Profiling Bovine Antibody Responses to Mycobacterium avium subsp. paratuberculosis Infection by Using Protein Arrays

Mycobacterial antigen preparation.A whole-cell sonicated lysate of M. avium subsp. paratuberculosis K-10 was prepared as described previously (57).

Recombinant protein production and purification.The cloning, protein production, and purification is described in detail previously (5). Briefly, maltose binding protein (MBP) fusions of M. avium subsp. paratuberculosis predicted coding sequences listed in Table 1 were constructed in Escherichia coli by using the pMAL-c2 vector (New England Biolabs, Beverly, MA). Primers were designed from the reading frame of each coding sequence and contained an XbaI site in the 5′ primer and a HindIII site in the 3′ primer for cloning purposes. The vector and amplification product were digested with XbaI and HindIII. Ligation of these restricted DNA fragments resulted in an in-frame fusion between the malE gene in the vector and the reading frame of interest. Following ligation, the products were transformed into E. coli DH5α and selected on LB agar plates containing 100 μg/ml ampicillin. Each MBP fusion protein (e.g., MBP-MAP4025) was overexpressed in 1-liter E. coli LB cultures by induction with 0.3 mM isopropyl-β-d-thiogalactopyranoside (Sigma, St. Louis, MO) and purified by affinity chromatography using an amylose resin supplied by New England Biolabs. A similar approach was used for production and purification of all MBP fusion proteins. E. coli DH5α harboring the parental plasmid pMAL-c2 was expressed, purified, and used as a control in all experiments. Purified protein from this control strain consists of an MBP fusion of the LacZ alpha peptide (MBP-LacZ). Selected fractions eluted from the amylose resin column were pooled and dialyzed using Slide-A-Lyzer cassettes (Pierce Biotechnology Inc.) in 1.5-liter phosphate-buffered saline (PBS; 150 mM NaCl, 10 mM NaPO₄, pH 7.4) with three exchanges at 4°C. Purified protein aliquots were stored at −20°C.

Quality control assessment of purified proteins.All expression clones were sequenced to confirm that the cloned insert both matched the native M. avium subsp. paratuberculosis gene and was in frame with expression signals built into the pMAL-c2 expression vector. Furthermore, final protein concentrations of pooled fractions postdialysis were determined by NanoDrop spectrometry measurements at 280 nm. All recombinant proteins were further evaluated by GelCode Blue (Pierce Biotechnology Inc., Rockford, IL) stained sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels to assess purity and to ensure that they migrated at the expected size.

Animals, antibodies, and host serum.Sera from a rabbit that was exposed to live M. avium subsp. paratuberculosis ATCC 19698 by intraperitoneal injection was used in these studies. Sera from three clinical cattle (509, 3235, and 3494) and three uninfected control cattle (109, 110, and 111) were obtained from a herd dedicated to Johne's disease research activities at the National Animal Disease Center. The infection statuses and health histories of these animals were monitored quarterly by fecal culture, ELISA, and IFN-γ release assays (Table 2). All sera from these animals were diluted 1:500. Animals 5902, 5903, and 5904 were experimentally infected by intratonsillar inoculation by using four weekly doses of 4 × 10⁶ CFU (strain K-10) in two tonsillar crypts as described in a previous study (57). Sera from all three animals experimentally infected with M. avium subsp. paratuberculosis were collected at day 321 post-intratonsillar infection and diluted 1:500 in these experiments. Sera from cow 202, experimentally infected with M. bovis (10⁴ CFU of strain 95-1315 in each tonsillar crypt), was diluted 1:500 and sera from M. avium subsp. avium-infected cow 6137 (10¹⁰ CFU of strain 3988 split between the tonsillar crypts) were also diluted 1:500 (Table 2). A monoclonal antibody to the MBP affinity tag (diluted 1:3,000) was developed at the National Animal Disease Center and has been used in previous studies (6).

Protein array production.PBS was used as the spotting buffer and diluent for all M. avium subsp. paratuberculosis proteins. The protein array was constructed on a nitrocellulose membrane using a Bio-Dot 96-well manifold apparatus (Bio-Rad, Hercules, CA). Diluted working stocks of purified proteins were stored in deep 96-well plates with a 1-ml capacity per round-bottom well (Nalgene Nunc International). Each well of the assembled 96-well dot blot apparatus was loaded with 100 μl of PBS containing the appropriate protein approximated to a final concentration of 0.5 mg/ml. The PBS-diluted samples were allowed to flow through the membrane by gravity flow. Each well was then washed in 200 μl of PBS with 0.1% Tween-20 (TPBS). A vacuum was engaged to pull the wash solution through, and the nitrocellulose membrane was removed from the apparatus and placed in a petri dish containing a blocking solution (PBS with 2% bovine serum albumin and 0.1% Tween 20). After 30 min in the blocking solution, the immunoblotting assay was performed. PBS was also included on the dot blot array as a spotting buffer control. MBP-LacZ was also added as a control to assess immunoreactivity to the affinity tag. Protein arrays were produced in groups of four with one of the arrays used as a control with the monoclonal antibody to MBP. This enabled tracking of spotting efficiency between experiments.

Immunoblot analysis of the protein array.Dot blot arrays were incubated with sera from several hosts exposed to M. avium subsp. paratuberculosis. Sera were diluted in blocking buffer and exposed to spotted arrays for 2 h at room temperature on a rocker platform. The rabbit sera were diluted 1:1,000. After three washes in TPBS, the nitrocellulose array was incubated with the appropriate horseradish peroxidase-conjugated anti-immunoglobulin G antibody diluted at 1:20,000 in TPBS-bovine serum albumin for 1.5 h. This was followed by a final three washes in TPBS. Assay development utilized SuperSignal detection reagents (Pierce) and Kodak BioMax MR film. A control strip containing unconjugated bovine antibody (heavy plus light chains; Pierce) was processed in parallel with arrays exposed to uninfected cow sera. The top spot (A) on the control strip contained 0.11 mg/ml bovine antibody, the B spot contained 0.02 mg/ml, the C spot contained 0.002 mg/ml, and so on.

Measurement of spot intensities.Quantitative spot intensities were obtained by performing a densitometric scan on autoradiograms containing antibody-probed arrays. LabWorks 4.5 software (UVP Bioimaging System; UVP Inc., Upland, CA) was used to obtain the integrated optical density values for each spot. 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 arrays in this study. Adjusted intensities were obtained following normalization of each spot with the corresponding spot on the MBP array produced from that set of experiments. Statistical analysis included the binomial test in which the null hypothesis P value was set to 0.5. Under this null hypothesis, 46 proteins were predicted to show a stronger intensity in M. avium subsp. paratuberculosis-infected animals than in non-M. avium subsp. paratuberculosis-infected animals and 46 proteins would have a weaker signal in M. avium subsp. paratuberculosis-infected animals than in non-M. avium subsp. paratuberculosis-infected animals. Any deviance from this null model was considered nonrandom.

Protein array content.All recombinant proteins represented on the M. avium subsp. paratuberculosis protein array were initially evaluated for purity by SDS-PAGE analysis. A selection of 25 proteins, representing a typical experiment, is shown in Fig. 1. The largest fusion protein purified was MAP1643 at 127 kDa (Fig. 1), and the smallest protein was MAP4228 at 50.4 kDa (data not shown). The protein with the highest yield obtained from all the purifications was MAP2155, which represents a putative transposase similar to IS6110. A total of 35 mg of this protein was purified from a 1-liter culture at a concentration of 9.57 mg/ml. The protein with the lowest yield obtained was also the smallest protein, MAP4228, which represents an IF-1 initiation factor. A complete listing of the proteins present on the array along with their sizes, purification yield and predicted function are listed in Table 1. Additional information surrounding each protein, including their similarity in other mycobacterial genomes, IPR number, and clusters of orthologous groups, etc., is listed in Table S1 in the supplemental material.

• Open in new tab
• Download powerpoint

FIG. 1.

SDS-PAGE analysis of purified recombinant fusion proteins. Two gels stained with GelCode Blue demonstrate the purity of 25 M. avium subsp. paratuberculosis recombinant fusion proteins plus the MBP-LacZ control. Most proteins show a single discrete band with limited contamination of E. coli host proteins. The MBP affinity tag is 42 kDa in size and the largest M. avium subsp. paratuberculosis protein, MAP1643, is 85 kDa in size. Therefore, this fusion protein migrates at the expected size of 127 kDa (lane 19). Protein size standards are indicated in kilodaltons in the left margin, and the migration position of the MBP-LacZ control is indicated in the right margin. Lane assignments: 1, protein size standards; 2, MBP-MAP1174c; 3, MBP-MAP1121c; 4, MBP-MAP3131; 5, MBP-MAP4198; 6, MBP-MAP2657; 7, MBP-MAP3121; 8, MBP-MAP1204; 9, MBP-MAP0961c; 10, MBP-MAP0853; 11, MBP-MAP3734c; 12, MBP-MAP2360; 13, MBP-MAP1388; 14, MBP-MAP0075; 15, MBP-LacZ; 16 and 17, protein size standards; 18, MBP-MAP3743; 19, MBP-MAP1643; 20, MBP-MAP3735c; 21, MBP-MAP3902c; 22, MBP-MAP0736; 23, MBP-MAP1609c; 24, MBP-MAP1087; 25, MBP-MAP1233; 26, MBP-MAP3761c; 27, MBP-MAP2740; 28, MBP-MAP0855; 29, MBP-MAP3751; 30, MBP-LacZ.

The 96-dot protein array was constructed on nitrocellulose filters. Represented among these spots are 89 MBP fusion proteins, 3 polyhistidine-tagged proteins, and 4 controls, including M. avium subsp. paratuberculosis sonicate (two spots; positive control), spotting buffer (one spot; negative control), and a LacZ alpha peptide fused to MBP (specificity control). The dot assignments for the protein array are shown in Fig. 2A. The three polyhistidine-tagged proteins (MAP0087, MAP2121c, and MAP3084c) are present on the upper right corner of the array (Fig. 2A, column 12). Note that MAP2121c, which represents a widely conserved and antigenic membrane protein (56), is represented twice on the array, once as an MBP fusion (spot 4F) and also a polyhistidine-tagged protein (spot 12B). Coding sequences that are present uniquely in LSPs previously reported by our laboratories (42) and others (45) are also represented on the array. In fact, nearly all the coding sequences in the initial two columns comprise the M. avium subsp. paratuberculosis-specific LSP4 segment (MAP0851 to MAP0866) reported by Behr and colleagues (45).

• Open in new tab
• Download powerpoint

FIG. 2.

Controls to test the performance of the 96-dot array. Shown are the dot assignments for the protein array (A) and two dot array (B) experiments. The dot assignments are consistent for all experiments in this study. The monoclonal antibody or serum sample used is indicated beneath each array. All MBP fusion proteins were detected by the MBP monoclonal antibody (MBP mAb), whereas the serum from uninfected cow 111 showed little to no reactivity with most of the proteins. Because background reactivity was so low with all three uninfected cow sera, a control strip (shown on the right side of the array) containing serial dilutions of unconjugated bovine antibody was processed in parallel with all arrays exposed to uninfected cow sera to ensure the experiment worked correctly. A whole-cell lysate representing a majority of the proteins produced by M. avium subsp. paratuberculosis is spotted in positions 12E and 12H for all dot arrays. Note that the MBP monoclonal antibody does not detect the three His-tagged proteins spotted in positions 12A, 12B, and 12C.

Antibody response in healthy control cattle.Because most proteins spotted on the dot array were produced as a fusion with the MBP affinity purification tag, probing with a monoclonal antibody to this affinity tag was performed to determine relative amounts of each protein spotted (Fig. 2B). The three uninfected cows (109, 110, and 111) showed weak reactivity or no reactivity to recombinant proteins or the whole-cell sonicate antigen (see Table S2 in the supplemental material). Of the three uninfected cows, cow 111 showed the most antibody reactivity to recombinant proteins on the array and is shown in Fig. 2B along with an array exposed to the MBP monoclonal antibody. These data demonstrate that all MBP-tagged fusion proteins are detectable by using the monoclonal antibody and that nonspecific reactivity is low for uninfected cattle.

Antibody profiles for M. bovis- and M. avium subsp. avium-infected cattle.In order to identify potentially cross-reactive epitopes, protein arrays were used to probe sera from M. bovis- and M. avium subsp. avium-infected cattle (Fig. 3). In general, there was more reactivity observed with M. avium subsp. avium-infected cow serum than with M. bovis-infected cow serum. These data agree with the fact that M. avium subsp. avium is more phylogenetically similar to M. avium subsp. paratuberculosis than it is to M. bovis and hence shows more antibody cross-reactivity than M. bovis. Five proteins were detected with sera from both types of infected animals (Fig. 3). The majority of M. avium subsp. paratuberculosis-specific coding sequences present within LSPs and spotted in columns 1 to 3 and elsewhere on the array were not detected with sera from either infected cow (Fig. 3 and Table 3). MAP1087 (11A) showed more reactivity in both the M. avium subsp. avium-infected and M. bovis-infected cows than any other protein present on the dot array. TBLASTn analysis of this protein with M. avium subsp. avium 104 showed 99% similarity over the full length of the protein but only 29% similarity over a partial length in the M. bovis protein (see Table S1 in the supplemental material). In conclusion, while sera from healthy, uninfected cows show little or no reactivity with the spotted proteins, some proteins do cross-react with sera from cattle exposed to other mycobacteria.

• Open in new tab
• Download powerpoint

FIG. 3.

Cross-reactive antibodies from cattle infected with mycobacteria other than M. avium subsp. paratuberculosis. Each array was exposed to serum from an infected cow as indicated beneath the image. Antibodies from the M. avium subsp. avium (M. avium)-infected cow were observed to be more cross-reactive than those from the M. bovis-infected cow. Dot assignments are as shown in Fig. 2A.

Analysis of the humoral immune response in hosts exposed to M. avium subsp. paratuberculosis.The dot arrays were initially tested using sera from a rabbit infected with live M. avium subsp. paratuberculosis (Fig. 4A). MAP1204, a homolog of Rv1478 in Mycobacterium tuberculosis, which represents a putative exported p60 homolog possibly involved in virulence (18), was among the antigens most reactive with the rabbit sera. Likewise, the major membrane protein, MAP2121c, was also strongly detected in this host. The protein arrays were next used to obtain antibody profiles for cows with clinical signs of Johne's disease as well as for experimentally infected cows (Fig. 4B). The antibody profiles for all cattle were generally similar regardless of the infection mode or health status. In addition, all experiments using cattle sera were subjected to densitometry analysis to provide quantitation for reactivity at each spot, reported as spot intensities. The mean intensity measurements for the experimentally infected cattle (n = 3) as well as the naturally infected clinical cattle (n = 3) are listed in Table 3. The strongest antigens that were consistently reactive with all the cattle serum included MAP1087 and MAP1730c (Table 3), which represent an ABC peptide transport permease and a putative GTPase protein, respectively (29). MAP1087 showed the strongest intensity values in three of the six M. avium subsp. paratuberculosis-infected cows and showed the second-strongest intensities in the remaining three cows (see Table S2 in the supplemental material). The protein showing the third-strongest intensity, MAP1204, had the highest intensity value among the proteins from the rabbit (data not shown). The MAP1204 gene codes for a putative invasion protein that possesses an Nlpc_P60 domain of unknown function that is found in several lipoproteins. PSORTb analysis software predicts that MAP1730c is located in the cytoplasm, MAP1087 is in the membrane, and MAP1204 is extracellular (19).

• Open in new tab
• Download powerpoint

FIG. 4.

Protein arrays used to assess the humoral immune response to M. avium subsp. paratuberculosis. The protein arrays were processed by immunoblot analysis with various primary sera, as indicated in the margins. (A) Arrays probed with sera from a rabbit exposed to M. avium subsp. paratuberculosis and anti-MBP antibody. mAb, monoclonal antibody. (B) Antibody profiles from one experimentally infected cow (5903) and three clinical cattle (509, 3235, and 3494). Sera from cow 5903 were collected at 321 days post-intratonsillar infection with M. avium subsp. paratuberculosis and represent an early stage of disease. Sera were collected from the other three naturally infected cows during the appearance of clinical Johne's disease. Antibody profiles are generally similar among all four cattle despite the representation of distinct infection routes and disease stages. Dot assignments are as shown in Fig. 2A.

To test the significance of these findings, a binomial test was performed on all proteins and the null hypothesis P value was set to 0.5. These calculations showed that rather than an equal 50/50 distribution of proteins showing a stronger signal with the sera from M. avium subsp. paratuberculosis-infected cows than with non-M. avium subsp. paratuberculosis-infected cows, as the null hypothesis would suggest, it showed a nonrandom distribution of 72 proteins that had higher intensity values with the M. avium subsp. paratuberculosis-infected cattle sera than with the non-M. avium subsp. paratuberculosis-infected cattle sera and 24 proteins for which the reverse was observed.

Immunogenicity of proteins encoded within LSPs.A total of 40 proteins on the array are produced by coding sequences present within the 17 LSPs identified in the M. avium subsp. paratuberculosis K-10 genome. These LSPs comprise 187 total coding sequences and span over 226 kb of genomic DNA (45). Seven of the LSPs are not present in any of the 96 M. avium complex isolates tested (45). Overall, MAP1730c, encoded within LSP9, had the most consistently strong intensities among all six sera from the infected cattle tested (Table 3; see Table S2 in the supplemental material). However, LSP9 did show some cross-reactivity with M. avium complex isolates in PCR screens (45). The three most-immunogenic proteins encoded within LSPs that were observed to be 100% specific for M. avium subsp. paratuberculosis isolates included MAP0865 and MAP0857c (LSP4) and MAP3817c (LSP16). These proteins had the sixth (MAP0865)-, eighth (MAP3817c)-, and ninth (MAP0857c)-strongest mean intensities (Table 3) and thus are excellent candidates for an antigen-based diagnostic test.