Comparison of recombinant MPB70 and SahH and a native 20-kDa protein for the detection of bovine tuberculosis by ELISA


 Background Bovine tuberculosis (bTB) is a zoonosis mainly caused by Mycobacterium bovis . Test-and-cull protocols and gross pathological examinations of abattoir animals as well as milk pasteurisation have been implemented for preventing the spread of tuberculosis from animals to humans worldwide, including in the Republic of Korea. Despite the importance of precise and rapid diagnostic tests, conventional methods, including intradermal skin tests and γ-interferon assays, are limited by the high rate of false negative results for cattle in the late infectious stage as well as laborious and time-consuming procedures. Therefore, antibody detection methods, such as enzyme-linked immunosorbent assay (ELISA), are urgently needed to supplement established approaches and to expand the diagnostic window. In this study, we developed a bTB ELISA by evaluating candidate recombinant and native proteins and various assay parameters.Results We produced recombinant MPB70 and SahH (rM70S) and a native 20-kDa protein (20K). The 20K ELISA showed 94.4% sensitivity and 98.2% specificity and had an optimal sample-to-positive (S/P) ratio cut-off of 0.531. rM70S ELISA showed 94.4% sensitivity and 97.3% specificity with an S/N ratio cutoff of 1.696.Conclusion The assays showed the same sensitivity but the specificity was higher for 20K ELISA than for rM70S ELISA. Both assays had acceptable diagnostic efficiency and are expected to be useful for bTB diagnosis in combination with established methods for herd screening and for expanding the diagnostic window.

owners and workers is an important issue [6][7][8][9][10], even though a human M. bovis infection has not been reported to date in South Korea.
Currently, test-and-destroy and abattoir surveillance as well as milk pasteurisation are used for preventing the transmission of animal tuberculosis from animals to humans in South Korea [4,5,11]. bTB has been detected by cell-mediated immunity-based diagnosis (CEMID), including intradermal skin tests (IST) and γ-interferon assays, based on the cell-mediated immune reaction to M. bovis [2,12,13]. However, CEMID has a bTB diagnostic window with a high false-negative rate and is laborious, requiring two farm visits for injection and interpretation, retesting at bTB-positive farms, and evaluations of non-reactors by IST [14][15][16]. To complement the CEMID, methods for the humoral immunity-based diagnosis (HUMID) of bTB are urgently needed.
In this study, we produced recombinant MPB70 and SahH (rM70S) and compared it with a native 20-kDa protein and a purified protein derivative (PPD) to establish a serological bTB assay with high sensitivity and specificity. To the best of our knowledge, this is the first evaluation of the combination of MPB70 and SahH antigens for bTB ELISA.

Results
Protein profile and antigenicity PPD were 22 kDa and 40 kDa. The three sources from M. avium and M. phlei did not show immunoreactive proteins against M. bovis-positive serum.
We targeted a 20-kDa major immunoreactive B cell antigen, MPB70. The 20-kDa antigen was purified at two laboratories, the Animal and Plant Quarantine Agency (APQA) and ChoongAng Vaccine Laboratory (CVL), by anion exchange chromatography [5]. We evaluated the size and protein profile by SDS-PAGE and antigenicity by western blotting (Fig. 1B). Non-reactive proteins with molecular weights exceeding 20 kDa were removed by purification.
rM70S was produced and evaluated with respect to size and antigenicity against M. bovis-positive serum (Fig. 1C). The sizes of MPB70 and SahH were 37 kDa and 75 kDa, respectively. The    The differentiation efficiency of 20K by the S/P ratio was the highest among the antigens. The ROC curve for 20K ELISA based on S/P values exhibited the highest diagnostic efficiency (Fig. 3). Optimal cutoffs for the OD value, S/N ratio, and S/P ratio were 1.149, 2.139, and 0.531, respectively (  ELISA based on the S/P ratio showed the highest diagnostic efficiency, similar to that of 20K ELISA (Fig. 3). Optimal cutoffs for the OD value, S/N ratio, and S/P ratio were 1.320, 1.696, and 0.143, respectively ( Table 2). The S/N ratio was the most appropriate criterion in M70S ELISA in terms of sensitivity (94.4%), specificity (97.3%), PPV (77.3%), and NPV (99.4%).

Discussion
We compared three bTB ELISA antigens confirmed as highly reactive to M. bovis-specific antibodies.
In particular, we used PPD as a crude protein antigen mixture, 20K as a purified protein antigen, and rM70S as a recombinant protein antigen. rM70S and 20K were more sensitive and specific in bTB ELISA than PPD. Interestingly, rM70S, which showed high sensitivity and specificity, was composed of  [12,[22][23][24][25][26]. In this study, we newly mixed MPB70 with SahH as a recombinant antigen. It was detected as an immunoreactive protein group by 2D-gel and immunoblot analyses. SahH encodes Sadenosylhomocysteinase, an enzyme that catalyses the reversible hydrolysis of Sadenosylhomocysteine to homocysteine and adenosine and is involved in the mycobacterial stress response [27,28]. SahH enhances attachment to IL-8 and promotes entry into neutrophils [29].
However, the serological characteristics of SahH have not been reported to date. In this study, the antigenicity of recombinant SahH was confirmed with a size of 75 kDa determined by western blotting (Fig. 1c). Reported values for the sensitivity and specificity of recombinant proteins are 63.0-83.2% and 75.5-98.0%, respectively [12,22,25,26]. In this study, rM70S showed higher or equivalent sensitivity and specificity than those in previous studies. The antigen combination is an important determinant of the sensitivity and specificity of ELISA. In this study, rM70S exhibited comparable sensitivity and specificity to those of the native 20 kDa ELISA antigen. Therefore, rM70S is an appropriate antigen combination with easy and simple production and standardisation.

Conclusions
Even though PPD showed the lowest specificity and sensitivity among the three ELISA antigens evaluated in this study, these values were acceptable for herd screening. PPD has been used as an IST reagent and can easily be applied to bTB ELISA. Overall, native 20K ELISA exhibited higher sensitivity and specificity than those of rM70S. However, rM70S is more convenient for commercialisation with respect to mass production, standardisation, and known concentrations [17,25].
bTB ELISA using purified and recombinant antigens will be useful for individual tests as well as herd screening [1,2,26]. In slaughterhouses, randomly selected serum samples can be evaluated, followed by the tracking of antibody-positive reactors [12,22,30]. For the control of bTB, ELISA will decrease labour and time requirements. Furthermore, HUMID can account for the gap in the diagnostic window of CEMID. Therefore, the combination of HUMID and CEMID provides a comprehensive diagnostic system for bTB.

Methods
Purified protein derivative Mycobacterium bovis AN5 was cultured on Sauton broth to harvest purified protein derivative (PPD), which was produced following the national standard protocol [11]. Briefly, after culture for 8 weeks at Diagnostic efficiency was compared based on sensitivities and specificities. Positive control and negative control sera were used for each ELISA for validation. ELISA results were analysed according to three criteria, the OD value, S/N ratio, and S/P ratio [1,17,22,23]. The optimal method for effectively diagnosing bovine tuberculosis was determined by comparisons of the sensitivity and specificity of each parameter.