OUP user menu

Disruption of the mycobacterial cell entry gene of Mycobacterium bovis BCG results in a mutant that exhibits a reduced invasiveness for epithelial cells

Bruno Flesselles , Naveen N. Anand , Jack Remani , Sheena M. Loosmore , Michel H. Klein
DOI: http://dx.doi.org/10.1111/j.1574-6968.1999.tb13738.x 237-242 First published online: 1 August 1999


Mycobacteria belonging to the Mycobacterium tuberculosis complex have the ability to invade and replicate in non-phagocytic cells, an event that requires the presence of bacterial surface components capable of triggering a cell response and the subsequent internalization of the microorganism. In this study, we report the sequencing of the mycobacterial cell entry gene (mce) of Mycobacterium bovis bacillus Calmette-Guérin (BCG) and the generation and characterization of a mutant BCG strain with an inactivated mce gene, by homologous recombination with double cross-over. This mutant strain does not express the mycobacterial cell entry protein (Mce) and exhibits a reduced ability to invade the non-phagocytic epithelial cell line HeLa as compared to wild-type BCG.

  • Bacillus Calmette-Guérin
  • Homologous recombination
  • Cell invasion

1 Introduction

In addition to invading and multiplying in professional phagocytic cells, members of the Mycobacterium tuberculosis complex can also enter non-phagocytic cells such as HeLa or Hep-2 cells, in culture, and replicate intracellularly [1,2]. Although the in vivo relevance of this latter phenomenon is not clear, it does imply that M. tuberculosis produces factors which promote its entry into mammalian cells, as was suggested with the cloning of a DNA molecule, which conferred invasiveness to a non-pathogenic strain of Escherichia coli [3].

Recently, the construction by homologous recombination of mutants deficient in some metabolic genes has been achieved in slow growing mycobacteria [46] and improvements in recombination techniques, along with the isolation of genes presumed to encode proteins involved in the survival of M. tuberculosis in macrophages, has led to the generation of M. tuberculosis strains that exhibit an attenuated phenotype [7,8].

In this study, we describe the sequencing of the mycobacterial cell entry gene (mce gene) from the attenuated strain Mycobacterium bovis bacillus Calmette-Guérin (BCG), as well as the generation of a knock-out mutant for the mce gene, engineered by homologous recombination with double cross-over, that exhibits a reduced ability to invade non-phagocytic cells.

2 Materials and methods

2.1 Bacterial strains, cells and growth conditions

E. coli HB101 used for cloning and plasmid propagation purposes was grown on solid or in liquid Luria-Bertani medium, containing 100 µg ml−1 ampicillin and/or 200 µg ml−1 hygromycin B. BCG Connaught was grown either in Middlebrook 7H9 supplemented with albumin dextrose complex (ADC) and 0.1% Tween 80 broth or on Middlebrook 7H10 solid medium. When necessary, hygromycin B was added to a final concentration of 50 µg ml−1. HeLa cells were grown in DMEM supplemented with 2 mM l-glutamine and 10% fetal bovine serum (cDMEM medium).

2.2 DNA techniques

BCG genomic DNA was prepared as described [9]. PCR amplification reactions were carried out on BCG chromosomal DNA, using primers P4414 (5′-GTATGTGTCGTTGACCACGCC-3′) and P4448 (5′-TCAGGTCGATCGGCATCGTAG-AAG-3′). PCR conditions were as follows: 2 min at 99°C, then 25 cycles of 45 s at 98°C, 45 s at 60°C, 90 s at 72°C, ending with a 10-min extension period at 72°C, in the presence of TsgPlus DNA polymerase (BioBasics). To screen BCG transformants, PCR experiments were directly performed on bacterial lysates obtained by resuspending individual colonies in water, boiling the suspensions for 10 min and using the same conditions.

BCG genomic DNA (8 µg) was digested with restriction enzymes and DNA fragments were separated by electrophoresis on a 0.8% agarose gel. These fragments were subsequently transferred to a nylon membrane (GeneScreen Plus, NEN Research), using a VacuGene apparatus (Pharmacia). Prehybridization, hybridization and washing of the blot under high-stringency conditions were carried out according to standard procedures [10]. DNA fragments used as probes to detect the mce gene were labelled with non-radioactive DIG-dUTP (Boehringer Mannheim) and the blots were processed according to the manufacturer's instructions.

Sequences of double-stranded plasmid DNA were determined on a ABI model 370A sequencer, using dye-terminator chemistry. The nucleotide sequence of the mce gene from BCG has been deposited at GenBank and assigned accession number AF113402.

2.3 Construction of the plasmid containing the disrupted mce gene

The hygromycin-resistance gene (hyg) of Streptomyces hygroscopicus [4] was isolated from plasmid pIDV6 (a gift from Dr M. Horwitz, UCLA, Los Angeles, CA, USA) as a 1.3-kb BspHI-NotI fragment. It was blunt-ended and inserted into the blunt-ended BsiWI site of the mce gene, contained in plasmid pBCGmceX, to yield plasmid pBCGmceX-H. The resulting plasmid was linearized with ApaI and electroporated into BCG as previously described [4].

2.4 Immunoblot analysis

4-Day cultures of BCG were harvested, resuspended in water, submitted to two 30-s sonication cycles in a Sonifer 250 sonicator (Branson, Danbury, CT, USA) and mixed with 4×loading buffer (0.1 M Tris-HCl, pH 6.8, 20% glycerol, 8% SDS, 8 M urea, 8%β-mercaptoethanol, trace of bromophenol blue). An aliquot of the mixture was boiled, resolved on a 12.5% acrylamide gel and transferred to a membrane as described [10]. A mouse monoclonal antibody raised against recombinant Mce expressed in E. coli was used at a concentration of 1 µg ml−1 to detect the Mce protein. An anti-mouse IgG horseradish peroxidase conjugate (Boehringer Mannheim) was used as secondary antibody to process the blot according to the supplier's recommendations.

2.5 Invasion assay

HeLa cells (105) were seeded in a 24-well plate, 24 h before infection. Bacterial samples (10–20×105 bacteria per well) were added to HeLa monolayers and the tissue culture plates were incubated for 2 or 6 h at 37°C in a 5% CO2 incubator. After washing the monolayers three times with Hanks’ balanced salt solution (HBSS), 1 ml of cDMEM containing 100 µg ml−1 of the bactericidal antibiotic amikacin was added to kill extracellular bacteria and the plates were further incubated for 2 h at 37°C in a 5% CO2 atmosphere. Subsequently, after three washes with HBSS, the viable intracellular bacteria were released by lysis of the monolayers with sterile water containing 1% Tween 80 and quantitated by plating serial dilutions onto Middlebrook 7H10 agar plates. Four separate experiments, each done in triplicate, were performed. Statistical analysis of the data was performed using the Student t-test.

3 Results and discussion

3.1 Cloning and sequencing of the mce gene from M. bovis BCG, M. tuberculosis H37Rv and M. tuberculosis H37Ra

PCR amplification of M. bovis BCG chromosomal DNA with primers P4414 and P4448, derived from the open reading frame (ORF) 1 sequence (181–810 bp) of plasmid pZX7 described by Arruda et al. [3], yielded a 572-bp amplified fragment which was used as a probe to analyze M. bovis BCG chromosomal DNA digested with XhoI or SacI by Southern blotting. This probe hybridized to a 5.2-kb SacI fragment and a 4.7-kb XhoI fragment (Fig. 1, lanes 1 and 4). The 4.7-kb XhoI fragment was isolated and cloned into plasmid pBluescript SK+ to give plasmid pBCGmceX, which was sequenced. A 1581-bp ORF containing the mce gene was identified by DNA sequence analysis (GenBank accession number AF113402). The gene sequence was similar to the published ones of M. tuberculosis H37Rv and H37Ra, except for two non-synonymous nucleotide changes, which resulted in two non-conservative substitutions in the protein, S386 for A386 and P432 for S432.


Southern blot analysis. Genomic DNAs from wild-type BCG (lanes 1 and 3) and BCGmce mutant BCG-65 (lanes 2 and 4) were digested by SacI (lanes 1 and 2) or XhoI (lanes 3 and 4) and probed with the labelled PMCE fragment containing the mce gene. Molecular size markers (in kb) are indicated on the left.

Analysis of the genome of M. tuberculosis [11] revealed the presence of the mce gene as being the third gene of an eight-gene operon and suggested that the start codon of the protein is likely located 220 nucleotides downstream of the first ATG codon, leading to a protein with a predicted molecular mass of 47.7 kDa, in agreement with the observed mass of the Mce protein (Fig. 2, lane 1).


Immunoblot analysis. The whole cell lysates of BCG wild-type (lane 1) and BCGmce mutant BCG-65 (lane 2) were analyzed for the presence of the Mce protein using a monoclonal antibody against recombinant Mce. Molecular size markers (in kDa) are indicated on the left.

Analysis of the genome sequence of M. tuberculosis indicates that there are four mce gene homologues, found in operons containing eight genes, the mce gene described in this study corresponding to mce1 [11]. Although the nucleotide sequences of the genes do not exhibit a very high homology (32% for mce2 with small (<150 bp) stretches with 85% homology, and not significant for mce3 and mce4), these genes have been identified as homologues by Cole et al. [11] and named accordingly, because of the exact same organization of the mce operons and a similar low degree of homology for all the other genes of these operons. The proteins encoded by these genes exhibit little similarity to Mce1 (64% for Mce2 with only small stretches of consecutive identical residues, 30% for Mce3 or Mce4). As reported by Cole et al. [11], these operons encode sets of proteins all predicted to contain signal sequences or hydrophobic residues at their N-terminus, suggesting that they may be secreted proteins or be expressed on the surface of the bacteria, in keeping with the potential role of Mce. Preliminary analysis of the BCG genome has indicated a similar structure for the mce operon in this organism, with proteins similar to that of M. tuberculosis. The exact function of these proteins is currently unknown.

3.2 Generation and characterization of a mce-deficient BCG mutant

Plasmid pBCGmceX-H containing the mce gene disrupted with the hyg gene (Fig. 3) was linearized with ApaI and electroporated into M. bovis BCG. Since this plasmid does not contain a mycobacterial origin of replication, hygromycin-resistant transformants could only result from integration of the plasmid into the mycobacterial genome.


Schematic organization of the disrupted mce gene used for allelic exchange. The mce open reading frame is represented by a gray box. Relevant restriction sites are indicated (X, XhoI; S, SacI; A, ApaI). The hygromycin-resistance gene (hyg) on a 1.3-kb fragment was inserted into the BsiWI site of the mce gene and is represented by a black rectangle. The solid bar represents the vector pBluescript SK+. Primers used for the amplification by PCR of the structure resulting from double cross-over are depicted by arrows P4414 and P4448. The PMCE probe used for Southern hybridization is represented by a black line.

Hygromycin-resistant transformants were screened by PCR using oligonucleotides P4414 and P4448 as primers. This set of primers would generate a 572-bp PCR product from the wild-type strain, while integration by homologous recombination with double cross-over would yield a 1.9-kb product. Both fragments would be amplified following random or single cross-over DNA integration. Three mutants out of 88 transformants (BCG-65, -73 and -83) yielded a 1.9-kb fragment only, while the 1.9-kb and 572-bp fragments were amplified from all the other transformants (data not shown).

Disruption of the mce gene by double cross-over was confirmed by Southern blot hybridization. Chromosomal DNA from wild-type M. bovis BCG and BCG-65 was isolated, digested with XhoI or SacI and analyzed by Southern blotting, using the PMCE probe (Fig. 1). For the XhoI digests (lanes 3–4), probing of the wild-type strain DNA highlighted a single band of 4.7 kb (lane 3) while DNA from BCG-65 showed a band of 6 kb (lane 4), as a result of the integration of the hyg gene within the mce gene. Probing of SacI digests with the PMCE probe (lanes 1 and 2) indicated the presence of a 5.2-kb band for the wild-type strain (lane 1), while the knock-out mutant displayed two bands of 4.8 and 1.7 kb (lane 2) as a result of the presence of a SacI site in the hyg gene integrated within the mce gene.

Protein lysates from wild-type and recombinant BCG strains were prepared, resolved on an acrylamide gel and immunoblotted with a monoclonal anti-Mce antibody. As shown in Fig. 2, BCG-65 did not express the Mce protein (lane 2), whereas the wild-type strain produced a protein approximately 48 kDa in size (lane 1). Therefore, the replacement of the native mce gene with a disrupted one resulted in a defective mutant unable to express the Mce protein.

3.3 Invasion assay

HeLa monolayers were infected with wild-type M. bovis BCG Connaught or the mce-deficient BCG-65 strain (BCGmce) in order to compare their invasive properties. The efficiency of invasion was expressed as the percentage of bacteria recovered from lysed HeLa monolayers compared with the original inoculum. Fig. 4 indicates that there is a 45% reduction in the relative ability of BCGmce to invade HeLa cells as compared to BCG wild-type (Student t-test analysis: P<0.001) and that the decrease in invasive potential was observed at both 2 and 6 h and, thus, did not depend on the duration of invasion.


Invasion assay. The ability of BCG wild-type and BCGmce to invade HeLa cells in vitro was compared. Results are expressed as the percentages of the initial inoculum that invaded HeLa cells and represent the mean result of four separate experiments, each performed in triplicate. Bars represent S.D.s. *, P<0.001 as compared to BCG wild-type.

Mycobacterial invasion of epithelial cells was first observed by Sheppard in HeLa cells [2] and its mechanism probably involves changes in the morphology of the host cell triggered by bacterial determinants. Despite the fact that M. tuberculosis resides primarily in macrophages in vivo [12], we compared the invasion of BCGmce and the wild-type strain into HeLa cells, as it was the cell line used by Arruda et al. [3] to characterize the mce gene. The 45% reduction in the relative ability of the BCGmce mutant to invade HeLa cells suggests an important role for the mce operon in the initiation of endocytosis. Indeed, assigning a direct role to the Mce protein in the invasive potential of mycobacteria must be tempered by the fact that the presence of the mce gene as the third gene in an eight-gene operon suggests that its inactivation is likely to affect the expression of the proteins encoded by the following genes. Therefore, the reduced invasiveness of BCGmce might be due to a lack of a cooperative mechanism involving multiple proteins. Expression studies of the proteins located downstream of the Mce protein will allow for testing of the hypothesis that the disruption of the mce gene interferes with the production of these proteins. Furthermore, to determine the exact role of the Mce protein in the mechanism of invasion, complementation of BCGmce by re-introduction of a copy of the wild-type mce gene on a plasmid must be considered.

4 Conclusion

We have sequenced the mycobacterial cell entry protein gene from M. bovis BCG and shown that it is virtually identical to that of M. tuberculosis. A mce knock-out mutant was engineered by homologous recombination and genetically characterized. The mutant strain exhibits a significantly reduced ability to invade the epithelial cell line HeLa. Whether the mechanisms of invasion of non-phagocytic cells and macrophages are similar remains to be determined conclusively, but the discovery of proteins or sugar moieties involved in such a process and the elucidation of the role of the Mce protein and the mce operon in the bacterial invasion of mammalian cells may help to understand the complex mechanisms used by M. tuberculosis during its entry into macrophages. Additional studies on the behavior of this mutant in vitro and in vivo may yield important information on the mechanisms of mycobacterial virulence.


We thank Bill Bradley for synthesis of the oligonucleotides, Diane England for DNA sequencing and Dr Ursula McGuinness for the gift of the monoclonal antibody against the Mce protein.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
View Abstract