The Hydrophobic Domain of the Mycobacterial Erp Protein Is Not Essential for the Virulence of Mycobacterium tuberculosis

Construction of the C-terminal deletion mutant.

In the course of a previous study, a spontaneous mutant lacking the most hydrophobic region of the carboxy-terminal domain was serendipitously obtained (Fig. 1A). Indeed, construct pML101, containing the M. tuberculosis erp::aph gene (lacking its C-terminal portion) in pNIP40b, was generated by amplifying M. tuberculosis genomic DNA with oligonucleotide primers erp-C1 and erp-C2 (18). Purified PCR product digested with XbaI was ligated into the XbaI site of dephosphorylated pNIP40b (16). Sequencing of pML101 revealed the deletion of an A residue at position 631, leading to a frameshift mutation and a premature stop codon. The corresponding protein lacks the last 74 amino acids, corresponding to a molecular mass of approximately 8 kDa. The M. tuberculosis erp::aph mutant strain (3) was transformed by either pML101 or pML10, a vector carrying the full-length erp gene (3), and selected on Middlebrook 7H11 (Difco) agar plates containing kanamycin (20 μg ml⁻¹) and hygromycin (50 μg ml⁻¹) at 37°C; these complemented strains are named M. tuberculosis erp::aph/TBΔC and M. tuberculosis erp::aph/TB, respectively.

• Open in new tab
• Download powerpoint

FIG. 1.

(A) Kyte-Doolittle plot of the of the erp gene product of M. tuberculosis. The deletion is indicated by an arrow. All data respecting protein features and sequences, together with the protein alignment, can be found at http://mic.sgmjournals.org/cgi/content/full/147/8/2315/DC1 . (B) Denaturing sodium dodecyl sulfate-polyacrylamide gel electrophoresis immunoblot, showing the TB-Erp and TB-ErpΔC protein present in the supernatant fraction of the M. tuberculosis erp::aph/TB and M. tuberculosis erp::aph/TBΔC strains. rErp is the purified recombinant TB-Erp (30 ng) and was used as an internal control.

These strains were grown in Middlebrook 7H9 medium (Difco) supplemented with 0.05% (vol/vol) Tween 80 and 10% ADC (Difco). Antibiotics were added as required. To quantify the expression of the complemented strains, 50 μg of total proteins from supernatant fraction was probed using an anti-TB Erp antiserum. Proteins were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis using 15% polyacrylamide gels (14) and transferred to Hybond P membranes (Amersham Pharmacia). A rabbit TB-Erp antiserum (3) was used as first antibody at a 1:5,000 dilution, and bound antibodies were revealed as previously described (18). No cross-reacting band was observed in the M. tuberculosis erp::aph mutant, whereas a clear signal of a similar intensity was observed in the case of M. tuberculosis erp::aph/TB and M. tuberculosis erp::aph/TBΔC, demonstrating that the ErpTBΔC protein is expressed stably and at a similar level (Fig. 1B). The reduction in size (8 kDa) was in agreement with the calculated mass of ErpTBΔC.

In vivo analysis.

The erp::aph mutant of M. tuberculosis is very severely hampered with respect to multiplication in vivo (3). Indeed, infection with the erp mutant of M. tuberculosis leads to a less that 1% bacterial M. tuberculosis erp::aph/TB load in the lung after long periods of time (3, 18). The virulence of the complemented M. tuberculosis strains was assayed using the mouse model of tuberculosis.

Mice (n = 5) were infected intravenously with 10⁵ CFU of either the erp::aph mutant or complemented strains in a total volume of 0.5 ml as previously described (12) and maintained in specific pathogen-free containment. Mice were sacrificed 1, 21, 42, or 80 days after infection. The course of the infection was monitored over time by homogenizing the lung and spleen and counting viable bacilli, as described elsewhere (6). On day 1, similar numbers of CFU from the spleen and for the lung were obtained for all strains, confirming that the inoculum sizes did not differ between the various complemented strains. As reported previously, the M. tuberculosis erp::aph mutant was severely attenuated, leading to a bacterial M. tuberculosis erp::aph/TB load in the spleen of about 4% and a load in the lung of 0.6% on day 42 (3). On days 21, 42, and 80, the bacterial loads observed in the spleen and in the lung for M. tuberculosis strains erp::aph/TB and M. tuberculosis erp::aph/TBΔC did not differ significantly, demonstrating that these alleles had similar functional roles in this context (Fig. 2A).

• Open in new tab
• Download powerpoint

• Open in new tab
• Download powerpoint

FIG. 2.

(A) Total CFU counts of M. tuberculosis erp::aph, M. tuberculosis erp::aph/TB, and M. tuberculosis erp::aph/TBΔC mutantstrains in the spleens and lungs of infected animals at various time points. Data are expressed as the log₁₀ value of the mean number of bacteria ± standard deviation recovered from each mouse (n = 5). Two independent experiments were carried out. (B) Representative photomicrograph of lung tissue sections harvested on day 80 from mice infected with the M. tuberculosis erp::aph/TB (panel 1), M. tuberculosis erp::aph (panel 2), or M. tuberculosis erp::aph/TBΔC (panel 3) mutant strain and from naïve mice (panel 4). Arrows, cellular recruitment and infiltration; stars, foamy macrophage; a, alveolus; v, blood vessel. All plates are shown at ×100 magnification. The qualitative histological analysis was repeated twice independently.

Histological analysis.

As CFU counts and histology analysis can sometimes produce discrepant results (13, 18), a histology analysis was undertaken to evaluate the damage caused by these three strains. The right superior, median, and inferior pulmonary lobes of the right lungs of three additional mice were embedded in paraffin, stained with hematoxylin and eosin, and examined under a light microscope as previously described (21). Sections were evaluated by three individuals, including a veterinary pathologist blinded to the time point and strain considered. On day 42 (data not shown), both infiltrations (which consist of lymphocytes combined with cells derived from the phagocyte mononuclear system that invade the lung tissue and that are not organized in a granuloma) and granulomas with foamy macrophages were observed in the lungs of mice infected with M. tuberculosis erp::aph/TB. In mice infected with the M. tuberculosis erp::aph mutant, little perivascular and peribronchial infiltration was observed. M. tuberculosis erp::aph/TBΔC infection resulted in damage compared to the results seen with M. tuberculosis erp::aph/TB infection, with both infiltration and granulomas with foamy macrophages appearing. The situation was similar on day 80 (Fig. 2B). These data demonstrate that the C-terminal domain of the Erp protein is not necessary for either organ multiplication or persistence or for induction of lung damage.

Despite intensive efforts such as genome sequencing and genetic tool development, the mechanism of virulence of the tubercle bacilli remains poorly characterized (11, 20). Some virulence factors have been characterized through the use of signature tag mutagenesis (5, 8), and several studies concerning the virulence factors of M. tuberculosis pinpointed the cell wall as a major player in this process. The surface of mycobacteria is a very complex structure composed of proteins, sugars, and lipids, some of which have been shown to play a role in virulence, as shown by studies of the fbpA gene, which encodes a mycoloyl transferase (1), the pcaA gene, which encodes a protein involved in mycolic acid cyclopropane ring synthesis (10), and the phthiocerol dimycocerosate locus, which is involved in the synthesis of lipid of the cell envelope (6).

In similarity to the results seen with gram-negative bacteria, this outer membrane contains porins and delineates a pseudo periplasm (4). Erp, which is specific to mycobacteria, is one of the numerous actors for the cell wall and likely plays an important function in its structure. Both the amino- and the carboxy-terminal domains are conserved between species, suggesting an important role for these domains.

In this study, we have shown that the most hydrophobic part of the carboxy-terminal domain is not essential for animal infection. Interestingly, this carboxy-terminal domain has been conserved through evolution even in M. leprae, in which a dramatic proteome reduction has been observed (7). It is remarkable that a region that represents more than one-fourth of the total protein length and that has been conserved through speciation appears to be dispensable for the virulence property of the tubercle bacillus. One may speculate that this region interacts with some uncharacterized cell wall products required in some situations other than virulence, such as desiccation or long-time persistence. Another hypothesis is that this hydrophobic domain is required in other mycobacterial species and that it is a relic of evolution in M. tuberculosis. Our current view is that Erp is a protein present at the surface of the bacillus. The most important domains are clearly the amino-terminal domain, which is required for secretion, and the central region, which influences disease outcome (2, 18). We expect that it will be an interesting challenge to characterize the molecular function of this surface protein.