Attenuation of Virulence by Disruption of the Mycobacterium tuberculosis erp Gene

See allHide authors and affiliations

Science  23 Oct 1998:
Vol. 282, Issue 5389, pp. 759-762
DOI: 10.1126/science.282.5389.759


The virulence of the mycobacteria that cause tuberculosis depends on their ability to multiply in mammalian hosts. Disruption of the bacterial erp gene, which encodes the exported repetitive protein, impaired multiplication of M. tuberculosis andM. bovis Bacille Calmette-Guérin in cultured macrophages and mice. Reintroduction of erp into the mutants restored their ability to multiply. These results indicate thaterp contributes to the virulence of M. tuberculosis.

Mycobacterium tuberculosis, M. bovis, andM. africanum cause tuberculosis, an infectious disease killing more than 3 million people per year worldwide (1). These bacteria constitute the M. tuberculosis complex (MTC), which also includes the tuberculosis vaccine strainM. bovis Bacille Calmette-Guérin (BCG) and the murine pathogen M. microti. Members of the MTC multiply within phagocytic cells in a specialized vacuolar compartment called the phagosome (2). Phagosomes containing mycobacteria do not acidify (3), and they escape fusion with lysosomes (4). Remodeling of the phagosome architecture by these pathogenic bacteria is thought to be critical for their intracellular multiplication and survival (5,6).

Mycobacterial proteins that are exported intracellularly are likely to participate in phagosome remodeling. Using a genetic approach to identify such exported proteins, we recently identified the M. tuberculosis erp gene (7). It encodes the exported repetitive protein (ERP), a protein previously identified in M. bovis as the PGLTS secreted antigen (8). Although no function has been ascribed to ERP, sequences homologous to erp are found exclusively in mycobacteria causing tuberculosis (7, 8) and leprosy (9).

To characterize M. tuberculosis ERP, we overproduced it inEscherichia coli as a fusion protein containing six histidines at the COOH terminus (ERP-6His) (10). ERP-6His formed cytoplasmic inclusion bodies and was purified by nickel-affinity chromatography under denaturing conditions (11). Renatured soluble ERP-6His (36 kD) was used to immunize rabbits, and high-titer anti-ERP polyclonal sera detected immunoreactive bands both in cell-associated fractions (36 kD) and trichloroacetic acid–precipitated culture filtrates (36 and 34 kD) of BCG and M. tuberculosis (Fig. 1A).

Figure 1

Expression, immunodetection, and localization of ERP. (A) Immunodetection of ERP in cell-associated (lanes 1 through 6) or concentrated supernatant (lanes 7 through 12) fractions of M. bovis BCG (lanes 1 through 3 and 7 through 9) andM. tuberculosis (lanes 4 through 6 and 10 through 12). Protein extracts were prepared from parental (lanes 1, 4, 7, and 10), mutant (lanes 3, 5, 8, and 11), and complemented (lanes 2, 6, 9, and 12) strains. (B through D) Surface analysis of parental (B), mutant (C), or complemented (D) M. bovis BCG cells processed for immunogold labeling with anti-ERP serum. (E and F) Immundetection of ERP in M. tuberculosis–infected, J774 murine macrophage sections incubated with anti-ERP serum previously adsorbed against purified ERP-6HIS (E) or anti-ERP sera (F); arrowheads in (F) show ERP labeling of an intraphagosomal mycobacterium (bottom) and a small cellular vesicle (top). Scale bars in (B) through (F), 200 nm.

To determine the subcellular localization of ERP, we performed immunocytochemical analyses of M. bovis BCG bacilli using rabbit antisera to ERP and a gold-labeled anti-rabbit conjugate (12, 13). Transmission electron microscopy revealed intense surface labeling at the periphery of the bacilli (127 ± 30 gold grains per bacterium), which indicates that ERP is a surface-exposed molecule (Fig. 1B). We next investigated whether ERP was produced during intracellular multiplication of M. tuberculosis within cultured macrophages. J774 murine macrophages were infected with a clinical isolate of M. tuberculosis and processed for immunoelectron microscopy (12, 14). Specific labeling of the mycobacterial cell wall and the phagosomal lumen was observed with sera from ERP-immunized rabbits (Fig. 1F) but not with anti-ERP serum previously adsorbed against purified ERP-6His (Fig. 1E). Furthermore, small vesicles near the phagosomes also contained the ERP label. These findings demonstrate that ERP is produced in M. tuberculosis phagosomes, and they suggest that it may traffic intracellularly.

We then examined whether ERP is essential for intracellular growth of the mycobacteria. A targeted null mutation was introduced at the erp locus in M. tuberculosis H37Rv and in the model tuberculosis vaccine strain M. bovis BCG. A sucrose-counterselectable suicide vector (pJQ2OO) was used to deliver a mutated allele of erp(erp::aph) into the chromosome of M. bovis BCG (15, 16). The corresponding M. tuberculosis mutant strain was constructed with the ts-sacB technology (17). Mutant strains resulting from allelic exchange (Fig. 2A) were referred to as BCGerp::aph and H37Rv erp::aph. By using a mycobacteriophage MS6-derived integrative vector pIPX70, we reintroduced a single copy of erp at the attBsite in BCG erp::aph and H37Rverp::aph (Fig. 2A). The genotype of these strains was analyzed by Southern (DNA) blotting with an erp-specific DNA probe. This analysis revealed that there was an insertion of 1.3 kb in the restriction fragment carrying erp and that an additional copy of erp was present in the chromosome of complemented strains (Fig. 2B). Analysis of cell-associated fractions and concentrated supernatants (Fig. 1A) from BCGerp::aph and H37Rv erp::aphcultures indicated that disruption of erp abolished the production of ERP. This was confirmed by the disappearance of immunogold labeling (32 ± 10 gold grains per bacterium) onM. bovis BCG erp::aph bacilli (Fig. 1C). In contrast, integration of erp at the Ms6attB site of erp::aph mutant strains restored the production of ERP, both at the cell surface (Fig. 1, A and D) (112 ± 57 gold grains per bacterium) and in the culture medium (Fig. 1A) of M. tuberculosis and M. bovis BCG. The parental, mutant, and complemented strains of BCG and H37Rv showed similar colony morphology, doubling time, and growth characteristics in Middelbrook 7H9/ADC broth or minimal Sauton medium. Taken together, these data show that erp is not essential for the growth of BCG and H37Rv under axenic conditions.

Figure 2

Disruption of theerp gene. (A) Chromosomal organization of theerp gene in the parental, mutant, and complemented strains of M. bovis BCG and M tuberculosis. Gene symbols are as follows: glf (UDP galactomutase), csp (secreted protein),ppe (repetitive protein of the PPE family),aph (aminoglycoside phosphotransferase), andattB (bacterial chromosome attachment site for the Ms6 mycobacteriophage). (B) Southern blot analysis of chromosomal DNA from the parental, mutant, and complemented strains of M. bovis BCG (B) and M. tuberculosis (T). The erp internal probe used for hybridization is depicted in (A).

To determine whether BCG erp::aph and H37Rverp::aph grow within phagocytes, we compared the multiplication capacity of mutant and parental strains in cultured bone marrow–derived macrophages (18). Analysis of colony-forming units (CFUs) indicated that multiplication oferp::aph mutants was impaired within murine macrophages, whereas parental and complemented strains grew normally (Fig. 3). Moreover, the H37RVerp::aph strain showed reduced cytopathic effects as compared to the parental or complemented strains. To determine whether the erp::aph mutation also affected multiplication within the host, we analyzed the persistence of BCGerp::aph and H37Rv erp::aphin mice. BALB/c mice were injected intravenously with 106viable units of parental, erp::aph mutant, anderp-complemented strains, and bacterial infection was monitored by counting CFUs over a 56-day period (19). We analyzed the lungs, liver, and spleen, which are the three organs known to contain most of the mycobacterial load after intravenous inoculation. The BCG erp::aph mutant was rapidly cleared from the lungs of infected animals, whereas the parental and complemented strains colonized and survived within this organ (Fig. 4A). In contrast, the H37Rverp::aph mutant survived but multiplied very slowly in the lungs as compared to the parental and complemented strains (Fig. 4A). The lungs are the primary site of infection in tuberculosis. Multiplication of erp::aph mutants was also greatly impaired in the spleen (Fig. 4B) and the liver. In addition, whereas parental BCG showed a “spreading” colony morphology after animal passaging, the mutant no longer spread and exhibited retarded growth (by 1 week as compared to the parental strain) (Fig. 4C). The significance of this observation is unknown, but a loss of the spreading phenotype has been correlated with the lowest levels of residual virulence among BCG substrains (20). The non-spreading phenotype we observed was lost after restreaking of the mutant on 7H11 plates, and reintroduction of erprestored the parental phenotype.

Figure 3

Effect of theerp::aph mutation on the intracellular multiplication of M. bovis BCG and M. tuberculosis. CFU counts in bone marrow–derived murine macrophages infected with the parental (squares), mutant (diamonds), or complemented (circles) strains are shown. Results are expressed as means and standard deviations of four different measurements.

Figure 4

Effect of the erp::aphmutation on the survival and multiplication of M. bovis BCG and M. tuberculosis in the lungs (A) and the spleen (B) of BALB/c mice. CFU counts are shown for the parental (squares), mutant (diamonds), and complemented (circles) strains. The results for each time point are the means and standard deviations of CFU counting performed on five mice. They axis reflects the number of viable bacteria recovered from a specific organ. (C) Morphologic differences between parental (erp), mutant (erp::aph), and complemented (erp::aph, attB::erp) BCG colonies recovered from the liver of infected mice at day 7. Bacteria were grown for 20 days (parental and complemented strains) or 27 days (mutant strain) on 7H11 plates, and the resultant colonies were photographed at the same scale (scale bar, 1 cm).

We have shown that the M. tuberculosis erp gene encodes a surface-exposed protein produced during phagosomal growth and that this protein is required for multiplication of the pathogen within the host cell. Thus, erp may be a good candidate for the rational attenuation of M. tuberculosis and the construction of new live-attenuated vaccines against tuberculosis. Indeed, the multiplication characteristics of H37Rv erp::aphin vivo are comparable to that observed for the tuberculosis vaccine strain M. bovis BCG. Sensitization of guinea pigs with mycobacterial purified protein derivative 1 month after immunization with BCG erp::aph resulted in a delayed-type hypersensitivity response comparable to that induced by the parental strain. This observation suggests that BCGerp::aph persists long enough to stimulate antimycobacterial immunity. Although the function of ERP is still unclear, this work opens new avenues for studying the pathogenetic mechanisms of the mycobacteria causing tuberculosis.

  • * Present address: SmithKline Beecham Biologicals, Laboratory of Molecular Bacteriology, Rue de l'Institut 89, B-1330 Rixensart, Belgium.

  • Present address: Department of Biological Sciences, University of Padova, Via Trieste 75, 35121 Padova, Italy.

  • To whom correspondence should be addressed. E-mail: bgicquel{at}


View Abstract

Navigate This Article