PPE Protein (Rv3873) from DNA Segment RD1 of Mycobacterium tuberculosis: Strong Recognition of Both Specific T-Cell Epitopes and Epitopes Conserved within the PPE Family

Abstract

Proteins encoded by DNA segment RD1 of Mycobacterium tuberculosis have recently been demonstrated to play important roles in bacterial virulence, vaccine development, and diagnostic reagent design. Previously, we characterized two immunodominant T-cell antigens, the early secreted antigen target (ESAT-6) and the 10-kDa culture filtrate protein (CFP10), which are encoded by the esx-lhp operon in this region. In the present study we characterized a third putative open reading frame in this region, rv3873, which encodes a PPE protein. We found that the rv3873 gene is expressed in M. tuberculosis H37Rv and that the native protein, Rv3873, is predominantly associated with the mycobacterial cell or wall. When tested as a His-tagged recombinant protein, Rv3873 stimulated high levels of gamma interferon secretion in peripheral blood mononuclear cells isolated from tuberculosis (TB) patients, as well as from healthy tuberculin purified protein derivative-positive donors. In contrast to other RD1-encoded antigens, Rv3873 was also found to be recognized by a significant proportion of Mycobacterium bovis BCG-vaccinated donors. Epitope mapping performed with overlapping peptides revealed a broad pattern of T-cell recognition comprising both TB-specific epitopes and epitopes also recognized by BCG-vaccinated donors. The immunodominant epitope (residues 118 to 135) for both TB patients and BCG-vaccinated individuals was found to be highly conserved among a large number of PPE family members.

The Mycobacterium bovis bacillus Calmette-Guérin (BCG) vaccine has undergone a number of genetic and biochemical changes during its propagation over the last 70 years (18). Although the current vaccine is effective for protecting against childhood forms of tuberculosis (TB), it has failed to prevent adult pulmonary manifestations of the disease in countries where TB is highly endemic (8, 17). The question of whether the contemporary BCG strains have been attenuated to impotence has therefore been posed (5).

In attempts to understand the limitations of BCG, workers have exploited modern molecular techniques to dissect the molecular differences between BCG, M. bovis, and Mycobacterium tuberculosis. Subtraction hybridization revealed three DNA segments (RD1, RD2, and RD3) that have been lost from BCG strains (15). More recently, DNA microarrays and bacterial artificial chromosome arrays have provided information about an additional 13 regions (representing 129 open reading frames [ORFs]) of the genome of M. tuberculosis that are lacking in various BCG strains (6, 12). In addition to the importance of these findings for our understanding of the genetic makeup of the current vaccines, the antigens encoded by these ORFs have attracted the attention of researchers in the TB field because they could feasibly be used to both supplement the BCG vaccine or allow the development of specific diagnostic reagents. Several lines of evidence suggest that the proteins encoded by one such DNA segment, the RD1 region, might be of particular importance since this region encodes a number of molecules that are strongly recognized by the host immune system. RD1 is present in virulent and clinical isolates of M. tuberculosis and M. bovis and was the first DNA fragment to be deleted from M. bovis during the attenuation process between 1908 and 1921, implying that it has a role in virulence (6). This hypothesis is supported by recent reports that M. bovis and M. tuberculosis strains which carry knockout mutations in the RD1 region exhibit reduced virulence (14, 31). Furthermore, reintroduction of RD1 into BCG resulted in a significant increase in virulence (21). Eight ORFs were predicted in this region; two of the ORFs were later found to comprise the lhp-exe operon that encodes two previously identified immunodominant T-cell antigens, the early secreted antigen target (ESAT-6) and the 10-kDa culture filtrate protein (CFP10) (3, 7, 27, 29). Next to lhp-exe is a gene pair that encodes a PE protein and a PPE protein. The PE and PPE families are two large, unrelated protein families that occupy approximately 9% of the coding capacity of the M. tuberculosis genome. These proteins have been hypothesized to represent the principal source of antigenic variation in M. tuberculosis (9).

In the present study, the PPE protein encoded by rv3873 in the RD1 region was selected for molecular and immunological characterization. Since previous proteomic studies had not identified the native protein encoded by this ORF in M. tuberculosis (13, 16, 23, 25), the expression of rv3873 was studied by reverse transcription (RT)-PCR and immunodetection. The immune recognition of the encoded protein (Rv3873) and the fine specificity of the response were evaluated by in vitro stimulation of gamma interferon (IFN-γ) production. The finding that Rv3873, like ESAT-6 and CFP10, is also a potent T-cell antigen recognized by M. tuberculosis-infected individuals illustrates the abundance of cellular immune targets in the RD1 region. We also identified an immunodominant T-cell epitope that is highly conserved among different PPE family members.

MATERIALS AND METHODS

RNA isolation and RT-PCR.

M. tuberculosis H37Rv (= ATCC 27294) was cultured in modified Sauton medium; 2 × 10⁶ bacteria per ml were grown for 7 days on an orbital shaker at 37°C, and the bacteria were harvested by centrifugation. The cell pellet was washed once with 0.8% NaCl-0.05% Tween 80, and subsequently the M. tuberculosis cells were disrupted by three cycles of freezing (in ethanol on dry ice) and thawing (at 37°C) in RLT buffer (Qiagen GmbH, Hilden, Germany) containing 2-mercaptoethanol, followed by beating with glass beads (diameter, 0.1 mm) for 5 min with a vortex mixer. The supernatant was collected by centrifugation, and RNA was isolated as described in the RNeasy mini protocol for isolation of total RNA from bacteria (Qiagen). Prior to cDNA synthesis, RNA was treated with RNase-free DNase I (Boehringer Mannheim) at 37°C for 20 min. RT-PCR was performed with the ThermoScript RT-PCR system (Invitrogen Life Technologies) by using rv387-specific primers Rv3873-F (ACCGGAGCTAAATACCGCAC) and Rv3873-R (GAAGTTGGTGGCCGTAAGGA) in accordance with the manufacturer's instructions.

Immunodetection of native Rv3873.

Short-term culture filtrate (STCF), total cell lysate, cytosol, membrane, and cell wall fractions of M. tuberculosis protein were prepared as previously described (23, 25). Standard sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) was performed by using 10 to 20% polyacrylamide gradient gels with the Protean IIxi system (24). Two-dimensional SDS-PAGE and immunodetection were performed as described elsewhere (23, 25). Antiserum K2020 used for identification of Rv3873 was raised by immunizing a rabbit with purified recombinant Rv3873 (rRv3873).

Recombinant proteins.

Rv3873 was cloned into the pQE-32 Escherichia coli high-level expression system (Qiagen) in frame with the N-terminal His tag as described elsewhere (P. Andersen and R. Skjoet, 20 May 1999, international patent application, PCT publication no. WO9924577) and was purified as follows. E. coli cells expressing high levels of rRv3873 were lysed by incubating the cells in lysis buffer (20 mM Tris-HCl [pH 8.0], 0.1 M NaCl, 1 mM EDTA, 0.1% deoxycholic acid, 0.2 mg of lysozyme per ml, 6.5 μg of DNase I per ml, 0.1 mg of Pefabloc SC Plus per ml, 1 μg of E-64 per ml, 1 μg of leupeptin per ml, 5 μg of Phosphoramidon per ml) at room temperature for 30 min with gentle agitation. Inclusion bodies were isolated by centrifugation at 20,000 × g for 20 min with a Sorvall centrifuge at 4°C. The pellet was washed once with 20 mM Tris-HCl (pH 8.0) with 0.5 M NaCl, 10% glycerol, 0.01% Tween 20, and 10 mM β-mercaptoethanol (buffer A) containing 3 M urea and then dissolved in buffer A containing 8 M urea. This solution was loaded onto a metal ion affinity column containing the Talon resin (Clontech), and the bound proteins were eluted stepwise with buffer A containing 5 to 100 mM imidazole in the presence of 8 M urea. Fractions were analyzed by SDS-PAGE, and the remaining fractions containing rRv3873 were pooled and dialyzed against CIEC buffer (20 mM sodium acetate [pH 5.3], 5 mM dithiothreitol, 0.01% Tween 20, 10% glycerol, 3 M urea). The remaining E. coli contaminants were removed by running the protein solution first through a Hitrap SP column (Amersham Biosciences, Uppsala, Sweden) in CIEC buffer and then through a Hitrap Q column (Amersham Biosciences) in a solution containing 20 mM bis-Tris (pH 5.8), 5 mM dithiothreitol, 0.01% Tween 20, 10% glycerol, and 3 M urea. Under these conditions only E. coli proteins bound to the columns and rRv3873 were in the run-through fractions. The purity of the purified rRv3873 was >99%, as determined by SDS-PAGE followed by silver staining and immunoblotting with anti-His (Clontech) and anti-E. coli antibodies to detect contaminants (DAKO, Glostrup, Denmark) (Okkels, unpublished data). The purified rRv3873 was stored at −20°C in 25 mM HEPES (pH 7.5) supplemented with 150 mM NaCl, 10% glycerol, and 0.01% Tween 20.

Synthetic peptides.

Synthetic overlapping 18-mer peptides covering the complete primary structure of Rv3873 either were synthesized by standard solid-phase methods with an ABIMED peptide synthesizer (ABIMED, Langenfeld, Germany) at the Department of Infectious Diseases and Immunohematology/Bloodbank, Leiden University Medical Center, Leiden, The Netherlands, or were purchased from commercial suppliers (Schafer-N, Copenhagen, Denmark, or Mimotopes Pty. Ltd., Clayton, Australia).

Human donors.

The healthy non-BCG-vaccinated donors used were high school students (ages, 16 to 18 years) from Denmark with no prior history of BCG vaccination. The healthy BCG-vaccinated donors were Danish adults who were 20 to 67 years old and had received BCG vaccinations in childhood. Donors in these two groups answered a questionnaire about BCG vaccination, travel history, contact with TB patients, and occupational exposure to mycobacteria. Subsequently, the donors who had known prior exposure or who were considered at risk of prior exposure were excluded. The TB patients (ages, 18 to 56 years) included both Danes and immigrants from Pakistan, Somalia, Venezuela, the Philippines, Vietnam, Bhutan, and Bosnia. All patients had microscopy- or culture-proven infections and were diagnosed as having pulmonary TB, extrapulmonary TB, or pleuritis. Blood samples were taken 0 to 6 months after diagnosis. The healthy natural converter group consisted of elderly Norwegians who had skin test converted in childhood (positive skin test conversion, >10 mm) without developing clinical disease or receiving any treatment. All participating donors gave informed written consent, and the study was approved by the Local Ethical Committee for Copenhagen and Frederiksberg (RH 01-282/96 and KF 01-369/98).

PBMC cultures and IFN-γ assays.

Peripheral blood mononuclear cells (PBMC) were freshly isolated by gradient centrifugation of heparinized blood on Lymphoprep (Nycomed, Oslo, Norway) and were stored in liquid nitrogen until they were used. The conditions used for PBMC culturing and stimulation of IFN-γ release by antigens have been described previously (28). Unless otherwise stated, purified protein derivative (PPD) and the recombinant antigens were used at a concentration of 5 μg/ml, while the synthetic peptides were used at a concentration of 10 μg/ml. Cell cultures to which no antigens were added were included as negative controls, and phytohemagglutinin at a concentration of 1 μg/ml was used as a positive control (results not shown). IFN-γ was detected by a standard sandwich enzyme-linked immunosorbent assay by using a commercially available pair of monoclonal antibodies (Endogen) according to the manufacturer's instructions. Recombinant IFN-γ (Endogen) was used as a standard.

RESULTS

Expression and characterization of M. tuberculosis rv3873.

In order to verify that the putative ORF rv3873 is functionally active, we carried out RT-PCR to detect the transcript. Total RNA was isolated from M. tuberculosis H37Rv grown for 1 week in broth culture. RT-PCR was performed by using a pair of rv3873-specific primers that are 352 bp apart. When the PCR products were analyzed by agarose gel electrophoresis, an amplicon of the expected size was obtained (Fig. 1A). The rv3873 origin of the amplicon was confirmed by nucleotide sequencing (data not shown). No PCR products were detected in the absence of reverse transcriptase (Fig. 1A). These data indicate that rv3873 is a functionally active gene that is expressed during exponential growth in vitro.

FIG. 1.

Expression of Rv3873 in M. tuberculosis. (A) RT-PCR detection of the transcript. Lanes 1 and 2, RT-PCR with and without reverse transcriptase, respectively; lane 3, 100-bp ladder. (B) Immunodetection of native Rv3873 in different subcellular fractions. Lanes 1 to 5, reactions with rabbit serum raised against rRv3873; lanes 6 to 10, reactions with rabbit preimmune serum; lanes 1 and 6, M. tuberculosis STCF; lanes 2 and 7, total cell lysate; lanes 3 and 8, cytosolic fraction; lanes 4 and 9, cell wall fraction; lanes 5 and 10, cell membrane. (C) Silver-stained two-dimensional SDS-PAGE analysis of total cell lysate (150 μg). The position of Rv3873 is circled. (D) Two-dimensional Western blot with anti-Rv3873. (B, C, and D) Molecular mass markers in kilodaltons are indicated on the left. (C and D) pH values are indicated below the panels.

To identify the native protein, the entire coding region of rv3873 was cloned into the pQE-32 E. coli high-level expression system (Qiagen) in frame with the N-terminal His tag. The recombinant protein was purified and used to immunize a rabbit in order to raise an anti-Rv3873 serum (K2020). The reactivity of K2020 was confirmed by Western blotting in which preimmune serum was used as a negative control (data not shown). Proteins from M. tuberculosis H37Rv lysate, STCF, cytosol, membrane, and cell wall fractions were separated by SDS-PAGE and blotted onto a nitrocellulose membrane. Both K2020 and the preimmune serum reacted with a protein band at about 47 kDa. In addition, K2020 recognized a protein band at approximately 39 kDa (Fig. 1B). This protein was mainly present in the lysate and cell wall fractions, and smaller amounts were present in the membrane fraction. The STCF and cytosol fractions did not contain detectable levels of this protein (Fig. 1B). The lysate was analyzed further by two-dimensional Western blotting. The 47-kDa protein, which reacted with both the preimmune serum and K2020 (Fig. 1B), was not detectable on two-dimensional Western blots (Fig. 1D) (the blot with the preimmune serum is not shown). This may have been due to differences in the methods used for protein sample preparation for standard SDS-PAGE and two-dimensional SDS-PAGE, or it could have reflected the fact that the 47-kDa protein had a pI that was not in the pH range used for isoelectric focusing (pH 4 to 7). K2020 reacted with a single spot at a mass of about 39 kDa and a pI of 4.2 (Fig. 1C and D). These values correlate well with the theoretical molecular mass (37.3 kDa) and isoelectric point (pH 4.0) of Rv3873 (http://genolist.pasteur.fr/TubercuList). Thus, we demonstrated that rv3873 encodes a protein that is associated with the hydrophobic environment of the plasma membrane and cell wall.

Human T-cell recognition of Rv3873.

After we demonstrated that the putative ORF rv3873 is indeed expressed in M. tuberculosis, we were interested in investigating whether the encoded Rv3873 protein is recognized by T cells during M. tuberculosis infection in humans. T-cell recognition was assessed by measuring IFN-γ release in PBMC upon in vitro stimulation with purified rRv3873. Ten non-BCG-vaccinated healthy donors all were negative for rRv3873, while PBMC from 10 of 24 TB patients (42%) recognized rRv3873 (Fig. 2). We also investigated the recognition of this molecule by healthy elderly individuals who had skin test converted in their youth but had never developed clinical signs of TB or received treatment (so-called healthy converters). Four of ten donors (40%) in this group responded to rRv3873.

FIG. 2.

IFN-γ responses to rRv3873 in different groups of human donors. PBMC from 10 non-BCG-vaccinated healthy donors (Control), 24 TB patients (TB), 10 healthy natural converters (Converter), and 40 healthy BCG-vaccinated donors (BCG) were stimulated in vitro with PPD or rRv3873. Each symbol represents one donor. The cutoff value for a positive response to rRv3873 (130 pg/ml) was the mean value for 36 antigen-free controls plus 3 standard deviations and is indicated by the dashed line.

Rv3873 is encoded by the RD1 region of the genome which is deleted in BCG and therefore might not be expected to be recognized by BCG-vaccinated donors. To evaluate the TB specificity of T-cell responses to Rv3873, PBMC from 40 BCG-vaccinated healthy donors (BCG group) with no known history of exposure to M. tuberculosis were analyzed for reactivity to rRv3873 and PPD. All of the donors recognized PPD, but interestingly, seven of the donors (17.5%) also reacted to rRv3873, indicating that this molecule is not completely specific for M. tuberculosis. The median amount of IFN-γ released by the Rv3873-reactive donors in the BCG group (203 pg/ml) was much smaller than that of the TB and the converter groups (833 and 1,374 pg/ml, respectively).

Analysis of the specificity of Rv3873 recognition in TB patients and BCG-vaccinated individuals.

One potential explanation for the recognition of rRv3873 by some BCG-vaccinated individuals is that this molecule belongs to the abundant PPE protein family, which contains 68 related members in M. tuberculosis H37Rv, many of which are also present in BCG (9). A search of the genome database of M. bovis (http://www.sanger.ac.uk/Projects/M_bovis/) by using TBLASTN revealed that whereas the C-terminal part of 208 residues of Rv3873 is only weakly related (<23%) to that of other PPE proteins, the N-terminal 160 residues of Rv3873 exhibit levels of sequence homology to several PPE proteins of approximately 50%. Second, by performing BLAST searches of the publicly available mycobacterial databases we identified Rv3873 homologues in other mycobacteria, including Mycobacterium marinum and Mycobacterium smegmatis, which are occasionally found to be causes of human sensitization (4). T-cell reactivity to rRv3873 may therefore result from both TB-specific recognition and cross-reactivity to epitopes shared by Rv3873 and other PPE proteins of BCG and in some cases even other atypical mycobacteria. To investigate this possibility in detail, 35 18-mer overlapping peptides spanning the entire sequence of Rv3873 were synthesized, and their ability to stimulate IFN-γ release was tested with PBMC isolated from seven TB-infected or -exposed individuals (four TB patients and three healthy converters) and 15 BCG-vaccinated donors. A total of 28 peptides were recognized by the TB patients, and 8 of these peptides were especially strongly reactive in PBMC from five or six of the donors; the median value was between 670 and 1,300 pg/ml for the positive responders (Fig. 3A). In comparison, the peptide recognition by BCG-vaccinated donors was at a much lower level. Altogether, 14 peptides were reactive, and the majority of them were recognized by 1 of the 15 donors. However, one peptide, Rv3873_118-135, was recognized by four donors (Fig. 3B). The cross-reactive peptide sequences were blasted against the M. tuberculosis and M. bovis genome databases. The Rv3873_118-135 peptide was found to be highly conserved among the PPE family proteins from both organisms (the highest level of sequence identity was 89%), whereas the rest of the peptides exhibited between 30 and 50% homology (Table 1). Several peptides showed sequence homology when they were blasted against the available genome databases of atypical mycobacteria, and the Rv3873_118-135 peptide was again found to be the most highly conserved (89% sequence identity in M. marinum and M. smegmatis). Interestingly, large portions of Rv3873 were TB specific, and only three of the eight dominant epitopes were recognized by BCG-vaccinated donors (Fig. 3).

FIG. 3.

Specificity of the T-cell response to Rv3873. PBMC from M. tuberculosis-infected individuals (n = 7) (A) and BCG-vaccinated healthy donors (n = 15) (B) were stimulated in vitro with a panel of 35 overlapping peptides (18-mers). Each peptide was designated by the residue numbers that it represented. The cutoff value for a positive response (130 pg/ml) is indicated by a dashed line. The most commonly recognized peptides are indicated by asterisks.

TABLE 1.

Sequence homology of Rv3873 peptides that were recognized by BCG-vaccinated healthy donors

Peptide	Homologous peptide in M. tuberculosis and M. bovisa	Source of homologous peptideb	% Sequence identity
Rv387363-80	KALAAATPMVVWLQTAST	Rv1790 PPE family	50
	:::::::::
	APIAAAKPMITWLQSAAE
Rv387373-90	VWLQTASTQAKTRAMQAT	Rv0442c, PPE family	50
	:::::::::
	AWLSTAAAQAEQAAAQAM
Rv3873118-135	VLTATNFFGINTIPIALT	Rv3021c, PPE family	89
	::::::::::::::::
	VLVATNFFGINTIPIALN
Rv3873137-154	MDYFIRMWNQAALAMEVY	Rv3021c, PPE family	44
	::::::::
	EADYVRMWVQAATVMSAY
Rv3873177-194	GASQSTTNPIFGMPSPGS	Rv0074c, unknown function	39
	:::::::
	GPDDWSANSVFGLPALGS
Rv3873187-204	FGMPSPGSSTPVGQLPPA	Rv3802c, unknown function	44
	::::::::
	RGAESPPSAVPPGVLPPG
Rv3873207-224	QTLGQLGEMSGPMQQLTQ	Rv2083, unknown function	44
	::::::::
	ALSGALGGVMGPLTQLPQ
Rv3873217-234	GPMQQLTQPLQQVTSLFS	Rv1196, PPE family	39
	:::::::
	QALQQLAQPTQGTTPSSK
Rv3873227-244	QQVTSLFSQVGGTGGGNP	Rv1829c, short-chain dehydrogenase	50
	:::::::::
	QQVTSMLDQVTAELGGID
Rv3873247-264	EEAAQMGLLGTSPLSNHP	Rv2297, unknown function	50
	::::::::
	MEMAMMGLLGTVVGASAM
Rv3873257-274	TSPLSNHPLAGGSGPSAG	Rv0704, rp1B	30
	::::::
	VMNPVDHPHGGGEGKTSG
Rv3873267-284	GGSGPSAGAGLLRAESLP	Rv3442c, rpsI	50
	:::::::::
	GGGGPSGQAGALRLGIAR
Rv3873277-294	LLRAESLPGAGGSLTRTP	Rv3406, probable dioxygenase	50
	:::::::::
	VLRAVSLPSYGGSTLWAN
Rv3873340-357	TRPGLVAPAPLAQEREED	Rv2118c, unknown function	50
	::::::::::
	LAPGAVAPAPLGRKREGR

^aThe upper sequence is the Rv3873 peptide sequence; the lower sequence is the M. tuberculosis-M. bovis peptide sequence that has the highest sequence identity. Identical residues are indicated by colons.

^bThe M. bovis and M. tuberculosis peptide sequences are identical.

Highly conserved dominant T-cell epitope Rv3873_118-135 as a source of Rv3873 cross-recognition.

The dominant cross-reactive peptide Rv3873_118-135 is conserved and exhibits 78 to 89% sequence identity with nine other putative PPE proteins common to M. tuberculosis and BCG. It is also well conserved in three PPE proteins of Mycobacterium leprae, as well as in the proteins encoded by seven ORFs of Mycobacterium avium, five ORFs of M. marinum, six ORFs of Mycobacterium ulcerans, and two ORFs of M. smegmatis (Table 2). To investigate the potential role of this peptide in the cross-recognition of Rv3873, we compared the PBMC recognition of Rv3873_118-135 with that of the intact rRv3873 in 32 BCG-vaccinated donors. Five of the seven rRv3873-reactive BCG-vaccinated healthy donors reacted to Rv3873_118-135 at a level similar to the level of reaction to the intact protein (Fig. 4A). These observations indicate that both BCG vaccination and sensitization to some strains of atypical mycobacteria may result in recognition of the dominant epitope Rv3873_118-135 and thereby in a lack of TB specificity of the intact antigen.

TABLE 2.

Sequences homologous to the Rv3873_118-135 sequence in other PPE proteins of mycobacteria

Organism(s)	Protein or region	Sequenceg
M. tuberculosis and M. bovisa	Rv3873118-135	VLTATNFFGINTIPIALT
	Rv0096116-133	VLIATNFFGINTVPIALN
	Rv0256c122-139	VLMATNFFGINTIPIALN
	Rv0280120-137	VLVATNFFGINTIPIALN
	Rv0286122-139	VLVATNFFGINTIPITLN
	Rv0453123-140	AMVATNFFGINTIPIAVN
	Rv1387121-138	VLVATNFFGINTIPIAIN
	Rv2123122-139	ALVTTNFFGVNTIPIALN
	Rv3018c121-138	VLVATNFFGINTIPIALN
	Rv3021c45-62	VLVATNFFGINTIPIALN
M. lepraeb	ML0539121-138	VLMATNFFGVNTIPIALN
	ML1828122-139	VLVTTNFFGINTIPIARN
	ML1991116-133	FLIATNFFGINTIPIALN
M. aviumc	nt115526-115579	ALVATNFFGVNTIPIAVN
	nt1282662-1282613	VLTATNFFGINTIPIAVN
	nt2451127-2451174	ALVATNFFGINTIPIAVN
	nt2547913-2547869	ALVSTNFFGVNTIPIALN
	nt4486820-4486873	ALVATNFFGINTIPIAVN
	nt4996522-4996472	VLLATNFFGINTIPIAVN
	nt5006081-5006040	ALVATNFFGVNTIPIAVN
M. marinumd		VLLATNFFGINTIPIALN
		ALTATNFFGINTVPIAVN
		ALTATNFFGINTVPITLN
		VLLGTNFFGINTIPIALN
		ALVATNFFGLNTIPITLN
M. smegmatise	Contig 3312	VLEATNFLGINTVPIALN
	Contig 3311	VLVATNFFGINTIPIALN
M. ulceransf		VLVATNFFGLNTIPIALN
		VLLATNFFGLNTIPIALN
		VLLGTNFFGLNTIPIALN
		VLLGTNFFGLNTIPIALN
		ALVATNFFGLNTIPITLN
		VLVSTNFFGVNTIPIAVN

^aData from TubercuList (http://genolist.pasteur.fr/TubercuList), and http://www.sanger.ac.uk/Projects/M_bovis/.

^bData from Leproma (http://genolist.pasteur.fr/Leproma).

^cData from http://tigrblast.tigr.org/ufmg/.

^dData from http://www.sanger.ac.uk/Projects/M_marinum/.

^eData from http://tigrblast.tigr.org/ufmg/.

^fData from http://genopole.pasteur.fr/Mulc/BuruList.html.

^gIdentical amino acids are indicated by boldface type; conservative substitutions are indicated by italics; and nonconservative substitutions are underlined.

FIG. 4.

(A) Recognition of the highly conserved peptide (Rv3873_118-135) by BCG-vaccinated donors. PBMC from BCG-vaccinated donors (n = 32) were stimulated with rRv3873 and the peptide epitope shared by members of the PPE family (Rv3873_118-135). Open circles indicate donors that recognized both rRv3873 and the conserved peptide. (B) Recognition of the homologous peptides by PBMC from two Rv3873_118-135-responsive donors. The results for one donor are indicated by solid circles, and the results for the other donor are indicated by open circles.

PBMC from two of the Rv3873_118-135-responsive BCG-vaccinated donors were also tested for recognition of a panel of homologous peptides from PPE family members present in BCG (Table 2 and Fig. 4B). One of the donors recognized several of the homologous peptides at levels similar to the level of Rv3873_118-135, whereas the other donor recognized the homologous peptides from Rv0286, Rv0453, and Rv1387 more strongly (>800 pg of IFN-γ/ml) than Rv3873_118-135 (450 pg/ml). These data confirmed the cross-reactive nature of Rv3873_118-135.

DISCUSSION

The findings of the present work demonstrate that the putative ORF rv3873 encodes a protein that is associated with the cell or wall. Although there are 68 PPE-encoding ORFs in M. tuberculosis, none of them have so far been identified by proteomic studies (13, 16, 23, 25). Several explanations for this have been suggested, as follows: (i) possible low levels of PPE proteins in M. tuberculosis under in vitro growth conditions; (ii) limited sites available for digestion by trypsin, the most commonly used enzyme in mass spectrometric analysis of a proteolytically digested protein; and (iii) use of suboptimal two-dimensional SDS-PAGE conditions for resolution of the PPE proteins. In the case of Rv3873 we mapped this protein on the two-dimensional SDS-PAGE proteomic map by highly sensitive immunodetection and observed that the native protein was well separated by the conditions employed in the proteomic studies. However, the amount of the protein in H37Rv was very small under the in vitro culture conditions (Fig. 1C), and attempts to analyze the immunoreactive protein spot from two-dimensional SDS-PAGE gels by matrix-assisted laser desorption ionization mass spectrometry were not successful, although in silico analysis of the primary sequence with the MS-Digest program (http://prospector.ucsf.edu) indicated that a sufficient number of potential trypsin cleavage sites are present in the protein.

In the present study we found that both rRv3873 and a large number of synthetic peptides derived from it are strongly recognized by human T cells primed during M. tuberculosis infection. The recognition of Rv3873 by healthy natural converters implies that this antigen may be expressed during active stages as well as subclinical or latent stages of infection. Analysis of PBMC recognition of synthetic overlapping peptides suggested that both TB-specific epitopes and epitopes that are cross-reactive with BCG are involved in the cellular immune response to Rv3873. Interestingly, a dominant epitope (Rv3873_118-135) was found to be highly conserved in a large number of PPE family members from various species of mycobacteria. The responses to this epitope may therefore not be primarily primed by Rv3873, as demonstrated by the recognition of this epitope by BCG-vaccinated donors. Our results also demonstrate that if the cross-reactive epitopes are removed, the remaining part of Rv3873 could still contain interesting moieties for inclusion in a cocktail for specific immune-based diagnosis of TB.

Recently, two other PPE proteins (Mtb39a/Rv1196 and Mtb41/Rv0916c) have been reported to be potent T-cell antigens that are recognized by M. tuberculosis-infected humans (10, 26). When administered as DNA vaccines in mice, these antigens provided substantial protection against challenge with aerosolized M. tuberculosis. Thus, since the completion of the genome sequence of M. tuberculosis in which the nomenclature for this family of antigens was introduced (9), the PPE family has contributed three candidates to the growing list of potential vaccine candidate antigens. This makes the PPE family an interesting target for systematic molecular and immunological characterization. In contrast to members of the ESAT-6 family, which have been identified in STCF of M. tuberculosis (3, 27), all three PPE antigens identified so far were absent in STCF. This demonstrates that secreted as well as cell-bound proteins contribute to protective immunity against M. tuberculosis (19), in agreement with a recent study involving vaccination with various subcellular fractions in animal models of TB (1).

In this study we identified a third important T-cell antigen that is encoded by the RD1 region. The genes encoding Rv3873, CFP10, and ESAT-6 are situated adjacent to each other in the M. tuberculosis genome. It has been suggested that the ORFs in RD1 may constitute a gene cluster that encodes a new transporter system in M. tuberculosis (11, 20, 30). The frequent occurrence of T-cell determinants in this region suggests that this functional unit may be particularly active in interacting with the host immune system and may play an important role in host-pathogen interactions. Recently, protein-protein interactions have been reported for ESAT-6 and CFP10 (22). It is therefore tempting to hypothesize that direct interactions between the proteins of this new transport system may lead to simultaneous availability of the RD1 proteins to the host immune system.

Recent data indicate that the clustering of immunodominant T-cell antigens in the RD1 region may not be a unique case and that there may be other such antigenic gene clusters in the M. tuberculosis genome. One example is Rv1196, another potent T-cell antigen belonging to the PPE family (10), which is genetically linked to an ESAT-6-like protein (Rv1198) that is strongly recognized by M. tuberculosis-infected individuals (2). It is therefore intriguing that in the present study several of the PPE family members present in BCG and suggested to be responsible for the cross-reactive immune responses to Rv3873_118-135 (Table 2 and Fig. 4B) were found to be encoded by ORFs that precede the genes that encode recently identified immunodominant ESAT-6 family members in the genome. This is the case for rv0286, rv3021c, and rv3018c, all of which are located upstream of the genes that encode recently identified potent T-cell antigens belonging to the ESAT-6 family (Rv0288, Rv3017c, and Rv3019c, respectively) (27, 28). If this phenomenon of antigen clustering is more widespread, it may lead to a new perspective on the in silico identification of novel antigens.