Search for Mycobacterium avium Subspecies paratuberculosis Antigens for the Diagnosis of Paratuberculosis

The aim of this study was to evaluate a wide panel of antigens of Mycobacterium avium subsp. paratuberculosis (MAP) to select candidates for the diagnosis of paratuberculosis (PTB). A total of 54 recombinant proteins were spotted onto nitrocellulose membranes and exposed to sera from animals with PTB (n = 25), healthy animals (n = 10), and animals experimentally infected with M. bovis (n = 8). This initial screening allowed us to select seven antigens: MAP 2513, MAP 1693, MAP 2020, MAP 0038, MAP 1272, MAP 0209c, and MAP 0210c, which reacted with sera from animals with PTB and showed little cross-reactivity with sera from healthy animals and animals experimentally infected with M. bovis. The second step was to evaluate the antigen cocktail of these seven antigens by ELISA. For this evaluation, we used sera from animals with PTB (n = 25), healthy animals (n = 26), and animals experimentally infected with M. bovis (n = 17). Using ELISA, the cocktail of the seven selected MAP antigens reacted with sera from 18 of the 25 animals with PTB and did not exhibit cross-reactivity with healthy animals and only low reactivity with animals with bovine tuberculosis. The combined application of these antigens could form part of a test which may help in the diagnosis of PTB.


Introduction
PTB is a prevalent and economically important disease that affects cattle and thus impacts on the cattle industry. It is caused by MAP.
Clinical PTB is characterized by chronic granulomatous enteritis with clinical signs of diarrhea, weight loss, decreased milk production, and mortality. However, most infected cattle show no clinical signs during the prolonged incubation stage of infection [1].
On the other hand, a number of theories have proposed that the principal infective agent of Crohn's disease, a chronic enteropathy in humans, is MAP [2][3][4]. The economic impact and possible link to Crohn's disease highlights the importance of the development of control programs at the herd level. To this end, it is necessary to improve the diagnostic methods of PTB.
Cattle are most often infected as young calves, before 6 month of age, but some studies have shown that infection may also occur in adult cattle. Fecal shedding of MAP generally starts after 2 years, and clinical symptoms appear after an incubation period of 2-10 years. In addition, the elimination of the agent through the stool is very variable [5].
Cell-mediated immune response wanes with progression of the disease and when this occurs, a humoral immune response becomes measurable. It has been shown that cattle are more likely to have a combined antibody and cellular response rather than a switch from cellular to antibody response [6][7][8]. Among tests to detect serum antibody to 2 Veterinary Medicine International MAP, ELISAs are the most widely used. Several commercial ELISA kits for bovine PTB are currently available, and multiple studies have compared their accuracy [9,10]. Comparative studies of ELISAs with different antigens have shown discrepancies in the ability of these tests to identify all infected animals [11]. Some authors have suggested that this may be due to the lack of representation of the entire range of immunodominant antigens for MAP in a given ELISA test [12]. Then, one of the crucial components of this test is the antigen used for the preparation of the ELISA test. The antigen most widely used for the serological diagnosis of PTB is PPA-3, which is the M. avium strain 18 protoplasmic antigen. Currently, antigen-based tests to detect MAP with a mixture of proteins include whole-cell sonicated extract, parcel purified antigen, and protoplasmic antigens. These antigens show variability in potency and cross-reaction. This diagnostic method has drawbacks due to the crossreaction with animals sensitized with this mycobacterium or other pathogens antigenically related to MAP [11]. Since the sequencing and analysis of the entire MAP genome was obtained [13], several specific proteins have been detected in the genome of MAP and the immunoreactivity of these proteins investigated [14]. Bannantine et al. [15] developed a spot protein array for initial antigen screening. Available diagnostic MAP antigens are reviewed in Mikkelsen et al. [16]. However, individual antigens are able to identify only a subset of PTB-infected animals. Then, a mix of antigens could be a good candidate for serological diagnosis.
In the present work, MAP antigens were obtained after fractionating proteins from the whole cell or membrane or secreted fraction, resolved with two-dimensional gels, printed in line onto nitrocellulose membranes, and analyzed with sera from animals with PTB.
MAP proteins recognized specifically by sera from PTBinfected animals were used to develop a cocktail of selected antigens to be evaluated by ELISA.

Material and Methods
2.1. Animals. Sera from a total of 43 animals from different groups (a) cattle coming from PTB-free herds and with MAP negative fecal cultures (negative control animals) (n = 10), (b) cattle naturally infected, with MAP-positive fecal cultures or ELISA-positive tests (n = 25), and (c) animals experimentally infected with M. bovis with lesions (n = 8) were used to evaluate the 54 recombinant proteins of MAP.
The final screening using a cocktail of seven selected antigens was carried out with sera from 68 cattle: 25 cows with MAP positive fecal culture or ELISA-positive tests, 26 cows from negative herds with no suspected cases of PTB and negative tests (fecal cultures, serology and INF-g), and 17 from animals experimentally infected with M. bovis and with lesion at the time of necropsy.

Selection of Antigens.
With the aim of identifing and characterizing immunoreactive proteins for their possible use in diagnosis, we extracted proteins of MAP by treatment of cells with sodium dodecyl sulfate at 50 • C [17]. The proteins were resolved by one-or two-dimensional SDS-PAGE, carried out in duplicate. Both duplicates were stained with colloidal Coomassie blue and the other was transferred to a nitrocellulose membrane (Amersham Hybond-ECL, GE Healthcare Life Sciences, Buckinghamshire, UK) to perform a Western blot with a PTB-positive bovine serum sample. Six proteins were immunoreactive with sera from animals with PTB and as nonimmunoreactive with sera from animals infected with M. avium. These proteins were cut from the gel and identified by MALDI TOF. This study identified these proteins as encoded by MAP 1962 (Glutamine synthetase), MAP 4143 (Elongation factor tu), MAP 0187c (SodA), MAP 3194, MAP 3205, and MAP 3206, as potential diagnostic antigens.
The other 48 proteins here evaluated were from CIDC-Lelystad, Central Institute for Animal Disease Control Department Bacteriology, and TSEs, Lelystad, the Netherlands.

Recombinant Antigens.
The 54 proteins were produced as recombinant. Briefly, they were cloned and purified essentially as described previously [18], where all proteins were PCR-cloned using 5 -and 3 -primers amplifying DNA fragments encoding the mature protein except for secreted proteins, which were cloned without the signal peptide. The DNA fragments were cloned into expression vectors. The antigens were cloned into the pET33b (Novagen Inc., Madison, WI, USA) and pRSET (Invitrogen Corp., Carlsbad, CA, USA) vectors, and expression was induced by the production of T7 RNA polymerase in BL21 (DE3) E. coli. These cells also produce T7 lysozyme to reduce the basal expression of the target genes. The secreted antigens described in Willemsem et al. [18] (MAP 2609, MAP 2942c and MAP 0210c) were cloned in pQE80 (Qiagen, Germantown, MD, USA). Antigens were purified using their histidine-tagged Nterminal region and Nickel-affinity columns, (1 mL HisTrap HP columns) (GE Healthcare Life Sciences, UK). Before that, solubilization of the recombinant antigen was established by a buffer containing 6 M Guanidine, 20 mM Tris (pH 8.0), 0.5 M NaCl, or 50 mM Imidazol, 0.25% CHAPS, 1 mM DTT, 0.5 mM PMSF, and 1% iso-propanol to solve proteins from inclusion bodies (Guanidine), reduce nonspecific binding (Imidazol), and decrease the amount of LPS (iso-propanol) or 8 M Urea. After affinity-purification, the antigens were dialyzed (10 kDa cut-off, except for Map 4000c, for which a 1 kDa cut-off was used) against a buffer containing 0 to 6 M Urea, 10 mM Tris (pH 8.0), 25% glycerol, 1 mM DTT, 0.5 mM PMSF, 1% iso-propanol or 0 to 7 M Urea, 10 mM Tris (pH 8.0), 10% glycerol, 1 mM DTT, 0.5 mM PMSF, and 1% iso-propanol. The urea concentration was the minimal concentration needed to keep the proteins solubilized and determined empirically.
The recombinant proteins were analyzed by Coomassiestained SDS-PAGE to test purity.

Coomassie Blue Staining.
A gel containing six of the recombinant purified proteins included in the cocktail is shown in Figure 1 (of 56.5 kDa) is shown on the first lane but is not visible because it is very close to the front of the run. The proteins were run in a 15% polyacrylamide gel and then incubated in a Coomassie blue solution 0.05% (Coomassie brilliant blue R250 0.05%, methanol 50%, acid acetic 10%) for 1.5 hours with agitation. The Coomassie blue solution was then removed and bleached with bleaching solution (methanol 50%, acetic acid 10%).

Evaluation of Humoral Response by a Line Print
Immunoassay. The panel of 54 proteins was evaluated as follows: 20 µL of each antigen was applied to a nitrocellulose membrane at a concentration of 100 µg/mL, using a semiautomatic aerosolizer (Camag Scientific Inc., Wilmington, Delaware). The membranes were blocked with 50 mL of blocking solution (5% milk TBS) for 1 h. The membranes were then placed in a "miniblotter" (Isogen BioSolutions, the Netherlands). This allowed parallel analysis of 45 sera. We evaluated serum dilutions of 1 : 100. After 1 h incubation, the serum samples were aspirated and washed three times for 10 min with TBS 1x. The membrane was then incubated for 1 h with protein G conjugated to peroxidase (1 : 1500). The membrane was washed three times for 10 min with TBS 1X and revealed by chemiluminescence (Pierce ECL western blotting, Thermo Scientific, USA). Line print immunoassay generated macroarrays of recombinant proteins which were evaluated with sera from animals with PTB. In order to identify potentially cross-reactive proteins, the macroarrays were also used to probe against sera from healthy animals and from animals experimentally infected with M. bovis. The macroarrays were subjected to densitometry analysis to provide quantification for reactivity at each spot, reported as spot intensities.

Bioinformatic and Statistical
Análisis. The panel of 54 recombinant proteins included in the present study was characterized in silico: molecular weight, location prediction, and homology with other mycobacterial proteins (Table 1). PSORTb analysis was used to predict protein localization based on a number of factors including transmembrane helices, signal peptide, motif search, and similarity to proteins with known subcellular location. (http://www.psort.org/psortb/). According to PSORTb prediction, the set of recombinant proteins contains 28 cytoplasmic proteins, 7 cytoplasmic and membrane proteins, 6 extracellular proteins, and 13 proteins with uncertain localization (Table 1).
BLAST similarity searches were performed locally on coding sequences by comparison with the GenBank nonredundant protein database (Table 1). This quantification software processes spot intensities on the array and determines the mean intensities of pixels within a spot as well as those of the background pixels around the spot. These local-background intensities are subtracted from the raw signals to obtain the local-background-corrected levels. The measured diameter that was selected for scans of different arrays was consistent for all the arrays in this study. Adjusted intensities were obtained following normalization of each spot. Five proteins (MAP 0038, MAP 1272, MAP 1693, MAP 0210c, and MAP 0209c) showed a stronger mean intensity with sera from MAP-infected animals than with sera from non-MAP-infected animals and were thus selected for a cocktail of antigens (Table 2).

Measurement of Spot
MAP 2020 and MAP 2513 were also selected because although they did not show a strong intensity with sera from MAP-infected animals, they recognized nine and six animals with PTB, respectively (Table 3), and because these proteins had not been previously evaluated or reported in the literature.

ELISA.
The antigens used for the ELISA test were PPA-3 (Allied Monitor, Inc. USA) and a cocktail with the seven antigens selected (MAP 2513, MAP 1693, MAP 2020, MAP 0038, MAP 1272, MAP 0209c, and MAP 0210c). The cocktail was prepared with 30 µg of each antigen for 1 mL of the mixture. The microtiter plates were coated at 4 • C overnight with 100 µL of 40 µg/mL PPA-3 or 20 µg/mL of the cocktail in carbonate buffer (pH: 9.6). Then, the plates were saturated with 100 µL of PBS/0.5% w/v gelatin for 1 h at 37 • C, then washed five times with PBS/0.1% Tween20 (PBS/T), and incubated for 1 h at 37 • C with 100 µL of 100-fold dilution of sera in PBS/T containing 0.5% (w/v) gelatin. The plates were then washed five times with PBS/T and incubated for 30 min at 37 • C with 100 µL of 1500-fold dilution of peroxidaseconjugated protein in PBS/T containing 0.5% (w/v) gelatin. Plates were washed five times with PBS/T, and 50 µL of peroxidase substrate was added. Optical density (OD) was measured at 405 nm.    Table 2. Serum samples with density values higher than the mean obtained with the control (PBS) were considered positive and the numbers of animals reactive with each protein are shown in Table 3. The antigens selected were those that showed stronger intensity with sera from MAP-infected animals than with sera from non-MAP-infected animals. These antigens were MAP 0038, MAP 0210c, MAP 1272, MAP 1693c, and MAP 0209, shown in bold in Tables 2 and 3. In addition, we selected two antigens, MAP 2020 and MAP 2513, because they recognized nine and six animals with PTB (Table 3), respectively, and because they had not been previously evaluated or reported in the literature.

Results
These results contributed to the development of an antigen mixture with seven antigens (MAP 0038, MAP 0210c, MAP 1272, MAP 1693c, MAP 2020, MAP 2513, and MAP 0209c).
The ORF of MAP 1272 codes for a protein that possesses an NLP/P60 domain of unknown function that is found in several lipoproteins. MAP 0210 codes for the P36/Erp protein of M. bovis, which has been studied in our laboratory [19]. PSORTB analysis software predicted that MAP 2513 is localized in the cytoplasm, MAP 0210c in the cytoplasmmembrane, and that MAP 1272 and MAP 0209c are extracellular. The remaining proteins were of unknown localization. All seven of these proteins have not been previously evaluated or reported in the literature, except for MAP 0210c, which has been studied by Willemsen et al. [18]. The cocktail of seven antigens was printed in the nitrocellulose membrane and evaluated by Line print immunoassay with sera from animals with PTB and healthy controls and animals experimentally infected with M. bovis (Figure 2). By line print immunoanalysis, 14 out of 25 of the sera from animals with PTB developed antibody response to the cocktail. This cocktail was not recognized with the sera from animals experimentally infected with M. bovis here evaluated and two serum samples from healthy animals gave a very weak signal (Figure 2). 6 Veterinary Medicine International  In addition, ELISAs with PPA-3 and with the cocktail with seven antigens were evaluated with sera from animals with PTB (n = 25), healthy animals (n = 26), and animals experimentally infected with M. bovis (n = 17).
The ELISA-PPA-3 test recognized 16 out of the 25 animals with PTB (64%) but also 12 of the 17 animals experimentally infected with M. bovis, while the ELISAcocktail detected 18 of 25 animals with PTB (72%) and only 3 of the 17 animals experimentally infected with M. bovis ( Figure 3). Both ELISAs did not have a reaction with sera from non-infected controls; however, when using sera from M. bovis-infected, ELISA-PPA-3 test recognized (12/17) 70,5% of these animals and ELISA-cocktail (3/17) 17,6% of these animals.
This new ELISA for bovine PTB showed 72% of sensitivity and had higher specificity than the ELISA with PPA-3 as antigen, using animals experimentally infected with M. bovis.

Discussion
The early and specific diagnosis of PTB is still a challenge. It has generally been believed that the early immune response to infection with MAP consisted primarily of a cellular immune response characterized by interferon gamma production, and this response would later be replaced by antibody production. However, some studies have shown that antibodies appear much earlier and therefore ELISA could be used as an early diagnostic tool [6][7][8]. Then it is necessary to characterize MAP antigens to increase the sensitivity and specificity of the ELISA test for PTB diagnosis. The MAP genome sequencing represented a significant  advance and will most likely contribute with new tools for diagnosis. The evaluation of a specific panel of antigens such as that studied in the present work is the first step in the selection of candidates to be studied at different herds with PTB of our country. This is an important area, since novel antigens that could improve the diagnosis of MAP-infected cattle are needed. The proteomic approach has been used to define specific antigens by 2D fraction of MAP proteins by several researchers [20][21][22][23][24]. Another approach to obtain specific antigens is to express recombinant proteins from cloned MAP-coding sequences and use them to construct a protein array [15]. Here we selected MAP proteins by 1D or 2D electrophoresis and developed macroarrays by line print of these proteins. These macroarrays were probed with sera from animals with PTB and healthy controls and animals experimentally infected with M. bovis.
After evaluation of 54 proteins of MAP with sera from PTB-infected, experimentally infected with M. bovis and healthy animals, we selected seven proteins, which were incorporated in an ELISA to develop an antigen-based diagnostic test (ELISA-cocktail). This new ELISA for bovine PTB showed 72% of sensitivity and had higher specificity than the ELISA with PPA-3 as antigen (Figure 3). This sensitivity was based on 25 animals from MAP fecal culture-positive or ELISA-positive cattle, because naturally infected cattle may represent various stages of MAP infection and serum samples only from culture-positive animals can express differentiated antigen patterns [25]. In addition, the use of only culturepositive animals to estimate the sensitivity of ELISA may increase the values of the test because most culture-positive animals are also ELISA-positive [11]. The results presented here suggest that several specific antigens can improve the detection of MAP infection. In fact, the profiles of antibody response varied considerably and then the antibody response to single antigens was not prominent, while simultaneous usage of several recombinant antigens is able to recognize the ongoing antibody response over time in the course of infection.
In summary, here we identified novel antigens of MAP by using multiple antigen print immunoassay. Based on this knowledge, we developed an antigen cocktail, which increased the correct diagnosis of MAP-infected animals in comparison with the results of ELISA-PPA-3. The study presents an antigen cocktail, which could be of diagnostic significance for further researches. However, the cocktail needs to be evaluated by larger sample sizes in order to estimate its sensitivity and specificity. In addition, the results here shown with the 54 proteins, indicated that other proteins not included were also good candidates. Then, new cocktails should be incorporated and evaluated to increase the sensitivity and specificity of the ELISA test for diagnosis of PTB.