Identification of novel immunogenic Mycobacterium ...

FEMS Microbiology Letters 177 (1999) 123^130

Identi¢cation of novel immunogenic Mycobacterium tuberculosis peptides that stimulate mononuclear cells from immune donors Alison J. Moran a; *, James L. Doran a , Jiong Wu a , Janice D. Treit a , Pattama Ekpo a , Valerie J. Kerr a , Alan D. Roberts b , Ian M. Orme b , Shirleen Galant c , Stanley R. Ress c , Francis E. Nano a a

Department of Biochemistry and Microbiology, Petch Building, University of Victoria, Victoria, B.C. V8W 3P6, Canada b Department of Microbiology, Colorado State University, Ft. Collins, CO, USA c Department of Medicine, Grootte Schuur Hospital and University of Cape Town, Cape Town, South Africa Received 2 June 1999 ; accepted 7 June 1999

Abstract Proteins which are secreted or associated with the cell envelope of Mycobacterium tuberculosis may contain protective T-cell epitopes. Prior to this study, a recombinant clone bank of enzymatically active M. tuberculosis-alkaline phosphatase fusions, were screened for immunogenicity in a murine T-cell model. Five of these were selected for further study, and the IFN-Q secretion and proliferation of human PBMC from purified protein derivative- (PPD)-positive and PPD-negative donors were measured in response to oligopeptides, Mtb-PhoA fusions and one full-length protein. Epitopes from four of the five selected antigens were immunoreactive in the human model and corresponded to cytochrome d ubiquinol oxidase, cytochrome c oxidase subunit II, MTV005.02 and MTV033.08. Thus, this strategy identified novel human immunogenic peptides as possible candidates for a subunit vaccine. ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords : Mycobacterium tuberculosis ; T-cell epitope; Vaccine candidate

1. Introduction The current tuberculosis vaccines are live, attenuated strains of Mycobacterium bovis, commonly referred to as bacillus Calmette-Gue¨rin (BCG). However, the e¤cacy of BCG is controversial [1] and as a live vaccine, BCG not only poses a threat of disease

* Corresponding author. Tel.: +1 (250) 721-6133; Fax: +1 (250) 721-8855; E-mail: [email protected]

to the immunocompromised, but vaccination also interferes with surveillance programs that rely on hypersensitivity reactions. A characteristic of immunity to the intracellular pathogens Mycobacterium tuberculosis and Listeria monocytogenes has been the requirement for live vaccines. This resulted in the investigation of antigens secreted during growth or exposed on the surface of the bacteria as an enriched source of protective Tcell epitopes. Support for the potential e¤cacy of a subunit vaccine has been demonstrated by the partial

0378-1097 / 99 / $20.00 ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 0 9 7 ( 9 9 ) 0 0 3 0 0 - 6

FEMSLE 8893 15-7-99

124

A.J. Moran et al. / FEMS Microbiology Letters 177 (1999) 123^130

protection against tuberculosis provided by vaccinating animals with M. tuberculosis culture supernatants [2] or puri¢ed secreted proteins [3]. M. tuberculosis culture ¢ltrate proteins induce delayed type hypersensitivity in mice [4] and stimulate cytokine production in both M. tuberculosis-immune mouse T cells [5] and human PBMC [6]. In the murine model of tuberculosis, there are protective CD4+ cells which emerge about a week after infection and have been shown to secrete both IFN-Q and IL-2 [5]. This study was undertaken to con¢rm whether antigens selected using a murine tuberculosis T-cell assay were recognised by a human peripheral blood mononuclear cell (PBMC) assay. Others have used M. tuberculosis-alkaline phosphatase (M.tb-PhoA) fusions, to identify exposed or secreted proteins [7] and we have isolated a number of fusions which were extremely immunoreactive in the murine assay. In this work, we have tested the proliferation and secretion of IFN-Q by human PBMC in response to di¡erent epitopes derived from ¢ve of these proteins.

2. Materials and methods 2.1. Media, and bacterial strains Luria-Bertani (LB) medium [8] was used for the growth of most recombinant Escherichia coli DH5K [8] and E. coli KS330 [9] strains, with appropriate antibiotics (ampicillin 250 Wg ml31 , chloramphenicol 10 Wg ml31 ). Tryptone-phosphate medium [8] was used for E. coli JM109 [10]. 2.2. Identi¢cation and partial puri¢cation of M.tb-PhoA fusion proteins Prior to this study, a clone bank of 1^3 kb fragments of M. tuberculosis H37Rv DNA in the pJDTphoA fusion vectors was used to select extracytoplasmically expressed alkaline phosphatase activity [11]. 368 fusion proteins were selected, analyzed by Western blot and partially puri¢ed from alkaline phosphatase-positive colonies. These proteins were then screened for their immunogenicity, in a murine T-cell model of tuberculosis [12] and certain fusion proteins elicited strong IFN-Q responses. The M. tuberculosis DNA upstream of phoA, was sequenced

[11] and compared by alignment to the M. tuberculosis sequence database provided by the Sanger Centre using the BLASTN program [13]. Of the 22 fusion proteins which gave maximal responses in the murine T-cell assay system [12] (i.e. mean IFN-Q s 20 000 pg ml31 from quadruplicate wells), ¢ve were chosen for further study. These were encoded on the plasmids pJTH1-92, pJTH1150, pJTH1-200, pJTH1-27 (HinP1I fragment of chromosomal DNA cloned into the BstBI site of pJDT1) and pJTA2-898 (AciI fragment cloned into the BstBI site of pJDT2). We took three di¡erent approaches to obtain antigens to test with human immune cells. These approaches were determined by the availability of the full-length DNA sequence of the genes and by the di¤culty in expressing some recombinant antigens. The antigens used in this study included PhoA fusion proteins, oligopeptides corresponding to the deduced amino acid sequence of PhoA fusion proteins, and one full-length recombinant M. tuberculosis protein. M.tb-PhoA fusion proteins were expressed in E. coli KS330. Cultures were grown at 37³C to midlog phase and induced with 1 mM IPTG for 3 h. Cells were collected by centrifugation, resuspended in lysis bu¡er (1% SDS, 1 mM EDTA, 25 mM DTT, 10% glycerol and 50 mM Tris-HCl, pH 7.5), sonicated and then separated by SDS polyacrylamide gel electrophoresis (PAGE) using 10% slab gels [14]. Proteins were transferred to nitrocellulose and strips removed for Western analysis. PhoA fusion proteins were located with the anti-PhoA monoclonal antibody VIAP1 (Caltag) and an anti-mouse IgG coupled to horse radish peroxidase (HRP) and imaged using the chemiluminescent NEN Renaissance substrate (DuPont). Parallel strips of nitrocellulose were washed with phosphate-bu¡ered saline (PBS) and water then eluted with 20% acetonitrile and lyophilized. Dried protein was resuspended in water and re-precipitated and washed with cold acetone. Samples were resuspended as needed, in PBS and diluted with tissue culture medium RPMI 1640. 2.3. Oligopeptides The computer program T Sites [15] was used to select regions from partial amino acid sequences that

FEMSLE 8893 15-7-99

A.J. Moran et al. / FEMS Microbiology Letters 177 (1999) 123^130 Table 1 Oligopeptide antigens Mtb827/cytochrome Antigen 1 Antigen 2 Antigen 3 Antigen 4 Antigen 5

d oxidase (AppC) AWYRLTKFFGKLFLINF WSEYSRFVGDVFGAPLA EYSRFVGDVFGAPLAME WIFGWNRLPRLVHLACI GRAELSSIVVLLTNNTA

Mtb27/MTV023.04c Antigen 6 GKTYDAYFTDAGGITPG Antigen 7 YDAYFTDAGGITPGNSV Mtb200/MTV033.08 Antigen 8 DFNRDSREVVHLATGMA Mtb31/19 kDa Antigen Antigen 9 AVAGAAILVAGLSGCSS

might contain T-cell epitopes. Oligopeptides were synthesized (Chiron Technologies) as 17-mers with a cyclic dipeptide (diketopiperazine) at the carboxy terminus (Table 1). The oligopeptides were suspended in PBS, aliquoted, and frozen at 370³C until used. Prior to assay, a vial was thawed and diluted in RPMI 1640 containing 10% AB serum. 2.4. Cloning and expression of cytochrome c oxidase subunit II (CtaC) Partial sequencing of pJTH1-92 identi¢ed an open reading frame (ORF). An M. tuberculosis cosmid library in pYUB328 [16] was screened by PCR, using primers 5P-CCCAGCTTGTGATACAGGAGG-3P and 5P-CGCTTGCCGGACGGCAGCACCAGC-3P. A positive clone (pG3) was identi¢ed and the ORF designated mtb-92. A 1.4 kb DNA fragment containing mtb-92 was ampli¢ed from pG3 by PCR using the primers 5PAAGCTTCGCCATGCCGCCGGTAAGCGCC-3P and 5P-TCTAGACACCACCACCACCACCACGTGACACCTCGCGGGCCAGGTC-3P, which introduced an XbaI site and a 6His-tag coding region upstream and HindIII site downstream of mtb-92. The 100 ml PCR reaction contained 250 mM dNTPs, 300 nM of each primer, 10 ml 10UPCR bu¡er and 1 U Taq DNA for 40 cycles (30 s at 95³C; 1 min at 62³C; 1.5 min at 72³C). The PCR products were ligated into pGEM-T (Promega) and the resulting plasmid was restricted with XbaI and HindIII and the fragment was ligated into pMAL-c2

125

(New England Biolabs). In the resulting plasmid, pMAL-MTB92-4.5, mtb-92 was in-frame with the gene encoding maltose binding protein (MBP) and expressed as an MBP-6His-MTB92 translational fusion protein. E. coli JM109 containing pMAL-MTB92-4.5 was induced at an OD = 0.5 to 0.6 (550 nm) with IPTG (1 mM) for a further 4 h. Cell pellets were resuspended in 50 ml of bu¡er A (20 mM Tris-HCl, pH 7.2, 200 mM NaCl, 1 mM EDTA) and sonicated (10U20 s pulses), spun at 10 000Ug (4³C for 30 min) and the supernatant diluted 5U in bu¡er A. This was applied to an amylose-resin column (New England Biolabs), washed with 10 bed volumes of bu¡er A, and bound MBP-6His-MTB92 was eluted with bu¡er A containing 10 mM maltose. 2.5. Antigens The predicted amino acid sequences of the proteins chosen, identi¢ed two terminal oxidases, CtaC and cytochrome d ubiquinol oxidase (AppC) and three hypothetical proteins. The Mtb92-PhoA fusion protein was used as Antigen 11 and was predicted to contain 295 amino acids of the M. tuberculosis CtaC protein (MTCY190.11c). The full-length protein (MBP-6His-Mtb92), expressed as a translational fusion, was used as Antigen 10. Mtb827-PhoA showed extensive similarity to other bacterial, AppC proteins. This fusion protein was poorly expressed and so ¢ve oligopeptides (designated Antigens 1 to 5) were designed. The full-length sequences of AppC (MTCY01B2.15c) and the three fusions, Mtb898-PhoA, Mtb27-PhoA and Mtb200PhoA (MTV005.02, MTV023.04c and MTV033.08 respectively), are now available [17]. Of these proteins, only Mtb898-PhoA was easily expressed. Therefore, oligopeptides were also designed for the latter two proteins. These were Antigens 6 and 7 for Mtb27-PhoA and Antigen 8 for Mtb200-PhoA. The fusion Mtb898-PhoA included the N-terminal 46 amino acids of the predicted protein, and was used as Antigen 12. No known immunogenic proteins elicited maximal responses in the mouse model, however, many were present at slightly lower response levels e.g. the fusion Mtb31-PhoA, corresponding to the 19 kDa antigen (mean IFN-Q = 14 638 pg ml31 from quadru-

FEMSLE 8893 15-7-99

126


plicate wells). For comparison, a well characterized, T-cell epitope containing sequence within the 19 kDa antigen was prepared as a 17-mer oligopeptide, excluding the outer two amino acids of the 19-mer used in the original study. This was designated Antigen 9. Apart from this peptide, none of the other peptides demonstrated any particular homology to the known stimulatory regions documented in MHCPEP [18]. Antigen 13 was the negative control, i.e. E. coli PhoA protein. 2.6. Subjects Sixteen donors of mixed racial and geographic origin were recruited in Cape Town, South Africa. Some subjects mounted a proliferative response to the puri¢ed protein derivative of M. tuberculosis (PPD (Central Veterinary College, Weybridge, UK)), despite a negative PPD skin test. Thus, proliferation in response was taken as the indicator of response status. Subjects were categorized as positive (14) or negative responders (two), accordingly. The positive responders divided into two groups, 11 which demonstrated strong proliferative responses to PPD and heat killed M. tuberculosis H37Rv and three which had statistically signi¢cant, but weaker responses, i.e. SI 6 5 for PPD, although responses were not necessarily lower for the heat killed cells. All positive responders were considered to be M. tuberculosis-immune and had no clinical history of tuberculosis. It is extremely di¤cult to ¢nd true PPD-negative responders where tuberculosis is endemic, thus, only two non-responders were identi¢ed and considered M. tuberculosis-naive. Although the HIV status of the donors is unknown, the PHA reactivity (Table 2), resulting from the polyclonal stimulation of all T cells, indicates comparable reactivity among all the subjects. Therefore, it is unlikely that any signi¢cant immune de¢ciency could be responsible for lack of reactivity to the test antigens. 2.7. Proliferation assay Twenty ml heparinized blood was taken from each donor and centrifuged at 250Ug to obtain a bu¡y layer. PBMC were obtained by Ficoll density centrifugation of 8 ml of a 1:1 mix of bu¡y layer and PBS; followed by washing 3U in PBS and ¢nal re-

Table 2 Responses to PHA Subject

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Median IFN-Q stimulation (pg ml31 )

SI for incorporation of 3 H-thymidine

BG day 3

PHA

(PHA/BG day 3)

188 214 319 182 292 186 177 233 202 178 212 92 183 175 129 190

8601 10855 7075 9700 13035 11965 16680 5780 10325 8770 5545 5410 3200 11810 7641 3825

389 368 433 458 232 1739 232 252 260 605 407 500 432 725 376 213

BG, background. Using the Mann-Whitney Rank Sum test, all subjects showed a statistically signi¢cant di¡erence between background and IFN-Q stimulation of P = 0.029; and between background and proliferation of P = 0.029, as measured by the incorporation of 3 H-thymidine. Medians were calculated from four replicate wells assayed on day 3.

suspension at 1U106 PBMC ml31 in RPMI 1640 which had been supplemented with 10% pooled, heat-inactivated human AB serum (Western Province Blood Transfusion Service, Cape Town, South Africa). The same batch of AB serum was used for the entire study. Prior to use, serum was tested to exclude its ability to stimulate the incorporation of 3 H-thymidine into the PBMC. In addition to penicillin (100 U ml31 ) and streptomycin (100 mg ml31 ), 10 U ml31 heparin and 5 Wg ml31 of polymyxin B were added to the RPMI 1640 media. Polymyxin B was added to neutralise any contaminating endotoxin in the protein preparations, although this would not have contributed to the peptides responses. PBMC proliferation was assessed in quadruplicate, in 96-well round bottom microtiter plates, with 200 ml of medium containing 105 PBMC with and without antigen or the mitogen phytohaemagglutinin (PHA). Antigenic stimulation was provided by PPD (3 Wg ml31 ), heat killed M. tuberculosis H37Rv (56³C for 60 min) at a 1:500 dilution (corresponding to 6U106 CFU/well), recombinant proteins

FEMSLE 8893 15-7-99


127

Table 3 Antigen stimulation of IFN-Q production by PBMC Sub BG M. tuberculosis PPD

H37Rv

Peptides Cytochrome d oxidase 1

1 2 3 4 5 6 7 8 9 10

192 201 308 181 270 186 177 149 240 174

4712 5047 2573 *958 1222 2215 4240 5710

7715 7524 3617 6225 4788 4127 6647 1937 7015 2633

CtaC

2

3

Mtb27 4

5

6

7

830 622 *1252

*1252

484

525

*1028

E. coli

Mtb20019 kDa Full

PhoA

PhoA

PhoA

8

11

12

13

9

10

415 3696

2292

502 2408 6639 3640 6031

1283 1642 1140 2608

493

586

463

1385

768 416 *1178

Mtb 898

Sub, subject ; BG, background. Values represent the median IFN-Q stimulation in pg ml31 . With the exception of subject 5, only medians which showed a statistically signi¢cant di¡erence from background of P90.05, when compared using the Kruskal-Wallis ranking, followed by a post hoc Dunn's test are reported. Responses for subject 5, which were greater than 3Ubackground and signi¢cant at P90.004, in a Mann-Whitney Rank Sum test, but not signi¢cant to P90.05 in the multiple comparison are marked with an asterisk.

(0.1 mg ml31 ) or peptides (1 mg ml31 ). PBMC cultures were incubated for 3 or 6 days (for mitogen or antigen, respectively) at 37³C in a 5% CO2 humidi¢ed incubator and 120 Wl of supernatant was removed and frozen at 320³C for the cytokine assay. One WCi of 3 H-thymidine (185 MBq/mmol, Amersham International, Buckinghamshire, UK) was added for the remaining 18 h of culture. PBMC were then harvested and 3 H-thymidine uptake measured by liquid scintillation counting. Proliferative responses were determined as stimulation indices (SI's) i.e. the median incorporation of 3 H-thymidine in wells containing antigen divided by that from wells containing medium and cells alone. 2.8. ELISA for IFN-Q Microtiter wells were coated with 1 mg ml31 biotinylated mouse anti-human IFN-Q antibody (Pharmingen), in 0.1 M NaHCO3 , pH 8.2. The wells were blocked with 3% BSA in PBS, and the samples added. HRP-conjugated streptavidin (0.16 Wg ml31 ), o-phenylenediamine (0.4 mg ml31 ) and H2 O2 (1 Wl ml31 ) were added to visualize reactivity. Plates were read at 450 nm, 20 min after addition of enzyme, OPD and H2 O2 . Samples were assayed in quadruplicate and IFN-Q concentrations determined

by comparison to standard curves (15 pg ml31 to 2000 pg ml31 of IFN-Q, Pharmingen). 2.9. Statistical analysis SigmaStat (Jandel, San Raphael, CA, USA) was used for all statistical analyses. In the proliferation experiment, three sets of quadruplicate background readings for each subject, formed the background control (n = 12) and for IFN-Q stimulation, two sets of background readings (n = 8) were used. Comparisons were made for proliferation and IFN-Q stimulation between background and oligopeptides and between background and all antigens including fusion proteins, peptides, positive control proteins (PPD and heat killed M. tuberculosis H37Rv) and the negative control, E. coli PhoA. Most of the data were not normally distributed. The non-parametric Kruskal-Wallis analysis of variance (ANOVA) ranked medians for a multiple comparison of the di¡erent antigens. As sample sizes were unequal, the extremely conservative post hoc Dunn's test determined statistical signi¢cance. Subject ¢ve displayed a large variation in the size of IFN-Q response to di¡erent antigens, thus medians not signi¢cant in the multiple comparison, but greater than three times background and signi¢cant to

FEMSLE 8893 15-7-99

128


naive subjects (PPD and heat killed M. tuberculosis H37Rv-negative) did not respond to any of the test antigens and none of the subjects showed a signi¢cant response to the negative control, Antigen 13 (E. coli PhoA). The signi¢cant IFN-Q and proliferative responses to the control and test antigens are shown in Tables 3 and 4, respectively. Unfortunately, only a small amount of oligopeptide was available, thus we were unable to do dose response curves, or use antigen concentrations as large as other authors [19,20]. However, by assaying with quadruplicate wells we were con¢dent that statistically signi¢cant immunodominant epitopes would be recognised. In comparison to the oligopeptide corresponding to the 19 kDa antigen, which achieved similar levels of recognition to that found by other authors [19], the oligopeptides corresponding to AppC were recognised by 60%, and the single oligopeptide from Mtb200 (Antigen 8) was recognised by 43% of the positive responders. AppC is predicted to be a membrane associated protein, and Mtb200, a secreted protein [21]. The PhoA fusion Mtb898 (Antigen 12) was predicted to be a lipoprotein and given that many lipoproteins are cleaved during the cell cycle, Mtb898

P 6 0.004 in a Mann-Whitney U test were marked with an asterisk.

3. Results and discussion An anti-tuberculosis subunit vaccine would need to include antigens that can elicit protective T-cell responses from people of diverse genetic backgrounds and it is important to identify a range of antigens with this potential. In this study, we identi¢ed four proteins with immunostimulatory activity Mtb92 (CtaC), Mtb827 (AppC), and hypothetical proteins Mtb898 (MTV005.02) and Mtb200 (MTV033.08). Two peptides tested from a ¢fth hypothetical protein Mtb27 (MTV023.04c), were not found to be immunostimulatory. The PBMC donors were screened and classi¢ed as to their PPD proliferative responses, rather than the extent of skin reactivity. Given the endemic nature of tuberculosis in South Africa, together with a policy of BCG vaccination, we were unable to recruit su¤cient numbers of completely naive subjects. Despite this, it is likely that the immunostimulatory antigens are mycobacteria speci¢c, since the two completely Table 4 Relative stimulation index for incorporation of 3 H-thymidine Sub

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

M. tuberculosis

Peptides

PPD

Rv

Cytochrome d oxidase 1

2

284 51 20 118

298 47 65 200 122 160 119 188 185 130 12

9 23 4 30 52

6

5

52

21

18

27 95

95 33 197 149 215

CtaC

3

Mtb27 4

5

6

13 16 5 29 40

70

21

23

7

7 39

Mtb898 E. coli

Mtb200 19 kDa

Full

PhoA

PhoA

PhoA

8

10

11

12

13

9 21

8 17 25 71 82

15

60

10 16 13

8 22

8

49

5

5 23

6

9

12 31 109

82 36

5

4 6

5

45 32 106 151 149 25 32

9

Sub, subject ; Rv, heat killed M. tuberculosis H37Rv. All SI's reported were calculated from medians which showed a statistically signi¢cant di¡erence from background when ranked using the Kruskal-Wallis multiple comparison, followed by a post hoc Dunn's test. Only medians which were signi¢cant in the Dunn's test to P 6 0.05 and had an SI s 3, were included.

FEMSLE 8893 15-7-99


may gain access to the macrophage cytosol and enter the MHC I processing pathway in a manner analogous to the secreted proteins. Interestingly, almost all the positive responders recognised the CtaC containing antigens. Both CtaC and AppC are terminal oxidases with highly conserved functions and structures. CtaC is required for vegetative growth and AppC for growth in oxygen-limited conditions. Generally, PPD-positive donors mounted greater responses to heat killed M. tuberculosis H37Rv than PPD. A similar pattern has been observed in other skin test-negative donors and in patients with clinical tuberculosis (S.S.R., unpublished observations). This may re£ect the larger repertoire of antigens in the preparation of heat killed M. tuberculosis H37Rv and suggests that a lack of PPD reactivity may be inadequate to categorise naive subjects in an endemic area. Finally, this study demonstrated that immunological screening of randomly selected, extracytoplasmically expressed M. tuberculosis-PhoA fusion proteins in a murine model of tuberculosis was able to identify proteins recognised by the PBMC from human M. tuberculosis-immune donors. Four antigens from this study could be included in future work to identify protective epitopes or diagnostics and the availability of genomic sequence should enable cloning of complete ORFs from these and other potentially immunoreactive, secreted or exposed proteins.

Acknowledgements This work was supported by grants to F.E.N. from the Glaxo-Wellcome Action TB Initiative, the Medical Research Council of Canada (PA12992), the Canadian Bacterial Diseases Network, the South African Medical Research Council and the British Columbia Lung Association. Work in the laboratory of I.M.O. was supported by N.I.H. grant AI75320. Thanks to Dr. James Sparling, the sta¡s at the Health Region Tuberculosis Clinic (Victoria) and MDS-Metro Laboratories for assistance in recruiting volunteers and drawing blood for preliminary studies. We also thank Christine Dunstan, Phillip Schellenberg, John Padoba and Marcia Watkins for their contributions, Jonathan Moran for help with statis-

129

tical analysis and Vladmir Brucic for help with MHCPEP.

References [1] Bloom, B.R. and Fine, P.E.M. (1994) The BCG experience : implications for future vaccines against tuberculosis. In: Tuberculosis Pathogenesis, Protection, and Control (Bloom, B.R., Ed.), pp. 531^557. ASM Press, Washington, DC. [2] Andersen, P. (1994) E¡ective vaccination of mice against Mycobacterium tuberculosis infection with a soluble mixture of secreted mycobacterial proteins. Infect. Immun. 62, 2536^ 2544. [3] Horwitz, M.A., Lee, B.-W., Dillon, B.J. and Harth, G. (1995) Protective immunity against tuberculosis induced by vaccination with major extracellular proteins of Mycobacterium tuberculosis. Proc. Natl. Acad. Sci. USA 92, 1530^1534. [4] Collins, F.M., Lamb, J.R. and Young, D.B. (1988) Biological activity of protein antigens isolated from Mycobacterium tuberculosis culture ¢ltrate. Infect. Immun. 56, 1260^1266. [5] Orme, I.M. (1987) The kinetics of emergence and loss of mediator T lymphocytes acquired in response to infection with Mycobacterium tuberculosis. J. Immunol. 138, 293^298. [6] Boesen, H., Jensen, B.N., Wilcke, T. and Andersen, P. (1995) Human T-cell responses to secreted antigen fractions of Mycobacterium tuberculosis. Infect. Immun. 63, 1491^1497. [7] Lim, E.M., Rauzier, J., Timm, J., Torrea, G., Murray, A., Gicquel, B. and Portnoi, D. (1995) Identi¢cation of Mycobacterium tuberculosis DNA sequences encoding exported proteins by using phoA gene fusions. J. Bacteriol. 177, 59^ 65. [8] Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) Molecular Cloning : A Laboratory Manual, 2nd Edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. [9] Strauch, K.L. and Beckwith, J. (1988) An Escherichia coli mutation preventing degradation of abnormal periplasmic proteins. Proc. Natl. Acad. Sci. USA 85, 1576^1580. [10] Yanisch-Perron, C., Vieira, J. and Messing, J. (1985) Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mp18 and pUC19 vectors. Gene 33, 103^119. [11] Mdluli, K.E., Treit, J.D., Kerr, V.J. and Nano, F.E. (1995) New vectors for the in vitro generation of alkaline phosphatase fusions to proteins encoded by G+C-rich DNA. Gene 155, 133^134. [12] Orme, I.M., Roberts, A.D., Gri¤n, J.P. and Abrams, J.S. (1993) Cytokine secretion by CD4 T lymmphocytes acquired in response to Mycobacterium tuberculosis infection. J. Immunol. 151, 518^525. [13] Altschul, S., Madden, T., Scha¡er, A., Zhang, J., Zhang, Z., Miller, W. and Lipman, D.J. (1997) GappedBLAST and PSIBLAST : a new generation of protein database search programs. Nucleic Acids Res. 25, 3389^3402. [14] Laemmli, U.K. (1970) Cleavage of structural proteins during

FEMSLE 8893 15-7-99

130


the assembly of the head of bacteriophage T4. Nature (London) 227, 680^685. [15] Feller, D.C. and de la Cruz, V.F. (1991) Identifying antigenic T-cell sites. Nature (London) 349, 720^721. [16] Jacobs, J., Kalpana, G.V., Cirillo, J.D., Pascopella, L., Snapper, S.B., Udani, R.A., Jones, W., Barletta, R.G. and Bloom, B.R. (1991) Genetic systems for mycobacteria. In: Methods in Enzymology, Vol. 204, pp. 537^555. Academic Press, New York. [17] Cole, S., Brosch, R., Parkhill, J., Garnier, T., Churcher, C., Harris, D., Gordon, S., Eiglmeier, K., Gas, S., Barry 3rd, C.E., Tekaia, F., Badcock, K., Basham, D., Brown, D., Chillingworth, T., Connor, R., Davies, R., Devlin, K., Feltwell, T., Gentles, S., Hamlin, N., Holroyd, S., Hornsby, T., Jagels, K. and Barrell, B. (1998) Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature (London) 393, 537^544.

[18] Brucic, V., Rudy, G. and Harrison, L.C. (1998) MHCPEP, a database of MHC-binding peptides : update 1997. Nucleic Acids Res. 26, 368^371. [19] de Groot, A., Roberts, C., Edelson, B. et al. (1996) A novel algorithm for the e¤cient identi¢cation of T-cell epitopes : Prediction and testing of candidate tuberculosis vaccine peptides in genetically diverse populations. In : Vaccines 96 (Brown, F., Ed.), pp. 127^134. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. [20] Silver, R.F., Wallis, R.S. and Ellner, J.J. (1995) Mapping of T cell epitopes of the 30-kDa alpha antigen of Mycobacterium bovis strain bacillus Calmette-Guerin in puri¢ed protein derivative (PPD)-positive individuals. J. Immunol. 154, 4665^4774. [21] Neilson, H., Engelbrecht, J., Brunak, S. and von Heijne, G. (1997) Identi¢cation of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Eng. 138, 293^298.

FEMSLE 8893 15-7-99