INFECTION AND IMMUNITY, Dec. 2002, p. 6996–7003 0019-9567/02/$04.00⫹0 DOI: 10.1128/IAI.70.12.6996–7003.2002 Copyright © 2002, American Society for Microbiology. All Rights Reserved.
Vol. 70, No. 12
Identification of Novel Mycobacterium tuberculosis Antigens with Potential as Diagnostic Reagents or Subunit Vaccine Candidates by Comparative Genomics P. J. Cockle,1 S. V. Gordon,1 A. Lalvani,2 B. M. Buddle,3 R. G. Hewinson,1 and H. M. Vordermeier1* TB Research Group, Department of Bacterial Diseases, Veterinary Laboratories Agency—Weybridge, New Haw, Addlestone,1 and Nuffield Department of Clinical Medicine, John Radcliffe Hospital, University of Oxford, Oxford,2 United Kingdom, and AgResearch, Wallaceville Animal Research Centre, Upper Hutt, New Zealand3 Received 15 March 2002/Returned for modification 3 June 2002/Accepted 20 August 2002
An independent review for the British government has concluded that the development of a cattle vaccine against Mycobacterium bovis holds the best long-term prospects for tuberculosis control in British herds. The development of complementary diagnostic tests to differentiate between vaccinated and infected animals is necessary to allow the continuation of test-and-slaughter-based control policies alongside vaccination. Vaccination with M. bovis bacillus Calmette-Gue´rin (BCG), the only available vaccine, results in tuberculin purified protein derivative sensitivity and has shown varying vaccine efficacies in cattle. Thus, identification of morespecific reagents to distinguish between vaccination and infection, as well as the identification of subunit vaccine candidates for improved tuberculosis vaccines, is a research priority. In the present study, we applied comparative genomics to identify M. bovis-Mycobacterium tuberculosis antigens whose genes had been deleted in BCG Pasteur. In total, 13 open reading frames (ORFs) from the RD1, RD2, and RD14 regions of the M. tuberculosis genome were selected. Pools of overlapping peptides spanning these ORFs were tested in M. bovis-infected (n ⴝ 22), BCG-vaccinated (n ⴝ 6), and unvaccinated (n ⴝ 10) control cattle. All were recognized in infected cattle, with responder frequencies varying between 16 and 86%. In particular, eight highly immunogenic antigens were identified whose potentials as diagnostic reagents or as subunit vaccines warrant further study (Rv1983, Rv1986, Rv3872, Rv3873, Rv3878, Rv3879c, Rv1979c, and Rv1769). Bovine tuberculosis (BTB) is caused by Mycobacterium bovis, which shows ⬎99.9% DNA identity with Mycobacterium tuberculosis, the main cause of human tuberculosis. Moreover, BTB is a zoonotic disease that was responsible during the 1930s and 1940s for approximately 6% of the total human deaths due to tuberculosis and ⬎50% of all cervical lymphadenitis cases in children (11). The introduction of pasteurization of milk in the 1930s dramatically reduced transmission from cattle to humans (11). However, BTB remains a small but significant cause of human morbidity and mortality, especially in developing countries (9). Cattle with a mycobacterial infection will predominantly mount a cellular immune response (6). Therefore, the skin test using tuberculin purified protein derivative (PPD) has become an integral part of the BTB eradication program. A compulsory eradication program based upon the slaughter of infected animals, detected by the single intradermal comparative tuberculin skin test, began in Great Britain in 1950, and by 1960 it had been implemented in all of Great Britain. These measures resulted in the dramatic reduction of BTB in Great Britain from incidence rates of around 40% of cattle infected with M. bovis to 0.41% of herds in 1996 (14). However, despite con-
tinued implementation of these control measures, the incidence of BTB in cattle has been steadily rising since 1988, possibly due to a wildlife reservoir of M. bovis (14). In addition to skin tests, blood-based diagnostic assays that measure antigen-induced production of lymphokines such as gamma interferon (IFN-␥) are also under consideration (27). However, specificity constraints are associated with the use of PPD in such assays. These arise due to the crude mixture of M. bovis proteins that PPD contains, including many that are crossreactive with other environmental mycobacterial species (e.g., Mycobacterium avium and Mycobacterium intracellulare) and, importantly, the vaccine strain M. bovis bacillus CalmetteGue´rin (BCG) (3, 7, 13). Other methods to reduce BTB in cattle are the development of cattle vaccines. A cattle vaccine would reduce the risk of cattle infection and hence result in lower tuberculin test frequencies and significant cost savings. Recently, a panel of scientists was commissioned by the British government to conduct an independent review of this problem, and they concluded that the development of an improved cattle vaccine provides the best long-term prospect for BTB control in British herds (14). It was also recommended that a complementary diagnostic test to differentiate between vaccinated animals and those infected with M. bovis (differential diagnosis) should be developed in parallel with the vaccine to ensure continuation of the test-and-slaughter-based control strategies (14). To date, BCG, an attenuated strain of M. bovis, is the only available vaccine for the prevention of BTB, although varying effi-
* Corresponding author. Mailing address: TB Research Group, Department of Bacterial Diseases, Veterinary Laboratories AgencyWeybridge, New Haw, Addlestone, KT15 3NB, United Kingdom. Phone: 44 1932 357 884. Fax: 44 1932 357 684. E-mail:
[email protected]. 6996
IMMUNOGENIC M. TUBERCULOSIS ANTIGENS
VOL. 70, 2002
cacies in cattle (3, 7, 26), as well as the fact that BCG vaccination compromises the specificity of tuberculin-based diagnostic reagents, has prevented its use in cattle. Diagnostic reagents which distinguish between vaccinated and infected cattle can be developed using specific, defined antigens that are present in virulent M. bovis but absent from the vaccine strain. Genetic analysis of BCG has revealed that several large genomic regions have been deleted during attenuation and subsequent prolonged propagation in culture (2, 4, 10). These regions have been characterized, and antigens from one of the regions, RD1 (16), have been studied extensively in several species, including humans and cattle (15, 17). For example, it has recently been demonstrated that protein or peptide cocktails composed of two RD1 region antigens, ESAT-6 and CFP-10, can be used to distinguish between BCG-vaccinated and M. bovis-infected cattle (5, 23, 25). At the same time, RD region antigens have also been described as subunit vaccine candidates (1). The present study has been designed to employ cattle models of M. bovis infection and BCG vaccination to identify highly immunogenic antigens from three genomic regions deleted during the evolution of BCG Pasteur (RD1, RD2, and RD14) (2, 4, 16). Five hundred and thirty-six overlapping synthetic peptides derived from the sequence of 13 antigens encoded in these regions were synthesized and used to diagnose infected or vaccinated cattle. The previously mentioned ESAT-6/ CFP-10 peptide cocktail (25) was also included as a “gold standard” with which to compare the immunogenicities of other antigens identified by the ability to stimulate IFN-␥ in whole-blood culture assays. We identified eight highly immunogenic antigens that warrant further investigation to determine their suitabilities as diagnostic reagents and/or as subunit vaccine candidates. MATERIALS AND METHODS Cattle. Ca. 6-month-old calves (Friesian or Friesian crosses) were obtained from herds free of BTB. The following groups of cattle were used in this study. (i) M. bovis infection. Nine calves were infected with a British M. bovis field strain (AF 2122/97) by intratracheal instillation of 2 ⫻ 104 CFU as described previously (7, 19). Twelve calves were infected with an M. bovis field strain isolated from a New Zealand infected cow, also using intratracheal instillation (5 ⫻ 103 CFU). BTB was confirmed in these animals by the presence of visible lesions in lymph nodes and lungs found in postmortem examinations and by the histopathological examination of lesioned tissues and the culture of M. bovis from tissue samples collected from lymph nodes and lungs. Heparinized-blood samples were obtained between 14 and 20 weeks after infection when strong and sustained in vitro tuberculin responses were observed. Data from a total of 21 experimentally infected cattle are presented in this study. One naturally infected animal was also used included in this group. (ii) BCG vaccination. Calves were vaccinated with BCG Pasteur by subcutaneous injection of 106 CFU into the side of the neck followed 8 weeks later by a booster injection using the same route and dose (7, 24). Heparinized-blood samples were taken between 4 and 6 weeks after the booster vaccination. Data from six calves are presented in this study. (iii) Uninfected controls. Heparinized blood from tuberculin skin test-negative calves from herds free of BTB (10 animals) was also obtained. These animals produced IFN-␥ in vitro after stimulation with tuberculin from M. avium, indicating that they had been exposed to environmental mycobacteria. Antigens and peptides. (i) Antigens. Bovine (PPD-B) and avian (PPD-A) tuberculins were obtained from the Tuberculin Production Unit at the Veterinary Laboratories Agency—Weybridge and used in culture at 10 g/ml. (ii) Peptides. A set of 536 synthetic peptides spanning 13 open reading frames (ORFs) (20 residues long with a 12-residue overlap) was prepared by multirod peptide synthesis. The peptides were used in mapping experiments in pools of 8
6997
to 11 at 5 g of each peptide/ml and at 25 g/ml when used individually. The peptides were purchased from Chiron Mimotopes (Clayton, Australia). ESAT-6and CFP-10-derived peptides were synthesized by solid-phase peptide synthesis and formulated into a peptide cocktail as described earlier (25). They were also used at 5 g of each peptide/ml. The purity and sequence fidelity of ESAT-6 and CFP10-derived peptides was confirmed by analytical reverse-phase high-performance liquid chromatography and by electron spray mass spectrometry, respectively. IFN-␥ enzyme-linked immunosorbent assay. Whole-blood cultures were performed in 96-well plates in 0.2-ml/well aliquots by mixing 0.1 ml of heparinized blood with an equal volume of antigen-containing solution (24). The supernatants were harvested after 48 h of culture at 37°C and 5% CO2 in a humidified incubator. The IFN-␥ concentration was determined using the BOVIGAM ELISA kit (Biocore AH, Omaha, Neb.). Results were deemed positive when the optical densities at 450 nm (OD450) with antigens minus the OD450 without antigens were ⱖ0.1. For comparative analysis of PPD-B versus PPD-A responses, a positive result was defined as a PPD-B OD450 minus PPD-A OD450 of ⱖ0.1 and a PPD-B OD450 minus unstimulated OD450 of ⱖ0.1. Bioinformatics. The DNA sequence of M. tuberculosis H37Rv was visualized using either the ARTEMIS display tool (20) or the TubercuList database (http:// genolist.pasteur.fr/TubercuList/). Basic Local Alignment Search Tool (BLAST) searches were performed from within TubercuList or using the National Center for Biotechnology Information BLAST server (http://www.ncbi.nlm.nih.gov/ BLAST).
RESULTS Selection of candidate antigens from the RD1, RD2, and RD14 regions of M. bovis. Thirteen ORFs from the RD1, RD2, and RD14 regions of M. bovis were selected for screening. These regions are deleted in BCG Pasteur, and proteins encoded within these regions hold promise as candidate antigens for the differential diagnosis of M. bovis-infected animals and BCG-vaccinated cattle and as potential vaccine candidates. The selection criteria were that the ORF should encode a protein that (i) showed no, or minimal, sequence similarity to other proteins in M. tuberculosis or other organisms, (ii) belonged to the PE or PPE protein family, (iii) had the potential for being induced or up-regulated in vivo (e.g., amino acid transporters), or (iv) had the potential to be secreted. The designations (Rv numbers) of the antigens encoded by the selected ORF, their sizes, and their putative functions are listed in Table 1. Immunogenicities of selected antigens in M. bovis-infected, BCG-vaccinated, and environmentally sensitized cattle. Five hundred and thirty-six overlapping peptides derived from the sequences of the selected antigens were synthesized. The peptides were then formulated into pools of approximately 10 neighboring overlapping peptides, which resulted in 52 peptide pools. Table 1 shows the pools in relation to the antigens they represent, as well as the total number of peptides per antigen required to ensure complete sequence coverage. Blood samples were obtained from 22 M. bovis-infected animals, 6 M. bovis BCG Pasteur-vaccinated animals, and 10 unvaccinated and uninfected controls. Whole-blood cultures in the presence of PPD-B, PPD-A, peptide pools, and a cocktail of 10 synthetic peptides derived from ESAT-6 and CFP-10 were established, and the amounts of IFN-␥ were determined after 48 h of culture. As expected, all M. bovis-infected and BCG-vaccinated animals responded more strongly to the bovine tuberculin, PPD-B, than to the avian tuberculin, PPD-A (median responses and ranges were as follows: M. bovis-infected cattle, PPD-B, 1.593 [0.274 to 3.500], and PPD-A, 1.313 [0.066 to 3.455]; BCG-vaccinated cattle, PPD-B, 0.886 [0.181 to 2.244],
6998
COCKLE ET AL.
INFECT. IMMUN. TABLE 1. RD region antigens selected for evaluation in this study Size (amino acids)
Peptide poolsb
Rv3872 Rv3873 Rv3878 Rv3879c
99 368 280 729
1 (11) 2–5 (45) 6–8 (34) 9–17 (90)
RD2
Rv1979c Rv1983 Rv1984c Rv1986 Rv1987 Rv1989c
481 558 217 199 142 186
18–23 (59) 24–30 (69) 31–33 (26) 34–35 (24) 36–37 (19) 38–39 (22)
Possible amino acid permease Member of the PE-PGRS subfamily of glycine-rich proteins Probable secreted cutinase Possible lysine transporter Possible chitinase Unknown
RD14
Rv1768 Rv1769 Rv1772
618 414 103
40–46 (76) 47–51 (51) 52 (12)
Member of the PE-PGRS subfamily of glycine-rich proteins Similar to Streptomyces coelicolor hypothetical protein Unknown
Deleted region
Designationa
RD1
Putative functionc
Member of PE-like protein family Member of M. tuberculosis PPE family Unknown alanine-rich protein Unknown alanine- and proline-rich protein
a
Rv designation of ORF as defined previously (8). Peptide pools covering the full sequence (total number of peptides required is shown in parentheses). c Putative function as suggested in reference 8. b
and PPD-A, 0.5115 [0.274 to 2.234]). Uninfected, nonvaccinated control animals responded strongly to avian PPD (PPDA), indicating that they were sensitized by environmental mycobacteria (median responses and ranges were as follows: PPD-B, 0.230 [0.090 to 0.684], and PPD-A, 0.686 [0.162 to 1.822]); these animals will be referred to below as PPD-A reactors. Next, we assessed the immunogenicities of the peptide pools shown in Table 1. Figure 1 depicts the results obtained with blood from one animal experimentally infected with M. bovis and one PPD-A reactor. The M. bovis-infected animal recognized at least one peptide pool from 12 of 13 antigens tested (Fig. 1A, C, and E). Figure 1A, C, and E also shows that the immune responses to even the larger antigens (which required up to nine peptide pools for complete coverage) could in most cases be attributed to one or two peptide pools (Fig. 1). This animal also responded to stimulation with the previously described peptide cocktail containing 10 peptides derived from ESAT-6 and CFP-10 (25) that was used for comparison (see the legend to Fig. 1). In contrast, none of the peptide pools induced IFN-␥ secretion in whole blood from the environmentally sensitized PPD-A reactor (Fig. 1B, D, and F). The peptide-induced IFN-␥ responses of all 38 M. bovisinfected and BCG-vaccinated cattle and PPD-A reactors (uninfected controls) to the 13 antigens are summarized in Fig. 2. When antigens were covered by more than one peptide pool, the result of the pool that stimulated the most IFN-␥ secretion is shown. Interestingly, all 13 antigens were recognized by M. bovis-infected cattle, albeit with the percentages of responding cattle (responder frequencies) varying between 21 and 86%. The most frequently recognized antigens were Rv3873, Rv3879c, and Rv1769, with responder frequencies of 82, 77, and 86%, respectively, whereas Rv1984c and Rv1772 were recognized only by 21 and 36% of infected calves. Interestingly, several of the most prominently recognized antigens were members of the PE/PPE protein family (e.g., Rv3873, with a responder frequency of 82%). Only two antigens (Rv1979c and Rv1769, with responder frequencies of 73 and 86%, respectively) were recognized by PPD-A reactor cattle. These two
antigens were also strongly recognized by BCG-vaccinated cattle, with responder frequencies of 67 and 100%, respectively. Surprisingly, considering the absence of the genes encoding these antigens in BCG Pasteur, 9 of the 13 antigens tested (Rv3873, Rv3879c, Rv1979c, Rv1983, Rv1987, Rv1989c, Rv1768, Rv1769, and Rv1772, with a range in responder frequencies of 17 to 100%) stimulated a positive response in BCG-vaccinated animals. The remaining four antigens (Rv3872, Rv3878, Rv1984c, and Rv1986, with a range in responder frequencies of 21 to 59%) were recognized by M. bovis-infected cattle only. In addition, 21 of 22 M. bovis-infected animals responded to a previously characterized peptide cocktail derived from CFP-10 and ESAT-6 (25). The responder frequencies of the eight most immunogenic antigens, as well as for the ESAT-6/CFP-10 derived peptide cocktail, are summarized in Table 2. Responses of peptide pools can be the result of a single peptide. The peptide pools we formulated contained between 8 and 11 peptides (Table 1 provides details of the peptide pools). To determine whether the IFN-␥ responses of the pools were due to single or multiple peptide constituents, we tested the individual peptides of pool 3 (representing residues 89 to 188 from Rv3873) and pool 26 (representing residues 161 to 252 from Rv1983) using blood from five M. bovis-infected animals. All three animals tested that recognized pool 3 responded predominantly to peptide 3.2 (Fig. 3A to C, residues 97 to 116), whereas both animals that responded only to pool 26 recognized predominantly peptide 26.2 (Fig. 3D and E, residues 169 to 188). These data suggest that the individual peptides imparting antigenicity can be identified from immunodominant pools and that pool immunogenicity can be attributed to single peptides. DISCUSSION In this paper, we have demonstrated the effective use of comparative genomics in combination with synthetic peptides to identify and screen 13 potential antigens encoded by ORFs located in the RD1, RD2, and RD14 regions of the M. tuber-
VOL. 70, 2002
IMMUNOGENIC M. TUBERCULOSIS ANTIGENS
6999
FIG. 1. Recognition of RD region gene products by an M. bovis-infected cow (A, C, and E) and a PPD-A reactor (B, D, and F). Whole blood was cultured in the presence of peptide pools of between 8 and 11 peptides representing RD1 (A and B), RD2 (C and D), and RD14 (E and F) at 5 g of each peptide/ml. The dashed horizontal lines indicate positive cutoffs (OD450 with antigens ⫺ OD450 without antigens ⱖ 0.1). The results (OD450 values) obtained with an ESAT-6/CFP-10 peptide cocktail (25) were as follows: M. bovis-infected cow, 1.086; PPD-A reactor, 0.016. The error bars indicate standard errors.
7000
COCKLE ET AL.
INFECT. IMMUN.
FIG. 2. IFN-␥ responses induced by RD region antigens in M. bovis-infected (n ⫽ 22), BCG-vaccinated (n ⫽ 6), and PPD-A reactor (n ⫽ 10) cattle. Only the results of the pool and antigen stimulating the greatest IFN-␥ response are shown. Green squares, M. bovis-infected cattle; red triangles, PPD-A reactors; blue circles, BCG-vaccinated cattle. The dashed horizontal lines indicate the positive cutoff (OD450 with antigens ⫺ OD450 without antigens ⱖ 0.1).
culosis genome. Our results indicated that six antigens in particular showed promise as diagnostic antigens because they were recognized either (i) by M. bovis-infected animals alone and not by BCG-vaccinated animals or control animals sensi-
tized by exposure to environmental mycobacteria (differential diagnosis [Table 2]) or (ii) by infected animals and vaccinated animals but not by exposed controls (specific diagnosis [Table 2]). Two more antigenic proteins (Rv1979c and Rv1769) can
IMMUNOGENIC M. TUBERCULOSIS ANTIGENS
VOL. 70, 2002 TABLE 2. Recognition of most immunogenic antigensa Potential application
Responder frequency (%) Designation
M. bovis reactorsb
BCG vaccinatedc
M. avium reactorsd
Differential diagnostics
Rv1986
41
0
0
Rv3872 Rv3878 Combined ESAT-6/ CFP-10e
50 59 82 95
0 0 0 0
0 0 0 10
Specific diagnostics
Rv1983
59
33
0
Rv3873 Rv3879c Combined
82 77 91
17 33 50
0 0 0
Vaccines
Rv1979c Rv1769
73 86
67 100
40 30
a
Only antigens recognized by ⬎40% of M. bovis-infected animals are listed. Results from 22 cattle experimentally infected with M. bovis. c Results from five or six BCG-vaccinated cattle. d Results from 10 environmental-mycobacterium-sensitized cattle. e Results obtained with ESAT-6/CFP-10-derived peptide cocktail (27). b
be considered vaccine candidates, since they were recognized by animals from all three categories (Table 2). It is unlikely that a single diagnostic antigen could impart enough sensitivity to provide population coverage, and therefore combinations of specific antigens will be needed. We analyzed our data, considering the two scenarios described above. First, we considered antigens suitable for differential diagnosis: combining Rv1986, Rv3872, and Rv3878 indicated that 82% of the infected animals would have been correctly identified by their responses to these three antigens. Second, we considered the three most immunodominant antigens (Rv1983, Rv3873, and Rv3879c) that had potential for specific diagnosis: taken together, these antigens would have identified 20 of 22 (91%) M. bovis-infected animals (Table 2). Interestingly, if we considered Rv3878 from the first category together with Rv3873 and Rv3879c from the second category, 21 (95%) of the 22 M. bovis-infected animals would have been detected, a figure identical to that for the ESAT-6/CFP-10 cocktail. All 13 antigens tested were recognized with responder frequencies varying between 21 and 86%. This hierarchy of response frequencies is interesting and worthy of further evaluation, as it could give insight into the nature of the immunogenicity of antigens from M. bovis. It is likely that a combination of several factors determines whether and to what degree mycobacterial proteins are immunogenic after infection. These factors could include (i) parameters intrinsic to the bacterium, such as the abundance of the protein and its subcellular location, posttranslational modification, participation in macromolecular complexes, and in vivo regulation, and (ii) factors relating to the immune system, including the location of the antigen with respect to the phagosome, proteolytic sensitivity, and the presence of motifs suitable for interaction with TAP transporters and different major histocompatibility complex alleles within the antigen. The approach described in the present report exploits pools
7001
of overlapping synthetic peptides derived from the sequences of these proteins. The number of peptide pools that represent the sequences of each ORF varies depending on the size of the antigen, as illustrated in Table 1. In the present study, we demonstrated that the combined results from Rv3873, Rv3878, and Rv3879c resulted in an overall responder frequency of 95%. These three antigens are represented by a total of 16 different peptide pools containing 169 individual peptides. However, the same frequency of recognition can be obtained using just 3 out of the 16 pools assayed (pools 3, 8, and 9), i.e., 33 peptides, suggesting the presence of the immunodominant epitopes within these pools. Indeed, the number of peptides needed to achieve responder frequencies similar to that with the complete set of overlapping peptides might be significantly lower, since the data shown in Fig. 3 demonstrate that only one or two immunodominant peptides can be responsible for the immunogenicity of an entire pool. Interestingly, the previously described peptide cocktail containing peptides derived from ESAT-6 and CFP-10 (25) was recognized by 95% of the M. bovis-infected animals tested. Moreover, the same animals responded to the combination of Rv3873, Rv3878, and Rv3879c. We have previously observed high frequencies of responders to ESAT-6 and CFP-10 in experimentally infected calves (reference 19 and unpublished observations). These response frequencies are much higher than the frequencies of responses to ESAT-6 observed in field samples (about 66 to 78%) reported by various laboratories (17, 18, 22, 24). The difference between the observed frequencies of responders to ESAT-6 in naturally infected and experimentally infected animals could be a result of sampling at an earlier stage of the disease in the experimentally infected calves, when ESAT-6 might be recognized preferentially (19). Therefore, future field trials will decide the relative merits of including the antigens described in this paper alongside ESAT-6 and CFP-10 with the aim of improving diagnostic sensitivity to a standard as good as, or better, than tuberculin PPD. As shown in Table 1, four PPE/PE genes were selected for testing (Rv3872, Rv3873, Rv1983, and Rv1768) and gave responder frequencies of between 45 and 82% when assayed in the M. bovis-infected cattle. Little is known about the functions or immunogenicities of these proteins, which account for approximately 10% of the total coding capacity of the M. tuberculosis genome (8). Here, we report that during M. bovis infection of cattle, a pool of peptides comprising part of Rv3873, a member of the PPE family of proteins, is recognized by 82% of the infected animals. However, the pool was also recognized in a BCG-vaccinated animal. This can be considered a surprising outcome given that the gene is deleted in BCG and that no homologous proteins were found elsewhere in the BCG genome. However, the unit of cross-reactivity is the epitope, and the molecular nature of cross-reactivity can be addressed only once these epitopes have been identified (12). When we used the sequence of the dominant peptide 3.2 to search for similar regions within other genes found in the M. tuberculosis genome, several regions with high identity were found. Table 3 shows the results using the BLAST program (http://www .ncbi.nlm.nih.gov/BLAST/) to identify similarity between mycobacterial proteins. The table highlights several sequences that contain amino acid identities of greater than 50%. All of
7002
COCKLE ET AL.
INFECT. IMMUN.
FIG. 3. IFN-␥ secretion induced by individual peptides from pool 3 and pool 26. Whole blood was collected from five cattle experimentally infected with M. bovis and incubated for 48 h with individual peptides (25 g/ml each) contained in peptide pool 3 (calves A to C) or peptide pool 26 (calves D and E). The results are expressed as delta mean OD450 (OD450 values with antigens ⫺ OD450 without antigens) of duplicate determinations, with a positive cutoff (horizontal lines) of 0.1. The error bars indicate standard errors.
the proteins identified are from the M. tuberculosis genome and are also members of the PPE family. The peptide covers an area of the gene that encodes two motifs identified in a number of PPE family members during their annotation (21; http://genolist.pasteur.fr/TubercuList/). Therefore, it is reasonable to speculate that the cross-reactive nature of the peptide is a result of similarity to other PPE family members located elsewhere in the genome of M. bovis BCG Pasteur. We conducted BLAST searches for the other identified cross-reactive antigens (e.g., Rv1979c) by comparing the whole genes in steps of 20 amino acids, representing the corresponding peptides, and were able to find numerous similar amino acid sequences in other mycobacterial proteins outside the deleted regions. The strategy of using peptides instead of recombinant proteins has advantages already discussed. However, in regard to the observed cross-reactivity of antigens between BCG-vacci-
TABLE 3. Homology between peptide 3.2 from Rv3873 and sequences from other mycobacterial proteinsa Designationb
Rv3873 Rv3021c Rv0286 Rv3018c Rv0280 a
Putative function
M. tuberculosis PPE family M. tuberculosis PPE family M. tuberculosis PPE family M. tuberculosis PPE family M. tuberculosis PPE family
Amino acid sequencec
AMATTPSLPEIAANHIT ALAEMPTLPELAANHLT ALAAMPTLAELAANHVI ALAEMPTLPELAANHLT AVAAMPTLVELAANHTL
The homology search was performed using the BLAST program. Designation of M. tuberculosis proteins as described previously (8). The sequences of these antigens in M. tuberculosis and M. bovis were found to be identical. c Nonidentical residues are underlined. b
IMMUNOGENIC M. TUBERCULOSIS ANTIGENS
VOL. 70, 2002
nated and M. bovis-infected animals, this peptide-based approach has other distinct advantages. If we take ORF Rv1987 as an example, it appears unsuitable as a differential diagnostic reagent due to the high cross-reactivity in the BCG-vaccinated cattle. However, the responder frequency of 57% in M. bovisinfected cattle is due to the comparable levels of recognition of the two representative pools, with responder frequencies of 47 and 53%, respectively. This is in contrast to the BCG-vaccinated animals, in which only the first of the pools is immunogenic, recognized by 50% of the cattle. Therefore, with a peptide-based approach, the diagnostic potential of this antigen can still be realized by using only the second peptide pool, still achieving a sensitivity of 53% in the M. bovis-infected cattle. In conclusion, this study demonstrates that the analysis of peptides derived from genes deleted in BCG Pasteur can lead to the identification of novel antigens for vaccination or diagnosis. In particular, we describe antigens that can form the basis of diagnostic reagents either to differentiate between infected and BCG-vaccinated animals or to improve the specificity of PPD per se. The evaluation of these antigens as diagnostic reagents alongside antigens like ESAT-6 and CFP-10 will be done in future large-scale field studies. In addition, this study also demonstrates for the first time that members of both the PE and PPE families of proteins induce cellular immune responses during tuberculous infection in a natural host species.
8.
9. 10.
11. 12. 13. 14. 15.
16. 17. 18.
ACKNOWLEDGMENTS This work was funded by the Department for Environment, Food and Rural Affairs, Great Britain. We express our appreciation to the staff of the animal service unit at the Veterinary Laboratories Agency and farm staff from AgResearch for their dedication to animal welfare. REFERENCES 1. Andersen, P. 2001. TB vaccines: progress and problems. Trends Immunol. 22:160–168. 2. Behr, M. A., M. A. Wilson, W. P. Gill, H. Salamon, G. K. Schoolnik, S. Rane, and P. M. Small. 1999. Comparative genomics of BCG vaccines by wholegenome DNA microarray. Science 284:1520–1523. 3. Berggren, S. A. 1981. Field experiment with BCG vaccine in Malawi. Br. Vet. J. 137:88–94. 4. Brosch, R., S. V. Gordon, M. Marmiesse, P. Brodin, C. Buchrieser, K. Eiglmeier, T. Garnier, C. Gutierrez, G. Hewinson, K. Kremer, L. M. Parsons, A. S. Pym, S. Samper, D. van Soolingen, and S. T. Cole. 2002. A new evolutionary scenario for the Mycobacterium tuberculosis complex. Proc. Natl. Acad. Sci. USA 99:3684–3689. 5. Buddle, B., N. A. Parlane, D. L. Keen, F. E. Aldwell, J. M. Pollock, K. Lightbody, and P. Andersen. 1999. Differentiation between Mycobacterium bovis BCG-vaccinated and M. bovis-infected cattle using recombinant mycobacterial antigens. Clin. Diagn. Lab. Immunol. 6:1–5. 6. Buddle, B. M., G. W. de Lisle, A. Pfeffer, and F. E. Aldwell. 1995. Immunological responses and protection against Mycobacterium bovis in calves vaccinated with a low dose of BCG. Vaccine 13:1123–1130. 7. Buddle, B. M., D. Keen, A. Thomson, G. Jowett, A. R. McCarthy, J. Heslop, G. W. De Lisle, J. L. Stanford, and F. E. Aldwell. 1995. Protection of cattle
Editor: S. H. E. Kaufmann
19. 20. 21. 22. 23.
24.
25.
26. 27.
7003
from bovine tuberculosis by vaccination with BCG by the respiratory or subcutaneous route, but not by vaccination with killed Mycobacterium vaccae. Res. Vet. Sci. 59: 10–16. Cole, S. T., R. Brosch, J. Parkhill, T. Garnier, C. Churcher, D. Harris, S. V. Gordon, K. Eiglmeier, S. Gas, C. E. Barry III, F. Tekaia, K. Badcock, D. Basham, D. Brown, T. Chillingworth, R. Connor, R. Davies, K. Devlin, T. Feltwell, S. Gentles, N. Hamlin, S. Holroyd, T. Hornsby, K. Jagels, B. G. Barrell, et al. 1998. Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature 393:537–544. Cosivi, O., F. X. Meslin, C. J. Daborn, and J. M. Grange. 1995. Epidemiology of Mycobacterium bovis infection in animals and humans, with particular reference to Africa. Rev. Sci. Technol. 14:733–746. Gordon, S. V., R. Brosch, A. Billault, T. Garnier, K. Eiglmeier, and S. T. Cole. 1999. Identification of variable regions in the genomes of tubercle bacilli using bacterial artificial chromosome arrays. Mol. Microbiol. 32:643– 655. Hardie, R. M., and J. M. Watson. 1992. Mycobacterium bovis in England and Wales: past, present and future. Epidemiol. Infect. 109:23–33. Harris, D., H. Vordermeier, M. Singh, C. Moreno, S. Jurcevic, and J. Ivanyi. 1995. Cross-recognition by T cells of an epitope shared by two unrelated mycobacterial antigens. Eur. J. Immunol. 25:3173–3179. Hubrig, T., and W. Kruger. 1958. Untersuchungen zur Tuberkuloseschutzimpfung bei Rindern. Monatsh. Veterinarmed. 13:513–519. Krebs, J. R. 1997. Bovine tuberculosis in cattle and badgers. Ministry of Agriculture, Fisheries and Food Publications, London, United Kingdom. Lalvani, A., A. Pathan, H. McShane, R. Wilkinson, M. Latif, C. Conlon, G. Pasvol, and A. Hill. 2001. Rapid detection of Mycobacterium tuberculosis infection by enumeration of antigen-specific T cells. Am. J. Respir. Crit. Care Med. 163:824–828. Mahairas, G. G., P. J. Sabo, M. J. Hickey, D. C. Singh, and C. K. Stover. 1996. Molecular analysis of genetic differences between Mycobacterium bovis BCG and virulent M. bovis. J. Bacteriol. 178:1274–1282. Pollock, J. M., and P. Andersen. 1997. The potential of the ESAT-6 antigen secreted by virulent mycobacteria for specific diagnosis of tuberculosis. J. Infect. Dis. 175: 1251–1254. Pollock, J. M., R. M. Girvin, K. A. Ligthbody, R. A. Clements, S. D. Neill, B. M. Buddle, and P. Andersen. 2000. Assessment of defined antigens for the diagnosis of bovine tuberculosis in skin test-reactor cattle. Vet. Rec. 146: 659–665. Rhodes, S. G., D. Gavier-Widen, B. M. Buddle, A. O. Whelan, M. Singh, R. G. Hewinson, and H. M. Vordermeier. 2000. Antigen specificity in experimental bovine tuberculosis. Infect. Immun. 68:2573–2578. Rutherford, K., J. Parkhill, J. Crook, T. Horsnell, P. Rice, M. Rajandream, and B. Barrell. 2000. Artemis: sequence visualisation and annotation. Bioinformatics 10:944–945. Tekaia, F., S. V. Gordon, T. Garnier, R. Brosch, B. G. Barrell, and S. T. Cole. 1999. Analysis of the proteome of Mycobacterium tuberculosis in silico. Tuber. Lung Dis. 79:329–342. van Pinxteren, L. A., P. Ravn, E. M. Agger, J. Pollock, and P. Andersen. 2000. Diagnosis of tuberculosis based on the two specific antigens ESAT-6 and CFP10. Clin. Diagn. Lab. Immunol. 7:155–160. Vordermeier, H. M., P. C. Cockle, A. Whelan, S. Rhodes, and R. G. Hewinson. 2000. Towards the development of diagnostic assays to discriminate between Mycobacterium bovis infection and BCG vaccination in cattle. Clin. Infect. Dis. 30:S291–S298. Vordermeier, H. M., P. C. Cockle, A. Whelan, S. Rhodes, N. Palmer, D. Bakker, and R. G. Hewinson. 1999. Development of diagnostic reagents to differentiate between Mycobacterium bovis BCG vaccination and M. bovis infection in cattle. Clin. Diagn. Lab. Immunol. 6:675–682. Vordermeier, H. M., A. Whelan, P. J. Cockle, L. Farrant, N. Palmer, and R. G. Hewinson. 2001. Use of synthetic peptides derived from the antigens ESAT-6 and CFP-10 for differential diagnosis of bovine tuberculosis in cattle. Clin. Diagn. Lab. Immunol. 8:571–578. Waddington, F. G., and D. C. Ellwood. 1972. An experiment to challenge the resistance to tuberculosis in BCG vaccinated cattle in Malawi. Br. Vet. J. 128:541–552. Wood, P. R., and J. S. Rothel. 1994. In vitro immunodiagnostic assays for bovine tuberculosis. Vet. Microbiol. 40:125–135.
