Despite the existence of effective treatment for most cases of tuberculosis (TB), still a serious health problem principally in developing countries, mortality and morbidly has a tremendous impact among adults and children. Additionally, TB affects the survival in acquired immunodeficiency virus infected people. In order to overcome this problem simple diagnostic tests and vaccines are essentials as they are the basis for treatment and control of the disease. In the last few decades, novel methodologies to diagnose TB have been developed, such as nucleic acid amplification by PCR (polymerase chain reaction) or radioactive or colorimetric probes as markers to speed mycobacterial culture detection. However, most of these technologies are based on sophistication, high cost, maintenance and needs of specialized personnel, but 90% of the countries with a high burden of TB have low income making it difficult for them to afford a diagnostic routine. Thereby, better tools need to be explored and immunological methods are attractive since more inexpensive, easy to performed and adapt to point of care test (POC). However, despite availability of rapid POC tests, such as immune chromatography platform [1] and microchip system [2], the drawback of antigen-antibody based reaction is the heterogeneity of the human immune response to mycobacterial antigens jeopardizing the accuracy of the tests, specially the specificity. It has become clear that an effective vaccine or diagnostic compound require multiple antigens or peptides, as well as appropriate isotype of antibodies, to improve coverage of genetically heterogeneous individuals [3].

The esat-6 family includes 23 low mass proteins, of which 13, due to their close relationship, can be divided into subfamilies. MT10.3 (Rv3019c, genebank ID: 888926), together with two other homologous proteins, comprise one of the subfamilies within this esat-6 gene family. The MT10.3 antigen as well as MPT-64, coded by the Rv1980c gene in the region of difference - RD2 but expressed only in mycobaterial cell actively dividing (genebank ID: 885925), have been described to be highly recognized by IgA from pleural fluid of pleural tuberculous patients [4]. However, IgG from pulmonary TB patient sera poorly recognized these antigens [5], no information on IgA serum-reactivity being noted.

In this study, we report cloning, expression and purification of two molecular recombinant fused constructions based on MT10.3 disrupted by MPT64_(173G-187A), the so called CE15, and MPT64_(91L-205A) fragments which are reactive as free peptides [6]. Furthermore, to obtain evidence that the changes introduced in the constructed fused proteins exert any effect on immune response in man, a batch of sera samples from active TB patients were examined for their recognition by IgA, IgG and IgM-ELISA, related to both single full proteins.

Bacteria and plasmid

The pGEM-T Easy plasmid was propagated in E. coli Top10 (Promega Corp, Madison, WI, USA) and pQE80L was propagated in E. coli BL21 (DE3) (QIAGEN Biotechnology Brasil Ltda, SP, BR). Mycobacterium tuberculosis H37Rv was obtained from the American Type Culture Collection (Rockvelle, MD, USA) and served as a source of DNA for amplification and cloning of the target gene fragments.

Primers

The homologous gene-specific forward (F) and reverse (R) oligonucleotides to amplify the target fragments of Rv3019c and Rv1980c genes of M. tuberculosis (GenBank) were designed as described in Table 1.

Table 1: Primers and PCR conditions used for cloning different sequences of MT10.3 and MPT-64. Underline: enzyme restriction site; F: forward primer; R: Reverse primer.

Cloning and expression of fused gene fragments of CE15_(173G-187A) and MPT64_(91L-205A) of Rv 1980c disrupting the Rv 3019c gene, MT10.3_(1M-40S) and MT10.3_(41S-96) in E. coli

The DNA coding sequences of each dominant gene fragment, MT10.3_(1M-40S), CE15_(173G-187A), MT10.3_(41S-96) and MPT64_(91L-205A) [7], were amplified by PCR with the respective primers and genomic DNA from M. tuberculosis H37Rv as the template (Table 1). Each amplicon gene fragment was processed with the purification kit (Bioneer Trade, Shanghai, China) and ligated to the pGEM T easy vector and transfected into the E. coli strain Top 10. The four plasmids, each containing the purified amplicons, were digested with the respective enzymes to generate the target gene fragments, by 1 hour incubation at 37°C. DNA was run on a 1% agarose gel and the band was excised and extracted with kit extraction (Zymoclean Gel DNA Recovery, Zymo Research Corp, USA). Then each digested DNA fragment was ligated into the pQE80L expression vector, one by one (1:10), with T4 DNA ligase (Fermentas, Waltham, MA, USA). Ligation products were transformed into E. coli BL21 (DE3) competent cells, and positive recombinants were selected for sequencing by Fiocruz Sequencing Platform (Fundação Oswaldo Cruz, RJ, BR). Schematic representation of these procedures is shown in figure 1a. The untransformed E. coli strain served as negative control [8-10].

Figure 1a:Schematic steps followed to obtain the fused proteins based on gene fragment MPT-64 CE15 peptide (42 bp located at the position 174-187 aa of the MPT-64) or MPT-6491L-205A sequence (348bp) insertion between MT10.3 (1M-40S) (120 pb) and MT10.3 (41S-96) (171bp), respectively. Fragments were amplified by PCR, from M. tuberculosis DNA, with homologous primers containing enzymes restrictions sites. As the CE15 fragment was not able to be ligated at pQE80L expression vector, the successful construction containing pQE80L-MT10.3 (1M-40S):MPT-64(91L-205A):MT10.3(41S-96) (F2) was used. The fragment MPT-64(91L-205A) was digested, and into SacI/SalI predigested F2 was inserted the same enzymes pre-digest CE15 PCR fragment generating the pQE80L-MT10.3 (1M-40S):CE15 (173G-187A):MT10.3 (41S-96) (F1).

Induction and expression of target proteins: cell extracts of E. coli containing the sequenced-verified recombinants, pQE80L-MT10.3_(1M-40S):CE15_(173G-187A):MT10.3_(41S-96) and pQE80L-MT10.3_(1M-40S): MPT64_(91L-205A):MT10.3_(41S-96), were grown in Luria Bertani (LB) liquid medium with 100 mg/L of ampicillin. After shaking the two cultures at 37°C, the absorbance (A₆₀₀) was adjusted to 0.6-0.8. Expressions were with 1 mM IPTG (isopropyl-b-D-thiogalactopyranoside, Fermentas, USA) and both cultures continued at 37°C and 35°C, respectively. After 1 hour of expression induction, the samples were collected at 1h intervals until 6 h. Bacterial cells were harvested and the expression products were resolved in 15% SDS-PAGE (Sodium Dodecyl Sulfate Poly-Acrylamide Gel Electrophoresis).

The bacterial cells were collected according to the best conditions for inducible expression. After cooling in ice, lised in denatured condition and centrifuged, the pellet and supernatant were analyzed for target protein expression with SDS-PAGE. The purification of tagged 6x histidine fusion proteins was performed under denaturing conditions with the Ni-NTA agarose protein purification column (Fermentas, USA) according to manufacturer’s instructions and analyzed by Western blotting with anti-His tagged monoclonal antibodies (GE Healthcare, Buckinghamshire, UK) [11,12].

Enzyme-linked immunosorbent assay (ELISA)

Serum: A batch of sera from 14 pulmonary tuberculosis patients (TB) as well as 22 sera from other respiratory disease patients (ORD), including patients with cancer, chronic obstructive pulmonary disease together with TB household and operational contacts were obtained from a previous study [5]. All had been vaccinated with BCG during childhood. Tuberculin skin test was performed in all ORD patients and induration ≥ 5mm was considered positive [14]. This study was approved by the Ethics Committee of the Oswaldo Cruz Foundation (number 560-10).

ELISA procedure: It was carried out to detect IgA, IgG and IgM antibodies in sera against the constructed proteins as well as the MPT-64 and MT10.3 full-length single proteins (kindly donated by Lionex Immunodiagnostic GmbH, Braunschweig, Germany), as described previously [4]. Briefly, the fused recombinant MT10.3_(1M-40S):CE15_(173G-187A):MT10.3_(41S-96) designed as F1, MT10.3_(1M-40S):MPT64_(91L-205A):MT10.3_(41S-96) signed as F2 and the full single antigens MT10.3 and MPT-64 were diluted at 0.5, 1.0 and 1.5 µg/well and 50 µl of the work solution were added in the respective wells incubated at 37^°C for 2 hours. The next day, 50 µl pool samples of TB and controls (comprising aliquots of 14 active TB and 22 of ORD sera, respectively) were added diluted from 1:50 to 1:800 to the respective wells and incubated at 37°C for 1 hour. Horseradish peroxidase (HRP) conjugate secondary antibodies were diluted, in a threefold dilution for IgA, IgG and IgM (1:500 to 1:40.000) (Pierce, Germany) and 50 µl was added per well, and a plate was placed at 37ºC/1 h. Following, there was addition of substrate for color development at a 50 µl per well (TMB; Pierce, Rockford, IL) and incubation at room temperature in the dark and stop with 50 µl of acid sulfuric solution. The reaction was measured by reading the optical density at 450 nm (OD450) with spectrophotometer Labsystems Multiskan MS Plate Reader (Stockholm, Sweden).

The optimal antigen coating concentration, as well as pools of TB and controls and second antibody (conjugates) dilutions were determined and used to test sera individually (Table 2). All tests were performed in duplicate and each sample was tested three times on different days. In each set of experiments, positive (pooled TB) and negative (pooled ORD control) sera specimens were adopted as references.

Table 2: Conditions used in the ELISA IgA, IgG and IgM to test the new proteins F1 and F2, as well MPT-64 and MT10.3 full proteins. ND: no done.

Cloning, expression and purification of F1-MT10.3_(1M-40S):CE15_(173G-187A):MT10.3_(41S-96) and F2-MT10_(1M-40S):MPT64m_(91L-205A):MT10.3_(41S-96G) recombinant fused proteins

The targeted gene fragments of M. tuberculosis were amplified by PCR and respectively ligated into the pGEM vector. Unfortunately the 42 bp DNA fragment (CE15) was not able to be fused between the two fragments of MT10.3 in the pQE80Lvector, maybe because of the small size. In order to overcome this problem, we used the successfully cloned pQE80L-MT10_(1M-40S):MPT64m_(91L-205A):MT10.3_(41S-96G), digested the MPT64_(91L-205A) peptide and ligated the digested CE15 PCR product, transfecting it into the E. coli BL21(DE3) strain (Figure 1a). This strategy generated a fused gene with 332 pb (F1), the F2 construction containing 684 bp. Sequencing of recombinant plasmids generated genetic identity sequences as that of M. tuberculosis H37Rv described in GenBank.

The screened E. coli cells, containing each pQE80L-F1 and pQE80L-F2, were better IPTG induced at 35°C growth incubation for 4 h as shown in the SDS-PAGE analysis. The recombinant proteins were mainly present in the insoluble fraction, and therefore, the purification was carried out under denaturing condition with 8 M urea lyses buffer (100 mM NaH₂PO₄; 10 mM de Tris-Cl; 8 M de urea, pH=8.0). The over expressed N terminal or C terminal 6x His-tagged recombinant proteins were confirmed by probing with anti-His tag monoclonal antibody by SDS-PAGE analyses (Figure 1b). Solubilized protein was loaded into HisLink Protein Purification Resin (Promega, USA) and eluted in imidazole (Quiagen, EUA) buffer at different concentrations (250 and 500mM) because the urea buffer pH gradient was not able to harvest most of the Ni-binding protein. Purified fused proteins were confirmed by SDS-PAGE and due to, the small size of F1, the WB resolved tagged protein was detected by anti-His tag monoclonal antibodies (GE Healthecare, USA) only after a short time membrane transference (20 min) (Figure 1b) instead of 1 h used for F2. The final renatured protein concentration was 0.5 mg/ml and 1 mg/ml for the F1 and F2 proteins, respectively, measured by Coamassie plus (Pierce, Rockford, USA). The constructed F1 proved less stable under stored condition (-20°C or -80°C). Afterwards, they were dialyze in phosphate buffer and used for immune reactive analyses.

Figure 1b: Result of agarose gel and SDS-PAGE Line 1: ladder 100pb, line 2: product PCR of F1 construct (332pb), line 3: F2 construct (684pb), Line 4: protein marker SDS-PAGE. Line 5: F1 protein (13.3 kDa); Line 6: F2 protein (24.7 kDa) purificated.

Characteristics of fused proteins and comparison with full single antigens sequences

The sequences of F1 and F2 were analyzed for linear epitopes by the Kolaskar & Tongaonkar method [9]. F1, with an estimated molecular weight of approximately 13.3 kDa, is bigger than the MT10.3 molecule (10.3 kDa), and, of the three linear epitopes identified in the MT10.3 full protein, one of them was modified as a result of the insertion of the CE15 peptide. The molecular weight of F2 was estimated at 24.7 kDa, with 11 linear epitopes predicted, of which two are different from that of full proteins, one (DSMLAVDSA) a unique epitope and the other a combination of two epitopes from MT10.3 and MPT64 with an addition of 2 amino acids, glutamic acid and leucine, (ASEQAVLSELLKVYQN), supplementary table 1.

ELISA

Using the matrix method, the optimal condition of the test is described in table 2. In order to analyze the individual immune response in TB patients to the constructed proteins (F1 and F2) and to MPT-64 and MT10.3 full antigens, cutoff values for each type of immunoglobulin was obtained from 22 ORD control patients (Figure 2). As shown in table 3, the best positivity for TB patients was detected for ELISA-IgA-F2 (66.7%) and F1 (58.3%), higher than those obtained for MPT-64 (16.7%) and MT10-3 (41.7%), maintaining the highest specificity (95.5%). Note that despite the average level of reactivity decreasing for all groups, TB patients compared with ORD controls displayed statistically significant average IgA responses against F1 (p=0.0007), F2 (p=0.0008) and the single MPT64 (p=0.0003). However, the UAC value was higher for F2 (0.920) followed by F1 and MPT-64 (0.894, each) and did not react to any TST negative control, suggesting IgA-F2 presents a better overall potential diagnosis than the other antigens. For F1-IgM there was no significant difference between pools of TB and ORD controls, so we didn’t test it in individual sera. Meanwhile comparing with both single full proteins, a substantial portion of TB sera had elevated F2-IgM (50% × 30%) but also increased in cross reactivity in ORD/TST⁺ (91% × 95.2%), exhibiting a statistically significant response only for the TB × ORD/TST^- groups (p=0.043). The ROC curve for each antigen demonstrated a slightly better F2-IgM performance than the other proteins with AUC value of 0.814. For IgG type, response against F1 (57.1%) in TB patients increased compared with that detected for single full proteins (MPT-64, 28.6% and MT10.3, 21.4%) and F2 (35.7%). Despite the significantly different average level of reactivity between both TST positive and negative ORD controls (p<0,001) and the ROC curve having a high AUC value (0.801), cross reactivity was jeopardized (87.5%) compared to MPT-64, MT10.3 and F1 (91.7%).

Figure 2: The receiver operates characteristic curve (ROC) analysis to ELISA: a) IgA b) IgG and c) IgM for the new antigens F1 and F2, as well as MPT-64 and MT10.3 full proteins.

Table 3: Differential levels in immune response by ELISA IgA, IgG and IgM to the new antigens F1 and F2, as well as MPT-64 and MT10.3, in serum samples from tuberculosis patients (TB) and controls with others respiratory diseases (ORD). SD: Standard Deviation; TST + or - : Tuberculin Skin Test Positive or Negative; AUC: Area Under Curve; a,b,cp<0.05) between groups tested (Mann Whitney test).

The MT10.3 (esx-R) gene belongs to the Esx-3 family gene of M. tuberculosis. There is limited information of the putative function of the Esx-3 system with suggested involvement in iron/zinc homeostasis, thus consistent with its essential role in M. tuberculosis [11,14]. MPT-64 is a secreted protein of ~24 kDa that, as well as MT10.3, is a T-cell recognized antigen, and in combination, there is an increased sensitivity in tuberculous pleural fluid IgA [4,12]. It is generally accepted that it will be advantageous to include several antigens, or peptides, in a serological test to improve sensitivity and specificity. This study aims to test the viability and reactivity of engineering MPT64 peptides disrupting the MT10.3 antigen. We constructed five fusion proteins, but only the F1 and F2 fusion exhibited satisfactory results. The other three constructions did not render the high sensitivity in the diagnostic test (data not shown). The F1 variant maintained the three MT10.3 epitopes, except one that was modified as a result of the insertion of the CE15 peptide, perhaps favoring TB recognition enhancement and loss in ORD/TST⁺ control but recognized by an ORD/TST^- patient, related to a single MT10.3. In order to increase reactivity to F1, we incorporated an additional single copy of CE15. However, this might have compromised the structure of the protein not being able to be expressed (data not show).

In fact, F1 was poorly expressed. It has been well reported that rare codon leads to depletion of internal tRNAs, decreasing the amount of fusion expression. The DNA sequencing confirmed the presence of 3 rare codons in F1 (CCC, CTA, CGA), a result of the combination of CE15 (CCC) and MT10.3 (CTA and CGA) [10]. A small size expressed protein has also been involved with poor expression, all this possibly contributing to F1 poor expression. F2 was constructed containing the CE15 (173G-187A) peptide and two other different epitopes, as opposed to single proteins and F1, of which one (DSMLAVDSA) was unique and the other a modified composition of the CE15 and MT10.3 peptides (ASEQAVLSELLKVYQN). This change probably was the reason for recognition improvement for F2-IgA among TB sera without jeopardizing the cross reactivity, hitherto unnoticed. In another study [13], a polyprotein containing dominant B-cell epitopes of MPT-64 and Mb8.4-TB16.3-MTB8 (38F) elicited immunoreactivity of IgG (64.37%) similar to our F2-IgA (66.7%) at high specificity. Our study displayed that MPT64 and MT10.3-IgG have similar sensitivity, as Silva et al. [5], even with a reduced number of patients. In the same study, IgG detection for antigens such as ESAT-6, 38 kDa and 16 kda (HspX) demonstrated lower reactivity (22.5% to 37.1%) to that of F2-IgA. This suggests that the selected peptides may overcome the unsatisfactory results of the majority of commercial serological tests.

The findings of this study were successful in achieving the proposed goal, which was through manipulation of the antigen molecule starting from new or modified peptides, the constructed quimerics could be cloned, expressed and the resulting proteins purified. Furthermore, their antigenicity was approached from the humoral perspective. In conclusion, the results indicated the potential of F2-IgA for serological diagnosis of pulmonary TB, and further study in large sampling is under way.

This work was parcially supported by the Brazilian Conselho Nacional de Desenvolvimento Científico e Tecnológico (Institutos Nacionais de Ciência Tecnologia, Universal and PROEP-CNPq) and Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ), Rio de Janeiro, Brasil. English review and revision by Mitchell Raymond Lishon, native of Chicago, Illinois, USA-UCLA, 1969.