Gene disruption experiments play an important role in the functional characterization of genes in mycobacteria and rely mostly on the use of one or two antibiotic resistance markers. We have developed a system for mycobacteria which features both the advantages of the use of antibiotic resistance markers for gene disruption experiments and the ability to efficiently rescue the marker leaving an unmarked mutation on the chromosome. This new genetic tool relies on the transposon γδ site-specific recombination system. A res-ΩKm-res cassette was used to generate an insertional mutation by allelic exchange both in Mycobacterium smegmatis and Mycobacterium bovis BCG. Upon expression in the mutated strains of tnpR, the transposon γδ resolvase gene, res-ΩKm-res, was excised efficiently leaving behind a single res sequence at the mutated locus. A plasmid was engineered allowing expression of tnpR from an easily curable mycobacterial vector. This system will be useful for simple construction of unmarked mutations or repeated use of the same antibiotic marker to generate multiple mutants.
Mycobacterial infections cause extremely serious infectious diseases and are responsible for millions of deaths annually worldwide. Over the last two decades, investigation into the biology of mycobacteria has benefited from the development of genetic tools (for review, see ) and completion of the genome sequences of several mycobacterial strains [2,3]. The next step is the functional characterization of gene products. In this task, systems allowing the production of defined mutations by allelic exchange contribute greatly to the characterization of the function of a defined gene [4,5]. Most of these systems rely on the transfer of an antibiotic resistance cassette onto the genome. The presence of a marker on the chromosome is often an advantage allowing (i) the easy selection of the mutation, and (ii) the tracking of a specific recombinant strain in experiments mixing several strains (such as competition experiments). For these reasons, they have been used extensively in functional characterization studies. However, the use of antibiotic resistance markers has several drawbacks. The insertion of an antibiotic resistance gene may have a polar effect on the expression of downstream genes complicating the characterization of each gene's role in an operon. Furthermore, integration of an antibiotic resistance cassette onto the chromosome excludes this marker for further genetic manipulation; and there are only few antibiotic resistance genes, especially for mycobacteria. This problem will become acute in mycobacteria where the sequenced genomes reveal many gene duplications with high sequence conservation leading to extensive redundancy . These redundancies may hamper the study of a gene function since a mutation in one gene may be compensated by the synthesis of a homolog [6,7]. Multiple mutations and complementation experiments are then required to decipher the biological role of a protein or a family of proteins. In these cases, the very small number of antibiotic resistance markers available for mycobacteria is a limitation. In fact, only kanamycin and hygromycin are currently used in genetic experiments with strains of the Mycobacterium tuberculosis complex.
To circumvent the use of antibiotic resistance markers, several groups have developed methods to produce unmarked mutations [8,9]. However, these methods lose the advantages of having a marker and also make the construction of the allelic exchange mutants more hazardous. Another possibility is to develop an antibiotic resistance cassette that may be rescued by genetic recombination and may subsequently be used for another round of mutation or other types of genetic manipulation. Such systems have been developed in various microbial species but not in mycobacteria . These systems rely on the use of site-specific recombination systems derived from bacteriophage or transposons. The antibiotic marker is flanked by a short nucleotide sequence in direct orientation. These sequences are recognized by recombinase, an enzyme that catalyzes the recombination on these two short sequences thereby excising the marker. Among the resolvase systems that have been well characterized are the TnpR and res site from transposon γδ . The TnpR resolvase mediates the very efficient resolution of the cointegrate intermediate generated during the transposition of γδ from a donor to a recipient DNA molecule . A system based on an antibiotic resistance cassette flanked by two copies of the res site in direct orientation and transcriptional fusion with the tnpR gene has been developed . When the promoter upstream of tnpR is expressed, the resolvase catalyzes the excision of the antibiotic resistance cassette to give a bacterial strain sensitive to the corresponding antibiotic. This system was used to identify bacterial genes transcriptionally induced during infection of an animal host .
Here, we describe the construction of a cassette containing the kanamycin resistance gene flanked by two transcriptional terminators and surrounded by two res sites from transposon γδ. This cassette was used to produce marked mutations in two genes of Mycobacterium bovis BCG and Mycobacterium smegmatis. We demonstrated the excision of the marker at a high frequency upon expression of the transposon γδ resolvase from a mycobacterial promoter in both fast and slow growing mycobacteria. This results in the production of unmarked mutations. We made a construct where the resolvase gene is expressed from a multicopy plasmid carrying a thermosensitive origin of replication and the counterselectable marker sacB. This new plasmid can be easily cured to produce a strain containing an unmarked mutation and no plasmid.
Materials and methods
Bacterial strains and culture conditions
M. smegmatis mc2155 and Escherichia coli strains were grown on Luria–Bertani (LB) broth or agar (Difco) at 37°C. M. bovis BCG 1173P2 (the Pasteur strain) was grown at 37°C on Middlebrook 7H9 medium (Difco) supplemented with ADC (0.2% dextrose, 0.5% bovine serum albumin fraction V, 0.0003% beef catalase) and 0.05% Tween 80 where indicated, or on solid Middlebrook 7H11 medium (Difco) supplemented with OADC (0.005% oleic acid, 0.2% dextrose, 0.5% bovine serum albumin fraction V, 0.085% NaCl, 0.0003% beef catalase). Kanamycin (Km), hygromycin (Hyg) and sucrose were added when required to final concentrations of 40µg ml−1, 50µg ml−1 and 2% (w/v) respectively.
General DNA techniques
Mycobacterial DNA was extracted from 5 ml saturated cultures as described previously . DNA pellets were resuspended in 100 µl of 10 mM Tris (pH 8) buffer. For polymerase chain reaction (PCR) amplification, 0.1 µl of the DNA preparation was used. Primers used for PCR amplification of the pyrF mutation were pyr1 (5′-TCGGATTCGGTTCGTTACGA-3′), pyr2 (5′-CCCAGCATGTGCAGGTCG-3′), res1 (5′-GCTCTAGAGCAACCGTCCGAAATATTATAAA-3′) or res2 (5′-GCTCTAGATCTCATAAAAATGTATCCTAAATCAAATATC-3′). PCR reactions were performed in a final volume of 50 µl containing 2.5 U of Taq DNA polymerase (Roche Molecular Biochemicals, Meylan, France), 10% Me2SO, and 1 mM of each primer. The amplification program consisted of one cycle of 10 min at 95°C, followed by 30 cycles of 30 s at 95°C, 30 s at 57°C, 30 s at 72°C, and a final 10 min at 72°C. Primers used for PCR amplification of the pks15/1 mutation were pks1K (5′-CCAACATCTCCTGCATGCGT-3′) and pks1L (5′-CTGGAAGTGGTGTGGTCGC-3′). PCR reactions were performed in a final volume of 50 µl containing 2.5 U of Taq DNA polymerase, 10% Me2SO, and 1 mM of each primer. The amplification program consisted of one cycle of 10 min at 95°C, followed by 30 cycles of 30 s at 95°C, 30 s at 57°C, 30 s at 72°C, and a final 10 min at 72°C. The PCR products were analyzed by electrophoresis in 0.8% agarose gels. For DNA sequencing, the 500-bp fragment obtained with primers pyr1 and pyr2 and the 497-bp fragment obtained with primers pks1L and pks1K were purified using the Qiaquick purification kit (Qiagen, Courtaboeuf, France) and sequenced using the same primers (ESGS Qbiogene, Evry, France).
The transposon γδres site was amplified from the genomic DNA of E. coli TG1 using primers res1 and res2. The res1 and res2 primers contained 5′ extensions corresponding to a Xba I restriction site for res1 and Bgl II plus Xba I restriction sites for res2. PCR reactions were performed in a final volume of 50 µl containing 2.5 U of Taq DNA polymerase, 10% Me2SO, and 1 mM of each primer. The amplification program consisted of one cycle of 5 min at 95°C, followed by 30 cycles of 1 min at 95°C, 1 min at 55°C, 1 min at 72°C, and a final 10 min at 72°C. The PCR product was analyzed by electrophoresis in 2% agarose gel. The purified 146-bp fragment was digested with either Xba I or Xba I and Bgl II and inserted either at the Xba I site or between the Xba I and Bam HI sites of vector pBluescript-SK+ (Stratagene, La Jolla, CA, USA) to give plasmids pCG118 and pCG121. Integrity of the res sequence in these two plasmids and its orientation was determined by sequencing. A 2.2-kb Bam HI fragment containing the Ωkm cassette from pHP45Ωkm  was inserted into the unique Bgl II site of pCG118 leading to pCG119 and pCG120 according to the orientation of the resistance cassette. Plasmid pCG119 was cut using enzyme Kpn I and then blunt-ended using the T4 DNA polymerase. The linearized pCG119 plasmid was then cut with Not I and the 2.4-kb fragment containing the Ωkm cassette plus one res site was gel-purified using the Qiaquick gel extraction purification kit. At the same time, plasmid pCG121 was cut with the enzyme Sac I and blunt-ended with the T4 DNA polymerase. A second cut was generated on this linearized vector with the Not I enzyme. The res-Ωkm-containing fragment was then inserted into the pCG121 cut vector to give plasmid pCG122. Thus, the plasmid was derived from pBluescript-SK+ and carried a res-Ωkm-res cassette where the two res sites are in direct orientation. This cassette may be released as a 2.6-kb fragment using various restriction enzymes (Fig. 1).
The transposon γδ resolvase gene, tnpR, was amplified from plasmid pIVET6  using primers tnpR1 (GGATGAGATCTCGACTTTTTGGTTACGCACGG-3′) and tnpR2 (5′-CCAATGCATCACTAATTCCACACTCAACC-3′). The PCR reaction was performed in a final volume of 50 µl containing 2.5 U of Taq Gold DNA polymerase (Roche Molecular Biochemicals), 10% Me2SO, and 1 mM of each primer. The amplification program consisted of one cycle of 5 min at 95°C, followed by 30 cycles of 1 min at 95°C, 1 min at 55°C, 1 min at 72°C, and a final 10 min at 72°C. The PCR product was analyzed by electrophoresis in 0.8% agarose gel. The 0.5-kb fragment was gel-purified using the Qiaquick gel extraction purification kit and inserted into the linearized pGEM-T easy vector according to the manufacturer recommendation (Promega, Lyon, France) to produce plasmid pCG123. The gene tnpR was purified on a 0.5-kb Bgl II–Kpn I fragment and inserted between the Bam HI and Kpn I sites of plasmid pMIP12  to give plasmid pCG124. A Pac I–Spe I 1.9-kb fragment was obtained from pCG124 and inserted between the Pac I and Spe I sites of pMIP12H . The resulting construct, pWM17, is an E. coli/mycobacteria shuttle plasmid containing a hygromycin resistance marker and the tnpR gene under the control of pBlaF*, a strong mycobacterial promoter (Fig. 1B).
A 1.6-kb fragment containing the hyg gene was obtained following digestion of plasmid pMIP12H with Kpn I and removal of the 3′ overhang by the T4 DNA polymerase. This fragment was inserted at the Pst I site (made blunt by the T4 DNA polymerase) of plasmid pPR23 to give pWM18. Finally, plasmid pWM19 was generated by inserting the 1.9-kb Pac I–Spe I fragment of pCG124 between the Bam HI and Spe I sites of pWM18 (the Pac I and Bam HI extremities were made blunt by the T4 DNA polymerase before digestion with Spe I) (Fig. 1B).
Results and discussion
Gene disruption in mycobacteria using a new antibiotic resistance cassette
The aim of this work was to develop a system that features both the advantages of the use of antibiotic resistance cassettes to produce allelic exchange mutants and the ability to rescue the marker afterward leaving only an unmarked mutation on the chromosome. To this end, a cassette was generated by inserting the ΩKm interposon between two copies of the res site from transposon γδ in direct orientation (Fig. 1A). The ΩKm interposon was previously shown to confer resistance to kanamycin in mycobacteria (data not shown). The res-ΩKm-res cassette was then used to produce insertional mutations in the fast grower M. smegmatis mc2155 and the slow grower M. bovis BCG.
We chose pyrF as the target gene for M. smegmatis. Mutations in this gene are feasible and the mutant strains have selectable phenotypes: they are auxotrophs for uracil and resistant to 5-fluoroorotic acid . To construct the M. smegmatis pyrF mutant, we used plasmid pY6001 described by Husson et al. . This plasmid contains a 5-kb fragment of the M. smegmatis chromosome carrying the entire pyrF gene. The res-ΩKm-res cassette was inserted into the unique Bam HI site of pY6001 to give pWM20.
M. smegmatis mc2155 was electrotransformed with this plasmid and the transformants were selected on LB plates containing 40 µg ml−1 kanamycin. Fifty-seven colonies were obtained after 4 days of incubation at 37°C and tested for growth on 7H11 plates both with and without the uracil supplement. Two clones appeared to be auxotrophs for uracil. One of them, named PMM9, was further analyzed by PCR to establish that it resulted from the replacement of the wild-type pyrF allele by the disrupted allele. Amplification with primers pyr1 and pyr2 gave no signal with PMM9 while it gave the expected 267-bp DNA product with the parental strain mc2155. In contrast, the same experiment with primers pyr1+res1 and pyr2+res2 produced a 340-bp and a 261-bp PCR fragment respectively with PMM9 but produced no amplification with mc2155 (Fig. 2A). These results established that PMM9 was an allelic exchange mutant.
For M. bovis BCG, we used the PMM3 mutant constructed for another study . Briefly, the plasmid, pWM08, was generated from vector pJQ200 by inserting a 4.6-kb fragment corresponding to the res-ΩKm-res cassette flanked by 1-kb arms corresponding to internal fragments of the pks15/1 gene of M. bovis BCG. This plasmid was transferred by electrotransformation in M. bovis BCG. From 76 KmR transformants, a mutant, PMM1, was obtained containing the pWM08 plasmid inserted via a single homologous recombination event at the pks15/1 locus. The second crossing-over event was then selected by plating a culture of PMM1 on medium containing 2% sucrose and kanamycin. Among the four KmR SucR colonies tested, two corresponded to allelic exchange mutants and then contained an insertion of the res-ΩKm-res cassette within the pks15/1 gene. One of them was named PMM3 (Fig. 2B).
These results indicate that the res-ΩKm-res cassette is suitable for the construction of allelic exchange mutant in both fast- and slow-growing mycobacteria.
Rescue of the res-ΩKm-res cassette and production of unmarked mutation
The res sites are sequences recognized by the transposon γδ resolvase. We produced a plasmid that expresses the transposon γδ resolvase, TnpR, in mycobacteria to address the question whether the res-ΩKm-res cassette may be rescued in mycobacteria. The tnpR gene was amplified by PCR from plasmid pIVET6 and was placed under the control of the strong mycobacterial promoter pBlaF* in the shuttle plasmid, pMIP12H (Fig. 1B). The resulting plasmid, pWM17, was used for electrotransformation of clones PMM3 and PMM9. Transformants PMM3:pWM17, PMM9:pWM17 were grown until saturation at 37°C in liquid medium with hygromycin before being plated in serial dilutions on solid medium with or without kanamycin. Colonies were counted after 4 days or 3 weeks of incubation at 37°C for M. smegmatis or M. bovis BCG strains respectively (Table 1). Whereas approximately 108 colonies were obtained on hygromycin plates for the two strains, approximately 2×106 and less than 104 colonies were counted on kanamycin plates for PMM9:pWM17 and PMM3:pWM17 respectively. To confirm that the HygR colonies were really kanamycin-sensitive, 92 clones of PMM3:pWM17 and 60 clones of PMM9:pWM17 were picked randomly from the hygromycin-containing plates and tested for kanamycin sensitivity. As expected, all of the tested clones were KmS. These results show that more than 96% and 99.9% of the PMM3:pWM17 and PMM9:pWM17 colonies respectively were sensitive to kanamycin which suggests that the res-ΩKm-res cassette was lost in these clones.
|CFU HygR ml−1||CFU KmR ml−1||Percentage CFU KmR/CFU HygR|
|M. bovis BCG|
|CFU HygR ml−1||CFU KmR ml−1||Percentage CFU KmR/CFU HygR|
|M. bovis BCG|
To see if resolution occurred between the two res sites in PMM3:pWM17 and PMM9:pWM17, we further analyzed one KmS clone of each strain by PCR using various combinations of primers flanking the insertion site of the res-ΩKm-res cassette within the pyrF and pks15/1 genes (Fig. 2). The PCR pattern results revealed that the res-ΩKm-res was excised leaving a copy of the 132-bp res site within the targeted gene. This was confirmed by sequencing the PCR products obtained with pyr1+pyr2 or pks1K+pks1L primers with PMM9:pWM17 and PMM3:pWM17 respectively (data not shown). These results demonstrate that resolution occurred between the two res sites in M. smegmatis- and M. bovis BCG-derived strains showing that the transposon γδ resolvase system is functional in mycobacteria.
We observed a higher frequency of resolution in M. bovis BCG than in M. smegmatis. The strength of the pBlaF* promoter was previously shown to be comparable in both strains , however we cannot exclude that TnpR is produced or accumulated in larger quantity in M. bovis BCG than in M. smegmatis. Another possibility is that the lower frequencies of KmS clones obtained in M. smegmatis than M. bovis BCG are due to the different incubation time required for growing these strains to saturation in liquid culture (3 days for M. smegmatis and 2 weeks for M. bovis BCG): the time during which recombination between the two res sites may occur in PMM3:pWM17 was therefore much longer than in PMM9:pWM17. However, in both cases, the large majority of clones recovered after expression of the resolvase are KmS mutants showing that the system is suitable to easily produce unmarked mutations on the mycobacterial chromosome.
Construction of an easily curable mycobacterial plasmid which allows the expression of γδ resolvase
To effectively generate site-specific mutated strains carrying no antibiotic resistance marker, the plasmid allowing the γδ resolvase expression has to be easily curable. For this purpose, a new plasmid, pWM19, was generated by cloning a fragment containing the γδ resolvase under the control of the pBlaF* promoter into pWM18, a plasmid derived from the thermosensitive counterselectable vector pPR23  and carrying gentamicin and hygromycin resistance cassettes (Fig. 1B). As shown in previous studies, pPR23-derived plasmids may be easily cured in mycobacteria by incubating a culture at a temperature higher than 39°C in the presence of 2% sucrose . We tested whether this construct would promote the expression of γδ resolvase in mycobacteria and thereby rescue the res-ΩKm-res cassette. Strain PMM9 was electrotransformed with either pWM19 or the parental plasmid pWM18 that does not carry the tnpR gene and transformants were selected at 32°C on hygromycin. Several transformants, PMM9:pWM18 and PMM9:pWM19, were grown until saturation at 32°C in liquid medium with hygromycin then plated in serial dilutions on solid medium with or without kanamycin. Colonies were counted after 5 days of incubation at 32°C (Table 1). As expected, excision occurred in the strain expressing the γδ resolvase. This rescue of the res-ΩKm-res cassette was dependent on the presence of tnpR on the plasmid since similar numbers of KmR and HygR were obtained with plasmid pWM18. In seven independent experiments realized with pWM19, we obtained a great variability in the percentage of res-ΩKm-res cassette rescue: from less than 5.0×10−5% to 6.2% of KmR colony-forming units (CFU) (Table 1). In fact, two different groups of results were obtained: either a percentage of KmR CFU below our detection level or a percentage in the same range as the one observed for pWM17, i.e. from 0.1 to 6.2%. An explanation for these results may be that several clones picked on the transformation plates were already KmS because recombination between the two res sites flanking the res-ΩKm-res cassette happened in these CFU very early after the transfer of pWM19 into PMM9. To address this issue, 16 independent CFUs from each of three transformations, PMM9:pWM17, PMM9:pWM18 and PMM9:pWM19, were picked directly from the transformation plates and patched on kanamycin- or hygromycin-containing plates. As expected, all the clones were HygR and KmR for PMM9:pWM18. For the two strains expressing the TnpR resolvase, we obtained two out of 16 clones already KmS for PMM9:pWM17 and four out of 16 clones KmS for PMM9:pWM19. Therefore, occurrence of these CFUs where the res-ΩKm-res was lost very early may explain our results (Table 1). Differences in number of Kms clones obtained with pWM17 and pWM19 directly on the transformation plates were likely due to the time between electroporation and plating where bacteria were incubated in growth medium without antibiotic to allow expression of the antibiotic resistance. This time was longer for pWM19 (16 h at 32°C) than for pWM17 (3 h at 37°C) because pWM19 has a thermosensitive replicon. However, with both plasmids, the resolution rate is high enough to obtain recombination and loss of the antibiotic resistance marker in most of the clones at the end of our protocol (Table 1).
To show that plasmids carrying tnpR can be easily cured in PMM9:pWM19, one Kms clone picked on a hygromycin plate was restreaked on solid medium containing 2% sucrose and incubated at 42°C for 4 days. Ten SucR colonies obtained after restreaking were tested for resistance to gentamicin or hygromycin and all were GmS and HygS which indicated the loss of pWM19. Hence, the experiments produced colonies which were unmarked site-specific mutants.
These results demonstrate that pWM19 is suitable to promote the rescue of the res-ΩKm-res cassette and can be very efficiently cured.
We demonstrated that the transposon γδ resolvase is functional in mycobacteria and catalyzes site-specific recombination between two res sites in direct orientation on the chromosome of M. smegmatis and M. bovis BCG. Based on this system, we have developed new genetic tools to produce unmarked site-specific mutations in mycobacteria. These tools combine the advantages of an antibiotic resistance cassette for selection of rare genetic events and the ability to remove the marker when required. This method should allow the successive use of the same antibiotic resistance cassette to produce multiple mutations on the chromosome of mycobacteria.
E.P. is the recipient of a Marie Curie Fellowship from the European Union. This work is supported by the Centre National de la Recherche Scientifique (CNRS, France) and the Ministère de la Recherche (Programme de Recherche Fondamentale en Microbiologie et Maladies Infectieuses et Parasitaires).