Tuberculosis (TB) ranks as the second major cause of death among infectious diseases around the world. In 2014, 9.6 million people around the world became sick with TB disease. There were 1.5 million TB-related deaths worldwide. Only 66% of the TB-cases worldwide are correctly diagnosed [1]. China is one of the 22 high TB burden countries in the world, second only to India [2]. About 550 million people was infected with Mycobacterium tuberculosis (Mtb). Early diagnosis of TB is especially important for improving early treatment. The gold standard in TB diagnosis remains to be the bacteria culture from sputum or body fluid specimens, which is time-consuming and low efficiency.

Serological tests that rely on the detection of TB-specific antigens or antibodies against TB-specific antigens possess several advantages: simple, inexpensive and feasible for the diagnosis of TB compared with traditional bacteria culturing. Moreover, it could be used to diagnose extrapulmonary TB. Many TB-specific antigens have been evaluated and applied to develop serological assays [3–5]. However, based on meta-analysis of commercial TB serological test from 1990 to 2010, current TB serological assays were failed to provide sufficient sensitivity and specificity [6]. In China, the diagnosis accuracy of these serological tests was major limited by reduced specificity. The low specificity can be attributed to cross-reactivity with other mycobacteria involving environment Mtb exposure. World Health Organization (WHO) warned against the use of current inaccurate serological test for TB clinic diagnose [7]. The WHO expert group strongly encouraged further study on identifying improved serological tests.

From the previous proteomic studies, many Mtb proteins can be detected by serum antibodies [8]. However, due to complicated post-translational modification, most Mtb proteins were difficult to express and purify [9]. For the matter of fact, epitopes of Mtb proteins are essential for the antibody reaction to reach the diagnostic accuracy. Protein array research revealed that several protein fragments from bacteria culture filtrate or protein lysate were immunoreactive with serum antibodies [10, 11]. Therefore, instead of expressing recombinant antigens, using epitope peptides is an alternative approach to develop novel biomarkers of serodiagnosis [12]. Recent studies demonstrated that peptide epitopes may even improve the detection efficiency in diagnostic assays for TB [13]. High-throughput proteomic study could help to screen diagnostic antigens in form of peptides or protein fragments that contain immunodominant epitopes. In this context, combinatorial phage display has emerged as a direct method for discovering novel antigens [14–16]. Since high quality peptides are easy to access through chemical synthesis, peptide arrays become an advanced technology. Without translating DNA to protein, designed libraries of peptides display on solid surface usually a cellulose membrane or a glass chip [17]. Systematically screening immunoreactive peptides for serodiagnosis of TB provided large data of specific TB epitopes [18]. These data also indicated that the segregation between TB and healthy individuals does not cluster into specific proteins, but into specific peptide epitope ‘hotspots’ at different location of the same protein [19]. Therefore, there needs large clinical specimens to validate the candidate peptides, which were screened out by peptide arrays.

Antigen array performed on TB protein fractionations revealed several TB lipoprotein fractionations including LppZ could be recognized by serum antibodies from TB patients [11]. LppZ(Rv3006) is a cell-wall lipoprotein that contains a post-translational modification involved in glycosylation. Expressing recombinant LppZ antigen is time-consuming and challenged by the need for correct folding and modification. An alternative approach of constructing a high-content peptide array, which displays the LppZ antigen in the form of linear peptide could be used to obtain anti-peptide antibodies in sera. These peptides contained dominant epitope sequence that could recognize by serum IgG.

Therefore, we utilize cell-wall protein LppZ as an example, through two-step screening on peptide array to characterize the immunogenicity of LppZ and identify novel antigens as potential candidates for serological diagnosis of tuberculosis. Here, the entire LppZ protein sequence were dissembled and synthesized on cellulose membrane for IgG reactive epitope mapping. Reactive epitopes ware evaluated by ELISA assay and truncation assay.

This study was focused on the use of peptide array screening to discover novel serological biomarkers for TB diagnose. The cell-wall protein LppZ was utilized as an example for proving the possibility of synthetic peptides to improve the early diagnosis of patients infected with TB. Even though there will be post-translation modifications of LppZ protein such as amidation, glycosylation, and phosphorylation or other modifications, previously studies have shown that all the isoforms of LppZ protein fragments were serologically reactive, suggesting that the epitopes recognized by serum antibodies are on the peptide backbone of the core protein and not on probable modification sites [10]. Therefore, overlap mapping was performed on LppZ protein basic sequence regardless of post-translation modification.

In the process of screening, we designed two-step approach for higher probability of success. At the first round of screening, we performed epitope mapping based on the entire LppZ sequence to obtain serum IgG-bound epitopes. The second round of screening was performed by large group of individual serum samples. Based on the peptide array’s result, the diagnostic capability of each candidate epitope was given out at certain diagnostic value (Additional file 1: Table S3). After two rounds of screening, two dominant epitopes, pep-LppZ-1 and pep-LppZ-13, were selected according to our selection criteria. Truncation assay revealed that the core reaction sequence was narrow down to 8 mer amino acids for each peptide. However, when the last unnecessary amino acid was deleted, the binding affinity has been reduced (Fig. 2). Therefore, protective amino acids were needed for immunoreaction.

Then we used traditional serology test method, ELISA to validate the diagnostic capability of these peptides in another group of serum samples, which matches the sample size of peptide array screening. The ELISA result agreed with the peptide array data in terms of diagnostic ability. Compare to other research, our method would be much simpler and the verification rate is much higher [10, 16, 18]. Since the second round screening could be considered analogous to carrying out a thousand of ELISAs at one time. It would save time and cost.

According to other research findings using LppZ recombinant protein or protein fragments, the positive rate of anti-LppZ antibodies in TB patients was around 30% to 55% [11]. The specificity was not provided. In our result, one peptide was not only sufficient to obtain sensitivity similar to protein component described above, but also with a high specificity (above 90%). Combinative use of two peptides could further improve specificity, but slightly drop of sensitivity. In clinical practice, not every TB suspect is able to take sputum smear test. In our research group, only 56 of 122 TB patients could run sputum smear test and 48 turned positive. In that case, the omission diagnostic rate was actually as high as 60.7% (74/122). In combination of smear test and peptide serology test, the detective rate could improve to 73.0% (Table 5).

In this study, we discover two derived peptides from one protein, which has a diagnostic potential in TB serology test. However, these peptides are not good enough to take place microscopy examine. The present studies provide an advanced strategy of screening TB diagnostic peptide. Integrating multiple diagnostic peptides would improve the existing performance of TB serologic test. Plenty of work had been done on using two or more markers in the form of recombination antigen or antibody groups or others to improve diagnostic efficacy [8, 16, 20, 21]. A research using ESAT-6, CFP10 and PPE68 fusion protein in diagnosing TB could increase the sensitivity to 73.3% [22]. Chemical synthesis of peptides would be a better strategy to study TB diagnostic panel. Peptide array has already revealed application prospect in other disease area. Printed peptide microarray which covering known tumor-associated antigens could predict prognostic in glioblastoma [23]. The array consists of random sequence peptides was capable to describe serum immunosignature pattern and might have potential as a diagnostic tool in Alzheimer’s disease [24]. High-density peptide array may have advanced prospect for comprehensive health monitoring [25]. In our study, we compared the peptide array as a diagnostic tool with the ELISA method. When the cutoff value was set at 4.48 and 3.0 for each pep-LppZ marker, the sensitivity and specificity of peptide array was parallel to ELISA. When we strict the cutoff of peptide arrays to 1.29 and 0.38 for each pep-LppZ marker, the sensitivity of peptide array had a remarkable improve to 75% with not much sacrifice of specificity (Additional file 1: Table S4).

Our study provides the proof that selected peptides can replace the target proteins in immunodiagnostic test for TB. Peptide array can be a superior tool for screening autoantibody based TB biomarker. High-throughput array technique could perform epitope mapping on a large proteome in an economic and effective way. In this study, two rounds of screening performed well and resulted in satisfactory verification rate. We recommend applying a small amount of sample in the second round of screening. It’s the key point of improving verification rate.