Trehalose-6-phosphate Phosphatase from Mycobacterium tuberculosis induces humoral and cellular immune responses


  • Editor: Patrick Brennan

Correspondence: Honghai Wang, Institute of Genetics, Fudan University, 220 Handan Rd, Shanghai 200433, China. Tel.: +86 21 65643777; fax: +86 21 65648376; e-mail:


Trehalose-6-phosphate phosphatase is an enzyme strictly essential for the growth of mycobacteria. Subcellular fractionation of Mycobacterium tuberculosis and M. bovis bacillus Calmette-Guérin (BCG) located the trehalose-6-phosphate phosphatase in the cell wall and membrane fractions. Trehalose-6-phosphate phosphatase induced an increased Th1-type immune response in mice, characterized by an elevated level of interferon-γ in antigen-stimulated splenocyte culture and a strong IgG2a antibody response. The trehalose-6-phosphate phosphatase was recognized by the sera of tuberculosis patients and BCG-vaccinated donors. The mycobacterial trehalose-6-phosphate phosphatase is an immunodominant antigen, and it may be a candidate for vaccine development for the control of tuberculosis.


Tuberculosis represents one of the world's greatest sources of mortality and morbidity, with approximately eight million new infections and 2.5–3 million deaths per year (Dye et al., 1999). The only TB vaccine currently available is an attenuated strain of Mycobacterium bovis termed bacillus Calmette-Guérin (BCG). The efficacy of BCG varied from 0 to 85% in different studies, and the development of an improved vaccine is urgently needed to counter the global threat of this disease (Colditz et al., 1994). An effective vaccine requires the ability to elicit protective immune response. The Th1-type immune response is believed to be necessary for protection against mycobacterial pathogens, such as Mycobacterium tuberculosis and M. bovis (Hovav et al., 2003). The finding of a new immunodominant antigen may contribute to the development of improved vaccine.

It has been shown that trehalose can protect proteins and cellular membranes from inactivation or denaturation caused by a variety of stress conditions, including desiccation, heat, cold and oxidation (Elbein et al., 2003). Furthermore, trehalose is an integral component of a number of different glycolipids in the pathogenic mycobacteria, and these compounds appear to be essential cell wall structures (Brennan & Nikaido, 1995; Woodruff et al., 2004). Murphy et al. recently reported that the OtsA-OtsB pathway for trehalose biosynthesis was the dominant pathway required for the growth of M. tuberculosis in laboratory culture and for the virulence of M. tuberculosis in the mouse model (Murphy et al., 2005). The OtsA-OtsB pathway for trehalose biosynthesis involves condensation of glucose-6-phosphate with UDP-glucose to form trehalose-6-phosphate (catalyzed by trehalose-6-phosphate synthase, OtsA), with subsequent dephosphorylation generating the free disaccharide [catalyzed by (TPP), OtsB] (De Smet et al., 2000). There are two ORFs, otsB1 (Rv2006) and otsB2 (Rv3372), encoding putative TPP in the genome of M. tuberculosis. Edavana et al. proved that the OtsB2 protein was a functional TPP, whereas the OtsB1 protein had no phosphatase activity (Edavana et al., 2004). Of the two otsB homologues, the otsB1 knockout mutant showed no obvious growth impairment, while the otsB2 deletion mutation is lethal to M. tuberculosis growth. The otsB2 is strictly essential for this pathogen growth and provides a tractable target for high-throughput screening (Murphy et al., 2005; Takayama et al., 2005).

To further reveal the biological function of TPP (encoded by otsB2), a cell fractionation experiment was performed to delineate the location of TPP within mycobacteria. TPP was found to be located in the cell wall and cell membrane. The immune responses induced by TPP were characterized in mice and TB patients.

Materials and methods

Bacterial strains and culture conditions

The Escherichia coli strains DH5α and BL21(DE3) were used for cloning and expression, respectively. Escherichia coli strains were cultured in L broth and on agar supplemented with 25 μg mL−1 kanamycin. Mycobacterium tuberculosis and BCG were grown in modified sauton's medium and on Middlebrook 7H10 agar supplemented with 10% oleic acid-albumin-dextrose complex.

Expression, purification of recombinant TPP and rabbit-antisera production

The gene of TPP (otsB2) was amplified from the M. tuberculosis H37Rv genome and ligated into pET28a vector. The recombinant plasmid was transformed into E. coli BL21(DE3). TPP was expressed under the control of T7 promoter and purified by Ni-NTA affinity chromatography. Procedures for antiserum production were based on the standard protocols (Sambrook et al., 1989). Recombinant protein TPP used for antiserum production was mixed with incomplete Freund adjuvant and given to immunized 3-month-old New Zealand White rabbits, followed by four boosts. Antisera were collected at the seventh day after the last boost.

Subcellular fractionation

Cultures of M. tuberculosis H37Rv and BCG were harvested in the early mid log growth phase and the supernatants were sterile filtered (0.22-μm cellulose acetate filter membrane). Filtrates were concentrated 50-fold in a lyophilizer and dialyzed against PBS. The concentrate was saved as the culture filtrate fraction. Pelleted cells of mycobacteria were washed, resuspended in PBS with protease inhibitor cocktail, and subjected to five cycles of sonication. Lysates were centrifuged twice at 11 000 g for 5 min at room temperature to remove unbroken cells and insoluble material. Lysate supernatants were centrifuged at 27 000 g for 1 h at 4°C to precipitate cell walls (Hirschfield et al., 1990). The supernatants were collected and ultracentrifuged at 100 000 g for 4 h at 4°C to separate cell membranes (pellet) from the cytoplasmic fractions (supernatant) (Lee et al., 1992). The cell wall and cell membrane fractions were washed three times with PBS. The cytoplasmic fraction was ultracentrifuged three more times for 4 h at 100 000 g to ensure that all residual membranes were removed. Fractions were checked for fidelity with antibodies recognizing the cytoplasmic 65 kDa protein (Andersen et al., 1986; Young et al., 1992), the 38 kDa membrane protein (Young & Garbe, 1991) and Ag85, which has been shown to be secreted and associated with the cell wall (Rambukkana et al., 1992; Wiker & Harboe, 1992; Florio et al., 1997).

Western blot analysis

All subcellular fractions were resuspended in phosphatebuffered saline (PBS) and an amount of protein equivalent to 500 μL culture was electrophoresed per lane of an sodium dodecyl sulphate (SDS)-polyacrylamide gel (10% acrylamide), electrotransferred to PVDF membrane, and subjected to detection with rabbit antiserum followed by a goat anti-rabbit IgG antibody (Amersham).

Animals and immunization

Experimentation strictly adhered to the 1986 Scientific Procedures Act. Female C57BL/6 mice (25 g, 6 week old) were divided into four groups and used throughout for in vivo studies. Mice received free access to food and water throughout the study. The harvested BCG was resuspended in PBS to an approximate concentration of 200 mg mL−1. One hundred microliter (plate count of 5 × 106 CFU) were injected subcutaneously into the mice of the BCG-immunization group (four mice), and the mice injected with 100 μL PBS were chosen as a control group. The mice of the antigen TPP-immunization group (four mice) were injected subcutaneously three times, 1 week apart, with 50 μg of TPP protein mixed with incomplete Freund adjuvant (IFA), and the mice injected with IFA were chosen as a control group.

Antibody subclass ELISA

For the BCG and PBS immunized mice, sera were collected 3 weeks postimmunization. For the TPP and IFA immunized mice, sera were collected 1 week after the last injection. All samples were tested in enzyme-linked immunosorbent assay (ELISA) designed for the identification of TPP-specific IgG subclasses, and the responses were compared with responses using the conserved mycobacterial antigen 85B (Ag85B) which was purified in our lab. ELISA plates (Nunc Maxisorp CertiWed, Xat bottom) were coated with an optimal concentration of 5 μg mL−1 purified protein in 100 μL of coating buffer (15 mM Na2CO3 and 35 mM NaHCO3, pH 9.6). Nonspecific binding sites were blocked with 1% bovine serum albumin (BSA) in PBST (PBS with 0.05% Tween 20). Plates were washed three times with 350 μL well−1 of PBST, and serum samples were diluted to 1 : 1250 with PBST (containing 1% BSA) and applied to plates in two-fold serial dilutions. Plates were incubated at 37°C for 1 h. After washing five times with PBST the plates were incubated (37°C, 60 min) with peroxidase conjugated anti-mouse IgG1 or IgG2a (diluted 1 : 1000) in PBST (containing 1% BSA). Colour was developed with orthophenylenediamine hydrochloride substrate (100 μL well−1) and the reaction was stopped by adding 7% H2S04 (50 μL well−1). Optical densities were read at 492 nm in an ELISA reader.


The number of IFN-γ-secreting, antigen specific T cells in fresh splenocyte preparations was determined by an enzyme-linked immunospot assay (ELISPOT) method. Briefly, spleens were aseptically removed and gently grinded through a 70-μm cell strainer, then single-cell suspensions were prepared with Lympholyte-M density gradient centrifugation (Cedar lane Lab. Hornby, Ontario, Canada). IFN-γ ELISPOT kits (U-Cytech BV, Utrecht, the Netherlands) were used according to the instruction manual. In brief, 96-well PVDF plates were coated with 50-μL anti-IFN-γmAb overnight at 4°C. The plates were then washed five times with PBST, and then blocked with 200 μL of blocking solution (PBS containing 10% calf serum) for 1 h at 37°C. The blocking solution was then discarded from the plates, and freshly isolated spleen cells were plated in duplicate at 5 × 105 cell per well in 100 μL of RPMI 1640 supplemented with 10% newborn calf serum (GIBICO, New Zealand), then the spleen cells were stimulated with TPP (1, 2, 5 μg mL−1) or Ag85B (2 μg mL−1) for 48 h at 37°C, 5% CO2. The cells were then removed by shaking them off the plates, and PBS was added to lyse the remaining cells. Next, biotin-labeled detector antibody solution was added. After incubation at 37°C for 1 h, the plates were washed five times with PBST, and then 100 μL well−1 of 3-amino-9-ethylcarbazole was added. The plates were once again incubated for 30 min at room temperature. Once spots could be seen under the inverted microscope, the plates were washed with distilled water, then air dried, and the spots were counted.

Study population and ELISA

Human serum samples used in the study were from 79 smear and culture positive patients with pulmonary and extrapulmonary tuberculosis before treatment, from 28 individuals vaccinated with BCG and from 30 non-BCG-vaccinated healthy controls. IgG levels against the antigen TPP were estimated by ELISA described above. The human sera were added at 1 : 100 dilutions, as determined in prior standardization experiments. The mean OD and SD of the non-BCG-vaccinated healthy controls were calculated. A mean OD of ±2 SD of non-BCG-vaccinated healthy controls was considered as the cut-off value. Any sample exhibiting absorbance above the cut-off value was considered as positive.


Subcellular localization of TPP in M. tuberculosis and BCG

A blast search against the available microbial genomes revealed that TPP was widely distributed in mycobacteria species, e.g. M. bovis AF2122/97 (100% identity), M. tuberculosis DC1551 (100% identity), Mycobacterium avium ssp. paratuberculosis K-10 (71% identity) and Mycobacterium leprae TN (67% identity). The subcellular localization of TPP recognized by rabbit antiserum was investigated by Western blot analysis of four different subcellular fractions: culture filtrate, cytosol-, cell membrane- and cell wall-enriched. The various preparations were resolved by PAGE and transferred on to PVDF membrane. The fidelity of the fractions was checked (Fig. 1a). As shown in Fig. 1b, the 42 kDa TPP was present in the cell wall and cell membrane fractions in M. tuberculosis H37Rv and BCG.

Figure 1.

 Subcellular localization of TPP in Mycobacterium tuberculosis H37Rv and BCG. (a) Sonicated preparations of M. tuberculosis H37Rv were fractionated into culture filtrate (CF), cell wall (CW), cell membrane (CM) and cytosol fractions (CY). These fractions were subjected to SDS-PAGE, electroblotted on to PVDF membrane and reacted with the antibodies specific for Ag85B (a), the 38 kDa membrane protein (b) and cytoplasmic 65 kDa protein (c). (b) The subcellular fractions from M. tuberculosis H37Rv (Line 1–4) and BCG (Line 5–8) were Western blotted to rabbit serum raised against recombinant TPP protein.

Distribution of TPP among M. tuberculosis strains

The genome sequence of M. tuberculosis H37Rv suggests the presence of a second gene encoding trehalose-6-phosphate phosphatase called otsB1 (Sanger website; Edavana et al. proved that the OtsB1 protein had no phosphatase activity (Edavana et al., 2004). The antibody prepared against the 42 kDa TPP (encoded by otsB2, Rv3372) did not cross-react with the 146 kDa OtsB1 (encoded by Rv2006). Some enzymes with a second or third homologous protein of M. tuberculosis may be pseudogenes or split ORFworks (Smith et al., 2004). We next looked for the presence of the 42 kDa TPP in different clinical isolates of M. tuberculosis (Fig. 2). The four strains of M. tuberculosis were epidemiologic independent examined by DNA fingerprinting in our lab previously (Zhang et al., 2005). Western blot analysis revealed that intact TPP was widely distributed among clinical strains of M. tuberculosis.

Figure 2.

 Distribution of TPP in mycobacteria. Rabbit serum raised to recombinant TPP protein was immunoblotted to sonicated antigen preparations obtained from BCG (1), avirulent strain of Mycobacterium tuberculosis H37Ra (2), M. tuberculosis H37Rv (3) and four clinical M. tuberculosis isolates (4–7).

TPP induced specific humoral response in mice

Figure 3 illustrates the IgG1 and IgG2a antibody titers in the different groups of mice. High titers of antibody were detected in the sera of mice immunized with the antigen TPP, and IgG2a titers were higher than the IgG1 titers, whereas for the BCG-immunized group the titers of IgG1 and IgG2a were low. Ag85B, a conserved secreted mycobacterial antigen which could induce effective humoral and cellular responses (Dhar et al., 2004), was included in the ELISA test. The titers of the Ag85B specific IgG1 and IgG2a antibodies were higher than TPP. Mice that were immunized with PBS or IFA showed no detectable titer of IgG1 and IgG2a antibodies.

Figure 3.

 Analysis of IgG1 and IgG2a antibody responses against purified TPP. (a, b) Mice were immunized with TPP emulsified with IFA (circles) or IFA(squares). Sera were assayed by ELISA using secondary antibodies specific for mouse IgG1 and IgG2a. (c, d) Mice were immunized with BCG (circles) or PBS (triangles). Levels of IgG1 and IgG2a antibodies against purified TPP (solid lines) or Ag85B (dashed lines) were determined in sera of mice 3 weeks postimmunization.

TPP induced specific cellular response in mice

The IgG2a isotype is associated with a Th1-type cytokine response during which IFN-γ is produced. This corresponded to the finding that splenocytes from mice vaccinated with TPP or BCG produced IFN-γ when stimulated in vitro with protein TPP (Fig. 4). IFN-γ production in response to purified TPP (1, 2 and 5 μg mL−1 in 4a, b and c, respectively) was increased three- to seven-fold in cultures of spleen cells from C57BL/6 mice immunized with protein TPP, and 20 –24-fold higher in mice vaccinated with BCG. The number of splenocytes from BCG-vaccinated mice producing IFN-γ when stimulated in vitro with 2 μg mL−1 TPP was 60% of that when stimulated with 2 μg mL−1 culture filtrates Ag85B. The antigen TPP located in the cell wall and cell membrane of BCG could induce IFN-γ, which was a typical character of Th1-type immune response.

Figure 4.

 The cellular immune response was measured in an ELISPOT assay with splenocytes isolated from C57BL/6 mice immunized with protein TPP (a) or BCG (b). Splenocytes (5 × 105 cell per well) were stimulated with 1, 2 and 5 μg mL−1 purified TPP, respectively. For the BCG-immunized mice the splenocytes were also stimulated with 2 μg mL−1 Ag85B.

Detection of antibody to TPP in sera from tuberculosis patients and BCG-vaccinated donors

The human humoral response to TPP was characterized by measuring serum IgG antibodies to the protein using ELISA. The mean OD±2 SD of non-BCG-vaccinated healthy controls was set as the cut-off value. With this criterion, the data were analyzed. As shown in Table 1, from the 30 control sera, one was positive, thereby giving 97% specificity. Thirty-three out of 79 TB sera and 10 out of 28 BCG-vaccinated individuals' sera were above the cut-off value, offering a test with 42% and 36% sensitivity, respectively. The sensitivity to pulmonary TB sera (47%) was higher than that to extrapulmonary TB sera (34%).

Table 1.   ELISA positivity for TB, BCG-vaccinated individuals, and non-BCG-vaccinated controls against antigen TPP
of positive
% of
  1. BCG, bacillus Calmette-Guérin; TPP, trehalose-6-phosphate phosphatase; ELISA, enzyme-linked immunosorbent assay.

Pulmonary TB470.70520.20742247
Extrapulmonary TB320.69450.18201134
BCG vaccinated individuals280.52170.08431036
Non-BCG vaccinated control300.12730.024813


In this study we analyzed the immunogenicity of the 42 kDa mycobacterial antigen, TPP. The DNA sequences of M. tuberculosis and BCG otsB2 genes are identical. The otsB2 gene was cloned and expressed for this study, and the purified protein TPP was enzymatic as reported previously (Edavana et al., 2004). Subcellular fractionation of M. tuberculosis and BCG located the 42 kDa TPP in the cell wall and membrane fractions. Our lab has previously found that TPP was more effectively induced in isoniazid-resistant strains of M. tuberculosis than it was in isoniazid-sensitive strains. TPP was a potential target for chemotherapy against M. tuberculosis and new marker in the detection of resistance of M. tuberculosis (Yue et al., 2004).

The induction of Th1/Th2 responses in mice has also been determined by analyzing the IgG1/IgG2a isotype profile of the antibody response and by measuring the levels of IFN-γ in spleen cell culture. In mice, Th1 responses are characterized by strong cell mediated immunity and an IgG2a antibody response, whereas Th2 responses are usually associated with high titer IgG1 and IgE antibody responses (Mosmann & Coffman, 1989). In mice, the immunization with the 42 kDa antigen TPP elicited specific helper responses with a Th1-like phenotype, characterized by higher titers of IgG2a than of IgG1 (at a serum dilution of 1/1,250, the ratio was 1.46 : 1) and splenocytes that secrete elevated levels of IFN-γ upon antigen stimulation.

For the BCG-immunized mice TPP failed to induce effective humoral response, in contrast to the recognition of TPP by IgG antibodies in the sera of BCG-vaccinated human donors, which may be contributed to by the different hosts and immune strategies. The detection of TPP- specific antibodies in human tuberculosis patients reinforced the importance of the antigen TPP. The antigen that induces humoral responses in TB patients and BCG-vaccinated donors may serve as candidate for vaccine development, such as Ag 85B (Rambukkana et al., 1992; Wiker & Harboe, 1992).

For the BCG-immunized mice the titers of Ag85B specific IgG1 and IgG2a antibodies and the number of splenocytes that secrete IFN-γ upon antigen stimulation were higher than that of TPP. Ag85B, a secreted mycobacterial antigen, could induce effective humoral and cellular responses. TPP was subcellular, localized in the cell wall and cell membrane of BCG. Bioinformatics analysis revealed that there was no recognizable N-terminal secretion sequence for TPP. For the BCG-immunized mice TPP failed to induce an effective humoral response, in contrast to the strong cellular response. The search for the effective epitopes on TPP and the construction of recombinant BCG that secrete the antigen TPP was under study for vaccine development.

There is a crucial requirement for cellular immune response to control the mycobacterium infection (Flynn et al., 1992; Cooper & Flynn, 1995; Andersen, 1997). It has been shown that innate and cell-mediated immune responses can act as a double-edged sword. On one hand, an appropriate immune response can eradicate and control the pathogens; on the other hand, the same effectors might enhance the pathogens' survival, damage the host tissues, and cause death (Rook & Hernandez-Pando, 1996; Reina-San-Martin et al., 2000). Thus, it is essential to define whether the mycobacterial antigen TPP that influences the immunological balance has positive effect for protection. Our further research was carried out to test the ability of the 42  kDa TPP to provide protection against M. tuberculosis or BCG challenges.


This work was supported by the National Basic Research Program of China (973 Program, No. 2002CB512804 and 2005CB523102) and the National High Technology Research and Development Program of China (863 program, No.2004AA215202).