Evaluation of PCR pncA-restriction fragment length polymorphism and PCR amplification of genomic regions of difference for the identification of M. bovis strains in lymph nodes cultures

Background A rapid accurate identification of Mycobacterium bovis is essential for surveillance purposes. Objectives A PCR pncA-Restriction Fragment Length Polymorphism (RFLP) and a multiplex PCR based on the detection of 3 regions of difference (RD-PCR): RD9, RD4 and RD1 were evaluated for the identification of M. bovis in lymph nodes cultures, in Tunisia, during 2013–2015. Methods Eighty-two M. tuberculosis complex strains were identified using the biochemical tests, GenoType MTBC assay, PCR pncA-RFLP and RD-PCR. Results The PCR pncA-RFLP showed that 54 M. bovis strains, identified by GenoType MTBC, had a mutation at position 169 of pncA gene. Twenty-eight strains did not show any mutation at this position 27 M. tuberculosis isolates and one M. caprae. The PCR pncA-RFLP had a sensitivity of 100.0% (95%CI: 93.3 -100.0) and a specificity of 100.0% (95%CI: 87.9–100.0) for identifying M. bovis. The RD-PCR showed that all M. bovis strains had the RD9 and RD4 deleted but presented RD1. RD-PCR also presented high sensitivity and specificity in detecting M. bovis strains (100.0%). Conclusions PCR pncA-RFLP and RD-PCR represent very accurate and rapid tools to identify M. bovis. They can be easily implemented in each laboratory due to their low cost and easy use.

Phenotypic and biochemical tests traditionally used to identify this species are time-consuming and inaccurate 9 . The WHO recommended identifying this species to estimate the burden of zTB in each setting and prescribe an adequate treatment 1 . Various methods have been developed for this purpose. Sequencing based genotyping methods have been used as a reference standard to well differentiate between MTBC species. A set of molecular markers has been used for this aim, such as 16S rRNA, oxyR, katG, pncA, gyrA, gyrB and hsp65 [10][11] . However, sequencing-based genotyping methods are expensive and require specific equipment.
At the national reference laboratory (NRL) for mycobacteria in Tunisia, the line probe assay: Genotype MTBC (Hain Lifescience, Germany) is used for molecular identification of M. bovis strains, whereas, this method is costly (34 $ for one test) Herein, two cost-effective PCR approaches are evaluated: a PCR pncA-Restriction Fragment Length Polymorphism (RFLP) and a multiplex PCR based on the detection of three Region of Difference (RD9, RD4 and RD1) for the detection of M. bovis in lymph nodes cultures, in comparison with the line probe assay: Gen-oType MTBC assay.

Materials and methods Ethical approval
This study is approved by the ethics committee of A. Mami pneumology hospital, Ariana, Tunisia.
Clinical specimens, strains identification and phenotypic Drug susceptibility testing (DST) Two hundred sixty-four lymph nodes samples (n=264) were tested at the NRL for mycobacteria in Tunisia, during 2013-2015. They were subjected to: acid-fast bacilli smear examination, a culture in liquid medium Mycobacteria Growth Indicator Tube 960 "MGIT960" (BD, USA), a culture in solid medium: Lowenstein Jensen "LJ", and a molecular diagnosis by GeneXpert MTB/RIF (Cepheid, USA). MTBC species identification was carried out by SD TB Ag MPT64 Rapid kit (Standard Diagnostics, South Korea), biochemical tests: niacin production, nitrate reductase activity, growth on thiophene-2-carboxylic acid hydrazide and the molecular assay GenoType MTBC. To study the specificity of evaluated methods, different PCR pncA-RFLP and RD-PCR DNAs were extracted from MGIT 960 cultures. One ml of MGIT was centrifuged at 12.000 rpm for 10 min. The pellets were suspended in 200 µl of Tris EDTA Buffer (10 mMTris-Cl pH 8.0, 1 mM EDTA) and heated at 95°C for 30 min. The suspensions were then cen-trifuged at 13,000 rpm for 15 min and the supernatants were kept and frozen at -20°C. The PCR mixture (25µl) for PCR pncA-RFLP method was prepared using 2.5 µl of buffer (10×), 0.1 µl of primers pncA F et R (25µM) 11 , 2 µl of dNTP (10 mM), 2.5 µl of MgCl2 (25 mM), 2.5 µl of DNA, 0.15 µl of Bioamtik Taq polymerase (500U) and water. The amplification was performed, according to Huard et al. 11 . The PCR products (664bp) were digested by BstEII enzyme (New England, UK).

Molecular identification by PCR pncA-RFLP
PCR pncA-RFLP showed that 54 M. bovis strains presented 2 bands of 170 bp and 494 bp after digestion by BstEII (Figure 1). It showed that 28 strains presented 3 bands of 103bp, 170bp and 391bp (Figure 1). Twenty-seven were M. tuberculosis and one strain was M. caprae, according to GenoType MTBC. No amplification of pncA was detected for NTM species.

Discussion
Mycobacterium bovis is an important cause of TB in humans. Accurate, rapid identification of this species is required to allow appropriate treatment and set a strategy to monitor the cattle's disease. For this purpose, two cost-effective PCR approaches were evaluated in comparison with the molecular assay: GenoType MTBC. The molecular identification based on the polymorphism at position 169 of pncA presented very high sensitivity and specificity in detecting M. bovis strains (100.0%). This method could also represent a rapid tool to detect the natural resistance to PZA. In fact, it is known that three MTBC species are intrinsically resistant to this drug: M. bovis, M. bovis BCG, due to the pncA C169G substitution and M. canettii 14,15 . The allelic variation at oxyR position 285 has also been proposed to differentiate M. bovis from M. tuberculosis but did not distinguish between BCG and non-BCG M. bovis strains 11,16,17 . In addition, a multiplex PCR was tested to detect the presence or absence of 3RD: RD9, RD4 and RD1 12, 13 . The RDs represent the loss of genetic materials in M. bovis BCG compared to M. tuberculosis H37Rv genome 11 . All M. bovis strains in this study (n=54) had RD9 and RD4 deleted but presented RD1. Consequently, the RD-PCR showed excellent sensitivity and specificity (100.0%) for identifying M. bovis isolates.
Compared with the conventional methods, pncA-RFLP and RD-PCR represent accurate and fast tools (few hours versus many weeks for biochemical tests) to identify and differentiate M. bovis from other MTBC members and NTM species. Furthermore, they have a low cost compared to GenoType MTBC (1.8$ versus 34$ for one test) and do not require expensive equipment and reagents as squencing. This study had some limitations: first, the two methods were tested using MTBC isolates and were not evaluated directly in lymph nodes samples. Second, mutation pncA C169G was also found in M. bovis BCG strains 8,14 . In addition, some PZA resistant M. tuberculosis isolates could display a mutation at this position. Consequently, these strains could be misidentified by pncA-RFLP as M. bovis. However, M. bovis BCG is rarely isolated from lymph node samples. In addition, a recent study in Tunisia has not reported any mutation at this position in PZA resistant M. tuberculosis isolates 18 . Finally, it was shown that some M. caprae strains and some M. tuberculosis isolates belonging to lineage 3 displayed the RD4 deleted 2,19,20 . Despite this finding, RD4 cannot be ruled out until further genomic deletion will be found to well distinguish between these species 19 .
Conclusions pncA-RFLP and RD-PCR represent a rapid, accurate tools to detect M. bovis in tuberculosis lymph nodes cultures compared with phenotypic and biochemical tests. They could be implemented easily in each laboratory owing to their easy use and low cost, in comparison with the DNA strip assay: GenoType MTBC and sequencing.

Conflicts of interest
None declared.

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-forprofit sectors.

Author contributions
Imen Bouzouita, (Ph.D): conception of the work, doing experiments, interpretation of data, drafting the work, final approval and agreement . Henda Draoui: conception of the work, doing experiments, critical revising of the manuscript, final approval and agreement. Samia Mahdhi: conception of the work, doing experiments, critical revising of the manuscript, final approval and agreement Leila Essalah: conception of the work, doing experiments, critical revising of the manuscript, final approval and agreement Leila Slim Saidi (Professor): conception of the work, critical revising of the manuscript, final approval and agreement