gold standard” MLVA assay, the 16S rRNA, the OMP31 and the BP26 genes were analyzed. Assays targeting Brucella non-coding genomic tandem repeats, such as the MLVA minisatellites [30] or the microsatellite loci proposed earlier [15, 49], have been developed based on the principle of the high genetic homogeneity levels of the different strains, which fail to be distinguished by classic conserved genomic region markers such as the 16S rRNA. Nevertheless, microsatellites, although are excellent markers for genotyping and genetic differentiation studies, on account of hypervariability and high mutation rates may be inadequate to assign the biovar correctly or an isolate at species level. Minisatellites (MLVA) [30] on the contrary, possess a better species identification capability and are moderately variable higher, resulting in higher discriminatory power. On the other hand, gene sequences possess lower mutation rates. In line with Gupta et al. [50] who concluded that for genus-based identification of Brucella species, 16S rRNA and 16S-23 rRNA gene are the best target, in our case examination of 16S rRNA produced one identical haplotype in all samples, thus not capable to discriminate the different Brucella species. Our results are in agreement with the same study [50] concerning identification based on BP26 gene targets.

Interestingly, BP26 worked better in lower quality samples, where not all minisatellite markers gave product (Table 1). In PCR based methods the quality and purity of Brucella spp. DNA is a crucial prerequisite before performing these methods, especially for multiplex PCR methods. Any inhibitor in DNA samples from any source can affect the result of a PCR based method. Sensitivity and specificity of most PCR-based methods are not well established and their real capability for use with clinical samples and hence diagnosis has not been validated [27]. Specifically, for investigation of Brucella presence and species identification, samples may originate from slaughterhouses or dead animals, sometimes travelling for days and eventually received at the labs highly degraded. Sequencing of the BP26 gene represents a very good alternative to MLVA minisatellite assay that according to our results constitutes an effective alternative marker in lower quality DNA samples.

Conclusions

The present research provides clear evidence for the continuing circulation of Brucella species in ruminants throughout Greece, a country that still remains therefore endemic in Brucellosis. All but one molecular techniques tested were proved informative and efficient. However our results clearly demonstrate a better performance of BP26 gene marker in lower quality samples, in terms of optical density ratio.

Methods

Sample collections

During a period of 3 years (2016-2018), 264 samples were collected from a total of 191 farmed ruminants, originating from farming units located in the mainland of Greece throughout the country. Those samples derived either from positive males to Rose Bengal Test (RBT) and/or Component Fixation Test (CFT) either from aborted fetuses of sheep, goats, and cattle. Thirty-nine (39) out of the 264 tissue samples, which derived from 30 animals, were found PCR positive using the screening Multiplex PCR detection molecular methodology of Garcia-Yoldi et al. [29] for the needs of the Greek Ministry of Rural Development [51] and were utilized for the needs of the present study. The 39 samples originated from various organs and tissues from 30 male and female domestic ruminants, i.e. 7 sheep, 7 goats and 16 bovines, including 7 testicles, 1 spleen, 5 lymph nodes, 11 embryonic rumens, 11 embryonic livers, 4 cotyledonary placentas. The above 39 samples were found positive by the Multiplex PCR Assay for Brucella spp. [29] in our laboratory [51]. All samples are displayed in detail in Table 1.

Extraction of genomic DNA from tissue sample

All procedures were performed under the Biosafety Level three (BSL3) guidelines [52]. Tissue samples were initially homogenized within 200μl sterile Phosphate Buffered Saline (PBS) Sigma- Aldrich using a tissue grinder. For homogenization, aseptic processing of all samples was performed by removal of extraneous material and further maceration and chopping into small pieces in PBS. DNA isolation was carried out using the High Pure PCR template preparation kit (Roche, Basel, Switzerland) following the manufacturer instruction with a final elution volume of 80 μl in each sample. The concentration and the quality of the extracted DNA were evaluated in a micro-volume Q5000 UV-Vis Spectrophotometer (Quawell, USA).

Molecular identification of Brucella species

Since our aim was also to evaluate the applicability of different molecular tools for Brucella detection, four different molecular techniques were applied for identification of bacterial species and assignment of Brucella positive samples to species level.

Initially the 16S rRNA gene was amplified using the universal for bacteria species primer pair 27F-1492R (Table 2) that may identify the greatest majority of bacteria species at least at genus level. This pair of primers amplifies nearly the complete 16S rRNA. The 27F and 1492R [53] primers corresponded to positions 8–27 and to positions 1492–1513 of Escherichia coli 16S rRNA, respectively [56].

Table 2 Primers used for Brucella identification
Full size table

Positive Brucella samples were then subjected to PCRs targeting the BP26 and the OMP31 genes, using the primer pairs BP26F-BP26R [54] and OMP31F-OMP31R [55] (Table 2). BP26 is a conserved gene capable of distinguishing the different Brucella species [57]. On the other hand, OMP31 gene is only present in B. melitensis. Cloning and sequencing of B. melitensis 16M OMP31 [55], a gene coding for a major Brucella outer membrane protein, verified that this gene is missing from B. abortus strains [58].

Finally, the MLVA typing panel 1 assay of minisatellite markers was applied [30, 59]. For MLVA analysis we worked with the set of primers (Bruce06, 08, 11, 12, 42, 43, 45, 55, Table 2), which amplifies 8 minisatellite markers with a good species identification capability [30]. All PCRs were performed in 20 μl final volumes, containing 0.6 pmol of each forward or reverse primer, 10 μl KAPA 2G Fast Hot Start readymix (Merck, Germany), approximately 50 ng extracted DNA and nuclease free water up to the final volume. Reactions were performed in a FastGene ULTRA Cycler (Nippon Genetics, Japan) and amplification program was as follows: after an initial denaturation at 95∘C for 3 min, 35 cycles were performed, of denaturation at 95∘C for 30 sec, annealing at 48-60∘C (Table 2) for 40 sec, and extension at 72∘C for 40-60 sec, depending on the product length (40 sec for products smaller than 1000 bp; 60 sec for products larger than 1000 bp), and final extension at 72∘C for 10 min. The amplified products were examined by electrophoresis in a 1.5% agarose gel stained with ethidium bromide (0.5 mg/ml) and photographed by photo documentation system, or in polyacrylamide gel electrophoresis stained with silver nitrate in cases of very small size products differences separation. To ensure reproducibility, each PCR was performed at least twice. Particularly for 16S rRNA, BP26 and OMP31 regions, successfully amplified products were purified using the NucleoSpin Gel and PCR Clean-up kit (Macherey-Nagel, Germany) following the manufacturer’s recommended protocol and sequenced bidirectionally in an ABI 3730xl automatic sequencer. Sequences produced, were aligned using the software MEGA 7 [60] and compared with conspecific and congeneric ones obtained from the GenBank database in Neighbor Joining phylogenetic trees that were created in the same software applying a bootstrap value of 1000 iterations.