INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Tuberculosis (TB) is a chronic pulmonary disease caused by infection with Mycobacterium tuberculosis. It is a major disease in developing countries as well as an increasing problem in many developed areas of the world, with about 8 million new cases and 3 million deaths each year (21). A resurgence of TB, largely due to the emergence of drug-resistant strains of M. tuberculosis (28) and the increased risk for TB in human immunodeficiency virus (HIV)-infected persons (9, 14, 29), has magnified the need for rapid, inexpensive, and accurate methods for the diagnosis of TB.

The most common immunologic method used for the diagnosis of M. tuberculosis infection is the purified protein derivative (PPD) (tuberculin) skin test. Although this test is used throughout the world, it is not optimal in terms of either sensitivity or specificity. Individuals vaccinated with bacillus Calmette-Guérin to prevent TB may show a false-positive PPD response. Furthermore, TB is a frequent occurrence in AIDS patients, and the sensitivity of the tuberculin skin test is substantially reduced during HIV infection (4, 13, 16). Direct detection of acid-fast bacilli in sputum can also be accomplished by bacterial staining, culturing, or PCR. Drawbacks to this approach include difficulty in obtaining sputum from children as well as the overall low sensitivity rate, particularly for extrapulmonary TB (12). An alternative approach for diagnosis involves the detection of serum antibodies.

Serodiagnostic tests based on the presence of antibodies against mycobacterial antigens in sera have been described (reviewed in reference 10). Antigens such as 38-kDa PhoS (1), the 30-kDa antigen (antigen 6, alpha antigen, MPB, or 85B) (27), 16-kDa HSP (31), LAM (18), and A60 (5) have been identified, purified, and tested, with various degrees of success. Diagnostic tests based on the 38-kDa antigen, antibodies to which are associated with severe and recurrent disease (3), have achieved sensitivities of as high as 70 to 80% and 95 to 100% specificities (6). However, this antigen has markedly lower sensitivity in smear-negative populations as well as in individuals infected with HIV (19, 32). Several studies have addressed the problems of detecting M. tuberculosis-specific antibodies in TB patients coinfected with HIV. For example, in a field test in Mexico, an enzyme-linked immunosorbent assay (ELISA) based on the 30-kDa mycobacterial antigen had a sensitivity of 70% in patients with culture-positive or smear-positive pulmonary TB and a specificity of 100% in 125 control donors (26). The same test was evaluated with HIV-positive and -negative patients in Uganda (11). Although the sensitivity and specificity in HIV-negative donors were the same as in the Mexico test, the ELISA was positive for only 28% of 128 sera from HIV-positive donors. Accordingly, there is a need for improved diagnostic methods for detecting TB in HIV-positive individuals.

More recently, the reactivity of high-molecular-weight antigens with patient sera has been described. In particular, an 88-kDa antigen was found to elicit a strong antibody response in patients with TB (23). Most importantly, the 88-kDa antigen, which is present in M. tuberculosis culture filtrate proteins (CFP), also has been described as a surrogate marker for TB in HIV-seropositive individuals, but its peptide sequence has been elusive (22). Monoclonal antibody IT-57 has been known to react with an 88-kDa M. tuberculosis antigen in the high-molecular-weight region of CFP, although the identity of the protein antigen has not been previously determined (20). In this report, we used two-dimensional (2D) gel electrophoresis and immunoblot analysis to identify a protein reactive with IT-57 and then nano-liquid chromatography-electrospray ionization tandem mass spectrometry (nano-LC/MS/MS) to identify and sequence a novel protein, Mtb81. Recombinant Mtb81 was expressed in Escherichia coli and evaluated by an ELISA. Based on serological analysis, Mtb81 appears to be a promising antigen for the serodiagnosis of TB, especially for patients coinfected with HIV.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

M. tuberculosis CFP. CFP from M. tuberculosis strain H37Rv were obtained from Colorado State University.

Antibody reagents. Murine monoclonal antibody IT-57 (20) was obtained from the United Nations Development Program/World Bank/World Health Organization Special Program for Research and Training in Tropical Diseases (Centers for Disease Control and Prevention, Atlanta, Ga.).

Antigens. The 38-kDa antigen was expressed in E. coli by using a hexahistidine tag similar to that used for Mtb81 (see below). The TB lysate was prepared by alternately homogenizing and sonicating three 100-mg ampoules of dessicated M. tuberculosis H37Ru (Difco, Franklin Lakes, N.J.) three times in 25 ml of 10 mM Tris (pH 8) containing 2% Nonidet P-40. The mixture was spun for 10 min at 13,000 × g. The supernatant was used as the TB lysate.

Study population. Blood was drawn from all patients after informed consent was obtained. Serum samples were obtained from individuals who had pulmonary TB alone prior to treatment (culture and/or acid-fast bacillus smear positive; see Results) or who had documented coinfections with HIV, as evidenced by a positive HIV type 1 and 2 antibody ELISA. These included samples from Uganda and South Africa that were obtained from Milton Tam (Program for Appropriate Technology in Health, Seattle, Wash.). Additional HIV⁺ TB⁺ serum samples from South Africa and samples from patients with lung cancer and pneumonia (China) were obtained from Robert Cole (AMRAD-ICT, French's Forest, Australia). This second group of HIV⁺ TB⁺ samples was selected on the basis of their negativity for the 38-kDa antigen in the AMRAD-ICT rapid test but smear- and/or culture-positive results (6, 34). To further evaluate the specificity of the Mtb81 antigen, we obtained sera from individuals who were PPD positive (>10-mm reaction zone) (culture, clinically, and radiographically negative for TB) and PPD negative (King County Tuberculosis Clinic, Seattle, Wash.). Samples from healthy blood donors (United States) were obtained from Boston Biomedica (West Bridgewater, Mass.), and sera from individuals who were HIV seropositive alone were obtained from Robert Ackridge (Fred Hutchinson Cancer Research Center, Seattle, Wash.).

Reverse-phase fractionation of M. tuberculosis CFP. Proteins were separated by reverse-phase chromatography on either a C₁₈ or a diphenyl column.

(i) C₁₈ fractionation. Approximately 75 mg of CFP was dissolved in water containing 0.1% trifluoroacetic acid (TFA), and the mixture was injected onto a C₁₈ reverse-phase column (22 by 250 mm; The Separations Group, Hesperia, Calif.) by using a preparative liquid chromatograph (Waters, Milford, Mass.). Fractions were eluted with a binary gradient of 0.1% TFA in water (solvent A) and acetonitrile (solvent B) at a flow rate of 10 ml/min. The gradient increased from 0 to 100% solvent B in 60 min, and fractions were collected at 1-min intervals. Fraction 38, which contained reactivity to antibody IT-57, was identified by Western blot analysis as described below.

(ii) Diphenyl fractionation. Approximately 5 mg of CFP was dissolved in water containing 0.5% TFA, and the mixture was injected onto a Vydac diphenyl reverse-phase column (catalog no. 219TP5415; The Separations Group). Fractions were eluted with a binary gradient of 0.5% TFA in water (buffer A) and acetonitrile (buffer B) on an AKTA Explorer 100 separation system (Amersham Pharmacia Biotech, Uppsala, Sweden). The column was equilibrated with 30% buffer B, and a linear gradient was run from 30 to 65% buffer B at 2 ml/min over the course of 30 min. Fractions were collected at 1.5-ml intervals and analyzed by Western blotting with antibody IT-57. Fractions containing a protein recognized by this antibody eluted at about 50% buffer B and were pooled for further analysis.

SDS-PAGE and 2D PAGE of CFP. Individual high-pressure liquid chromatography (HPLC) fractions were separated by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) on 4 to 20% gradient gels (Bio-Rad, Hercules, Calif.) in accordance with the manufacturer's instructions and transferred to nitrocellulose. HPLC fraction 38 was concentrated to approximately 400 µl, and 40 µl was added to 400 µl of rehydration solution containing 8 M urea, 0.5% (wt/wt) 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS), 15 mM dithiothreitol (DTT), and 0.2% (wt/vol) Parmalyte (pH 3 to 10). The solution was placed in a rehydration cassette, and 18-cm pH 3 to 10 Immobiline DryStrips (Pharmacia Biotech, Uppsala, Sweden) were allowed to hydrate overnight. The hydrated strips were rinsed and focused by using a Multiphor II electrophoresis system with an Immobiline DryStrip kit and an EPS 3500 XL power supply (Pharmacia Biotech) according to the following gradient: 0 to 300 V for 5 min, 300 to 3,500 V for 6 h, and thereafter to 80,000 V · h.

Immunoblotting. Proteins subjected to SDS-PAGE or 2D PAGE analysis were transferred to nitrocellulose membranes (Hybond C Extra; Amersham Corp., Arlington Heights, Ill.) and blocked with 0.5 M NaCl in phosphate-buffered saline (PBS) containing 0.05% Tween (PBST). Blots were washed with PBS and probed with antibody IT-57 at a 1:50 or 1:70 dilution of culture supernatant in 0.5 M NaCl in PBST for the 1D or 2D gels, respectively. After overnight incubation, blots were washed and probed with horseradish peroxidase-conjugated donkey anti-mouse immunoglobulin G (Jackson Immuno Research, West Grove, Pa.). Western blots were developed by use of a Pierce Super Signal enhanced chemiluminescence protocol (Pierce, Rockford, Ill.).

Silver stain protocol. Polyacrylamide and 2D polyacrylamide gels were silver stained by the method of Blum et al. (2) with slight modifications. After the gels were sufficiently developed, the chemical process was stopped by the addition of 5% acetic acid. Prior to storage at 4°C, the gels were rinsed in nanopure water and placed in 0.1% acetic acid solution.

In situ digests and mass spectrometric analysis. The silver-stained gel pieces were excised and digested in situ (30). An aliquot of the tryptic peptides was loaded onto a C₁₈ microcapillary column (75 µm [inner diameter] by 12 cm) and eluted with a binary gradient of acetonitrile and 0.1 M acetic acid, with the concentration of acetonitrile increasing from 0 to 80% over 12 min, into a triple-quadrupole mass spectrometer (TSQ7000; Finnigan MAT, San Jose, Calif.) equipped with an electrospray ionization source. Mass spectra were acquired every 1.5 s over a mass range of 300 to 1,400 atomic mass units. Candidate peptide masses were identified by comparing the tryptic digest to a control digest. Collision-activated dissociation (CAD) mass spectra were recorded for the peptide-free acids. The CAD spectra were interpreted de novo (17) or by use of peptide sequence tags (24). The H37Rv genomic sequence used was from the published literature (7).

Mtb81 cDNA cloning, expression, and purification. Mtb81 cDNA representing the entire open reading frame was generated by PCR from genomic M. tuberculosis DNA using the primers 5'-CTAAGTAGTACTGATCGCGTGTCGGTGGGC-3' and 5'-CAGTGAGAATTCACTAGCGGGCCGCATCGTCAC-3'. Hexahistidine-tagged Mtb81 was obtained using custom PCR primers with the hexahistidine tag and Mtb81 cDNA as a template. The amplified product was subcloned into the pET28 vector system (Novagen, Madison, Wis.), and the sequence was confirmed. His-tagged Mtb81 protein was expressed and purified by standard methodology and affinity purified from inclusion bodies by detergent extraction and Ni²⁺-nitrilotriacetic acid affinity chromatography.

ELISAs. ELISAs were performed with 96-well microtiter plates (Corning Costar, Cambridge, Mass.) coated with Mtb81, 38-kDa antigen (200 ng/well), or TB lysate (100 ng/well). Coating was done overnight at 4°C. Plates were then aspirated and blocked with PBS containing 1% (wt/vol) bovine serum albumin for 2 h at room temperature, followed by a wash with PBS containing 0.1% Tween 20 (PBST2). Serum (diluted 1/25 in PBST2 for Mtb81 and 1/100 for the 38-kDa antigen and the lysate) was added to the wells and incubated for 30 min at room temperature. Following incubation, the wells were washed six times with PBST2 and incubated with protein A-horseradish peroxidase conjugate (Sigma Chemical Co., St. Louis, Mo.) at a 1/20,000 dilution for 30 min. Plates were then washed six times with PBST2 and incubated with tetramethylbenzidine substrate (Kirkegaard & Perry Laboratories, Inc., Gaithersburg, Md.) for a further 15 min. The reaction was stopped by the addition of 1 N sulfuric acid, and plates were read at 450 nm with an ELISA plate reader (Biotek, Hyland Park, Va.). The cutoff for the assays was the mean of the negative population plus three standard deviations of the mean.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Antibody IT-57 recognizes a single HPLC fraction of CFP. To identify a diagnostic Mtb antigen, CFP were separated by reverse-phase chromatography on a C₁₈ column. Each fraction was individually tested by immunoblot analysis to identify fractions containing reactivity against IT-57. Fraction 38 reacted with IT-57; all other fractions were negative (Fig. 1). To simplify this mixture of proteins, fraction 38 was further analyzed using 2D Western blot analysis. This methodology combines the resolving power of 2D gel electrophoresis, which can separate proteins to within 1 kDa in the molecular mass dimension and a pI of 0.1 in the isoelectric dimension, with the specificity of immunoblot analysis. Because ultra-high-sensitivity protein sequencing can be achieved from silver-stained gels (33), we analyzed fraction 38 by 2D SDS-PAGE in duplicate. One gel was transferred to nitrocellulose, immunoblotted with antibody IT-57, and developed by chemiluminescence, and the other gel was silver stained for subsequent mass spectrometric analysis. Primary antibody binding conditions were achieved using a high-salt and detergent blocking protocol. Because of the blocking conditions used, we used gold to stain the membrane for a total protein stain to allow an accurate correlation between the immunoblot and the silver-stained 2D gel. Gold stain has a sensitivity similar to that of silver stain. The spot of reactivity on the immunoblot that corresponds to the visible spot on the silver-stained gel is shown in Fig. 2.

View larger version (30K): [in this window] [in a new window]   FIG. 1.   Immunoblot analysis of reverse-phase fractionated M. tuberculosis CFP. CFP were separated on a C18 column, and fractions were collected at 1-min intervals. Pooled (A) and individual (B) aliquots were separated by SDS-PAGE and, after being transferred to a membrane, immunoblotted with antibody IT-57. Numbers to the left of lanes are kilodaltons.

View larger version (51K): [in this window] [in a new window]   FIG. 2.   2D Western blot analysis of HPLC fraction 38. Fraction 38 was separated by 2D gel electrophoresis in duplicate. One gel was silver stained (A), and the other was transferred to nitrocellulose and immunoblotted with antibody IT-57 (B). Data from the two gels were correlated, and the 81-kDa spot of reactivity in the immunoblot is labeled in the corresponding silver-stained gel (arrow labeled "1"). Numbers to left of panels are kilodaltons.

Identification of Mtb81. To determine the identity of the protein in spot 1 (Fig. 2), the gel piece was excised and digested in situ with trypsin (30), and the total peptide mixture was extracted, concentrated, and analyzed by nano-LC/MS. By comparing the tryptic digest of spot 1 to that of a control tryptic digest, we identified two peptides, with (M + 2H)²⁺ ions at m/z of 444 and 531, which were present in spot 1 but absent in the control.

View larger version (26K): [in this window] [in a new window]   FIG. 3.   Identification of proteins in spot 1 by nano-LC/MS and nano-LC/MS/MS analyses. The silver-stained gel spot 1, shown in Fig. 2, from the 2D Western blot analysis of HPLC fraction 38 was excised, digested in situ with trypsin, and analyzed by nano-LC-MS. Ion current chromatograms for the candidate peptides at m/z 531 and 444 as well as the total ion current chromatogram are shown in panels A, B, and C, respectively. CAD mass spectra of (M + 2H)2+ ions at m/z 531 and 444 are shown in panels D and E, respectively. Ions observed in each spectrum are underlined. Deduced amino acid sequences are shown, where X refers to Leu or Ile, which were not differentiated in this experiment.

Serological evaluation. To evaluate Mtb81 as a diagnostic antigen, the purified recombinant protein was tested in an ELISA with sera from individuals from Uganda and South Africa with pulmonary TB and coinfected with HIV, as determined by an HIV type 1 and 2 antibody ELISA. Also tested were sera from the same regions but from individuals with only pulmonary TB. The results showed Mtb81 antibody reactivity in sera from 25 of 27 HIV-positive, TB-positive patients (92%) tested, compared to 38 of 67 (56%) for patients with TB alone (Fig. 4 and Table 2). The reactivity did not appear to be a function of HIV seropositivity, since it was not detected in any U.S. HIV-positive, TB-negative individuals tested (0 of 11). In contrast, the 38-kDa antigen showed reactivity with sera from 18 of 27 patients (67%) in this same group (HIV and TB positive), with ELISA signals barely above the cutoff, and 27 of 67 patients (40.3%) with TB alone, with higher ELISA signals (Table 2 and Fig. 4). To further evaluate the specificity of Mtb81, we also tested sera from individuals who had class II TB, who were PPD positive (>10 mm), but who had no evidence of active disease (culture, clinically, and radiographically negative) and sera from individuals who were PPD negative and had no history of exposure to TB (class 0) or who had a medical history of TB exposure but were PPD negative (class I) (Table 2 and Fig. 4). Sera from healthy blood donors were also tested. Sera from 2 of 57 PPD-positive individuals were positive in the Mtb81 ELISA, whereas sera from 0 of 29 PPD-negative individuals were reactive. No Mtb81 ELISA-positive results were seen for the 45 healthy individuals, indicating that Mtb81 may have a high positive predictive value for TB. These results demonstrate the presence of an antibody response against Mtb81 in sera from patients with active TB both in the presence and in the absence of HIV infection; this response was not detectable in sera from PPD-negative individuals and was detected in sera from only 2 of 57 individuals with PPD-positive responses.

View larger version (14K): [in this window] [in a new window]   FIG. 4.   Distribution of seroreactivity against Mtb81 among sera from healthy donors (n = 45) (), PPD-negative individuals with no medical history of exposure to TB (class 0; n = 11) and PPD-negative individuals with a medical history of exposure to TB (class I; n = 18) (), PPD-positive individuals who were culture negative and had no clinical or radiographic evidence of TB (class II; n = 57) (), HIV-seropositive individuals with no evidence of TB (n = 11) (×), HIV- and TB-positive individuals (n = 27) (+), and TB-positive, HIV-negative individuals (n = 67) (). OD 450, optical density at 450 nm.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

This study combines 2D gel electrophoresis and immunoblot analysis, together with tandem mass spectrometric sequencing, to identify a serological marker for TB. Advances in mass spectrometry have made it uniquely suited for the sequencing of proteins. The sensitivity is 10,000 times higher than that of N-terminal Edman analysis, with sensitivity of sequencing by tandem mass spectrometric data at the low-femtomole to low-attomole level (8, 25). This is an essential criterion for protein identification in situations in which very limited quantities of material are available. In addition to high sensitivity, tandem mass spectrometric techniques are unaffected by a blocked N terminus, can directly identify posttranslational modifications, and can be used to identify individual components in a complex mixture without the need for extensive biochemical purification. Therefore, this approach is well suited to antigen identification.

Knowledge of serological markers for TB patients coinfected with HIV would facilitate the development of diagnostic tests for TB. A study designed to characterize the target of the antibody response revealed that an 88-kDa antigen from CFP reacted with 70% of sera from TB patients (23). This same 88-kDa antigen was able to detect antibody in 74% of sera from HIV-positive patients who later developed TB (22). The authors suggested that the 88-kDa antigen, which was also recognized by antibody IT-57, can serve as a surrogate marker for identifying HIV-infected persons with subclinical TB. Because Mtb81 described in this report does not react with IT-57, is not clear whether these antigens are indeed the same. However, since IT-57 has been shown to react with the protein KatG and KatG knockout strains of M. tuberculosis lose IT-57 reactivity yet still show reactivity with patient sera (J. T. Belisle, personal communication), it is possible that IT-57 reactivity does not correlate with patient serum reactivity. It is interesting to note that in this study, even after reverse-phase fractionation and 2D gel electrophoresis, Mtb81 and KatG copurified. One method to determine if Mtb81 and the 88-kDa antigen described by others are the same is to preincubate patient sera with Mtb81 prior to analysis with the purified 88-kDa antigen fraction to ascertain whether Mtb81 can block reactivity. Sequence analysis suggests that Mtb81 is a putative malate synthase based on its 61.3% identity in a 724-amino-acid overlap with an E. coli enzyme (7). Further work is necessary to identify the exact nature and function of this protein.

Mtb81 appears to be highly specific for the diagnosis of TB. Sera from non-TB patients with no medical history of exposure to TB and who were PPD negative tested negative in the Mtb81 ELISA. This was also the case for previously exposed individuals who were PPD negative. Sera from 2 of the 57 PPD-positive individuals (culture negative and with no clinical or radiologic evidence of TB) reacted with Mtb81. Sera from both of these individuals, who had a history of TB infection, also reacted with TB lysate and the 38-kDa antigen. The sensitivity of the Mtb81 ELISA does not appear to be significantly impaired in HIV-seropositive individuals. Although in the studies described, Mtb81 demonstrated a higher sensitivity in HIV-seropositive individuals (92%) than in HIV-seronegative individuals with TB (56%), further studies are necessary to confirm this finding with HIV and TB smear-negative individuals and to assess the utility of Mtb81 for extrapulmonary TB. The fact that Mtb81 was able to detect specific antibody in approximately 70% of sera from individuals with HIV-TB coinfections that were previously shown to be seronegative for the 38-kDa antigen indicates the utility of Mtb81 as a complementary antigen in the design of a cocktail of antigens for the serodiagnosis of TB. Some evidence for this notion is shown in Table 2, where a higher sensitivity was seen when Mtb81 and the 38-kDa antigen were combined. These results indicate the need to perform expanded studies with Mtb81 in combination with other novel antigens previously described (15), including the 38-kDa antigen, particularly for HIV-TB coinfections in different geographical regions.

In summary, Mtb81, identified by serological proteome analysis, is a promising antigen for the serodiagnosis of TB. This methodology should be useful in the identification of other pertinent diagnostic markers relevant for other diseases. In addition, the data suggest that the combination of Mtb81 and the 38-kDa antigen or, potentially, other novel M. tuberculosis antigens (15) would lead to optimal sensitivity for the serodiagnosis of TB and potentially improve clinical sensitivity for TB-positive individuals coinfected with HIV.