Identification of the Origin of Replication of the Mycoplasma pulmonis Chromosome and Its Use in oriC Replicative Plasmids

Abstract

Mycoplasma pulmonis is a natural rodent pathogen, considered a privileged model for studying respiratory mycoplasmosis. The complete genome of this bacterium, which belongs to the class Mollicutes, has recently been sequenced, but studying the role of specific genes requires improved genetic tools. In silico comparative analysis of sequenced mollicute genomes indicated the lack of conservation of gene order in the region containing the predicted origin of replication (oriC) and the existence, in most of the mollicute genomes examined, of putative DnaA boxes lying upstream and downstream from the dnaA gene. The predicted M. pulmonis oriC region was shown to be functional after cloning it into an artificial plasmid and after transformation of the mycoplasma, which was obtained with a frequency of 3 × 10⁻⁶ transformants/CFU/μg of plasmid DNA. However, after a few in vitro passages, this plasmid integrated into the chromosomal oriC region. Reduction of this oriC region by subcloning experiments to the region either upstream or downstream from dnaA resulted in plasmids that failed to replicate in M. pulmonis, except when these two intergenic regions were cloned with the tetM determinant as a spacer in between them. An internal fragment of the M. pulmonis hemolysin A gene (hlyA) was cloned into this oriC plasmid, and the resulting construct was used to transform M. pulmonis. Targeted integration of this genetic element into the chromosomal hlyA by a single crossing over, which results in the disruption of the gene, could be documented. These mycoplasmal oriC plasmids may therefore become valuable tools for investigating the roles of specific genes, including those potentially implicated in pathogenesis.

The mollicutes are the smallest free-living microorganisms capable of autoreplication (33). They are related to gram-positive eubacteria from which they evolved by a drastic reduction of genome size, being thus considered the best representatives of the concept of a minimal cell. The genomes of four mollicutes have been sequenced to date: Mycoplasma genitalium (580 kbp, [15]), Mycoplasma pneumoniae (816 kpb [21]), Ureaplasma urealyticum (752 kpb [19]), and Mycoplasma pulmonis (964 kbp [5]). Postgenomic approaches aiming at deciphering the role of specific genes discovered by in silico analysis now require efficient strategies to genetically manipulate these organisms. The transposon Tn916 or Tn4001 can transpose in all the mollicute genomes, at least for the species amenable to transformation, such as M. genitalium, M. pneumoniae, and M. pulmonis. In M. genitalium and M. pneumoniae, a global approach of gene inactivation with the transposon Tn4001 was used to define the genes that are essential for the life of these bacteria (22). Derivatives of this transposon, obtained by insertion of a chloramphenicol acetyltransferase gene and the tetM tetracycline resistance determinant, were also shown to transpose in M. pulmonis (12). The use of these genetic elements for gene inactivation, which relies on the random insertion of the transposon in the genome of the organism, does not allow specific targeting of a gene of interest. In addition, the complementation of mutants would require the cloning and expression of genes in these organisms, experiments for which there are no genetic tools to date. If we except the cryptic replicating plasmids previously described for Mycoplasma capricolum and Mycoplasma mycoides subsp. mycoides (23, 24), there no other plasmids known to replicate in Mycoplasma species. In contrast, replicative plasmids have been obtained for the mollicute Spiroplasma citri by cloning the origin of replication of the chromosome (oriC) into artificial plasmids (34, 41). The resulting oriC plasmids can be used to inactivate targeted genes by homologous recombination (8, 41), to support heterologous gene expression in S. citri (35), and to functionally complement mutants obtained by transposon insertion (18).

The first aim of this work was to compare the putative oriC region of mollicutes for which sequence data are available. This analysis served as the basis for the development of oriC plasmids that replicate in M. pulmonis. The minimal sequences required for the replication of these plasmids were found. To evaluate the use of these vectors for inactivation and/or complementation of genes of interest, we have chosen to disrupt the M. pulmonis hemolysin A gene.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

The M. pulmonis strain used in this work was UAB CTIP (5). For subcloning experiments and propagation of plasmids, the Escherichia coli strain TG1 {supE hsdΔ5 thiΔ (lac-proAB)F′[traD36 proAB⁺ lacI^q lacZΔM15]) was used. Mycoplasmas were grown in modified Hayflick medium N (16) without thallium acetate and supplemented with BBL IsoVitalex Enrichment (Becton Dickinson, Sparks, Md.). For growth in solid medium, mycoplasmas were incubated at 37°C under anaerobic conditions. E. coli cells were grown in Luria-Bertani (LB) broth or in LB agar at 37°C. E. coli cells transformed with plasmids were grown in LB medium supplemented with 50 μg of ampicillin/ml and 5 μg of tetracycline/ml.

Cloning procedures.

The pBOT1 plasmid which contains a 2-kbp oriC region from S. citri was described elsewhere (35). All plasmids constructed in this study were based on the pSRT2 plasmid (C. Lartigue et al., unpublished data) (Fig. 1), which harbors the tetM gene (25) from the transposon Tn916 inserted into the pBS plasmid (Stratagene, La Jolla, Calif.). In pSRT2, the tetM gene is under the control of the S. citri spiralin gene promoter (40). The 1.9-kbp oriC region of M. pulmonis, containing the dnaA gene with its upstream and downstream intergenic regions (Fig. 1), was PCR amplified with the primers OR1 and OR2 (Table 1). This amplified DNA fragment was cloned into the unique BamHI site of pSRT2, resulting in the pMPO1 plasmid (Fig. 1). The downstream DnaA box region was PCR amplified with the primers OR3 and OR2, and the upstream DnaA box region was amplified with the primers OR1 and OR4 (Table 1). To clone these two DNA fragments side by side at the BamHI site of the vector pSRT2 for obtaining the pMPO4 plasmid, the upstream and the downstream DnaA box regions were amplified with the primers OR1 and OR5, and OR2 and OR6, respectively. To clone these two fragments spaced by the tetM determinant, the upstream DnaA box region was first amplified with the primers OR7 and OR8 and cloned at the SphI site of pSRT2. The downstream DnaA box region was then amplified with the primers OR2 and OR3 and cloned at the BamHI site, resulting in the plasmid pMPO5 (Fig. 1).

FIG. 1.

M. pulmonis oriC region and plasmid constructs developed in this work (pMPO1, pMPO2, pMPO3, pMPO4, and pMPO5), using the vector pSRT2 for cloning. The plasmid pMPO1 was obtained by cloning the 1.9-kpb oriC region into the unique BamHI site of the pSRT2 plasmid. The orientations of the three genes (recD, dnaA, and dnaN) found in the vicinity of the predicted oriC region are indicated by arrows, and those of the putative DnaA boxes are indicated by triangles above the intergenic regions. The plasmids pMPO2 and pMPO3 were constructed by cloning the intergenic regions downstream and upstream from dnaA, respectively, into the unique BamHI site of the pSRT2 plasmid. The plasmid pMPO4 was constructed by cloning these two fragments side by side into the BamHI site of the vector pSRT2, and the plasmid pMPO5 was constructed by cloning these two fragments with the tetM determinant in between them into the BamHI site of the vector pSRT2. P, promoter of the S. citri spiralin gene.

TABLE 1.

Sets of primers used in this work for PCR amplification

Primer name	Sequencea	Annealing tempb (°C)	Amplified region	Size of amplified fragment
OR1	5′-ATTAGGGATCCGCACTCTGGTCAGCGCTAGATC-3′	56	Complete oriC region	1.9 kbp
OR2	5′-TCGCGGATCCGCCTAACTTGAAAATAAGCTCC-3′
OR2		58	Downstream DnaA boxes	326 bp
OR3	5′-GGCTGGATCCGAAAACTTATCCAAGG-3′
OR1		58	Upstream DnaA boxes	262 bp
OR4	5′-CGCTAGGATCCCTATTTTGTCAAGGC-3′
OR1		56	Upstream DnaA boxes	262 bp
OR5	5′-CGCTAGAATTCCTATTTTGTCAAGGC-3′
OR2		56	Downstream DnaA boxes	326 bp
OR6	5′-GGCTGAATTCGAAAACTTATCCAAGG-3′
OR7	5′-AATAGGCATGCGCACTCTGGTCAGGCGTAGATC-3′	56	Upstream DnaA boxes	262 bp
OR8	5′-CGCTAGCATGCCTATTTTGTCAAGGC-3′
HE1	5′-CCGCGAATTCGAAAAAGAAGCTGTTGG-3′	56	hlyA gene	688 bp
HE2	5′-GGCCCGAATTCGAGAGATGTATTCAATG-3′
HEL	5′-CCCCGAATTCGATAATTGCCTTGCAG-3′	56	hlyA gene and flanking regions	1.05 kbp
HER	5′-CAGGTGAATTCGAATTAGCGGCTGAGTC-3′

^aSequences recognized by the restriction enzymes mentioned in the text are underlined.

^bAnnealing temperature used during the PCR.

Transformation of mycoplasmas by polyethylene glycol (PEG).

M. pulmonis cells from 10-ml cultures were transformed by a PEG-mediated method previously described by others (12). Ten micrograms of plasmid DNA was used in each transformation. After plating, the cultures were incubated at 37°C under anaerobic conditions and were examined from the third day of incubation for colony development. Transformants were subcultured in 2 ml of Hayflick liquid media, in which the tetracycline concentration was gradually increased from 2 to 50 μg/ml. Cloning of transformed mycoplasmas was performed by standard methodology with three steps of filter cloning procedures using membranes with a pore size of 0.45 μm (39).

DNA isolation and Southern blot hybridization.

Mycoplasma genomic DNA was obtained from a 10-ml mycoplasma culture by modifications of the method described by Ferris et al. (14). Briefly, cells were collected by centrifugation (20,000 × g, 20 min, 4°C) and resuspended in 1.0 ml of 10 mM Tris-HCl (pH 8.0) containing 50 mM EDTA. To this suspension, 64 μl of a 10% sodium dodecyl sulfate solution and 160 μl of proteinase K at a concentration of 10 mg/ml were added. The lysate was incubated at 56°C for 60 min and then deproteinized twice with phenol and once with phenol-chloroform-isoamyl alcohol. The nucleic acids were ethanol precipitated. The DNA was further purified by RNase A treatment followed by phenol chloroform deproteinization and ethanol precipitation and finally resuspended in 50 μl of Tris-EDTA buffer.

For Southern blot hybridization, approximately 2 μg of genomic DNA or 20 ng of plasmid DNA were digested by the appropriate restriction enzyme, and submitted to electrophoresis in a 0.8% agarose gel. After alkali transfer of the DNA to a positively charged nylon membrane, hybridization was performed in the presence of 20 ng/ml of digoxigenin-labeled probes. Hybridization was detected by incubation of the membranes with anti-digoxigenin antibodies coupled to alkaline-phosphatase and the use of the chemoluminescent substrate CPD Star (disodium-4-chloro-3-[4-methoxyspiro{1,2-dioxietane-3,2′-(5′-chloro)tricyclo(3.3.1.1^3,7) decan}-4-yl]phenyl phosphate) (Roche Molecular Biochemicals).

PCR-based detection of disruption events of the hlyA gene.

Mycoplasma cells were collected from 1.0 ml cultures by centrifugation (10,000 × g, 20 min, 4°C) and washed once with PBS. After resuspension in 100 μl of lysis buffer (10 mM Tris pH 8.0, 1 mM EDTA, 0.25% Nonidet P-40, 0.25% Tween-20, 100 μg/ml proteinase K), the lysate was first incubated at 56°C for 60 min, then at 95°C for 10 min. Nucleic acids were deproteinized once with phenol and twice with phenol-chloroform. After ethanol-precipitation, the DNA was resuspended in 50 μl of Tris-EDTA buffer. To detect M. pulmonis hlyA mutants resulting from disruption of the chromosomal hlyA by insertion of plasmid sequences, a PCR-based strategy was developed. The wild-type hlyA gene can be PCR amplified by using the primers HE-L and HE-R located upstream and downstream of the gene, respectively (Table 1; see Fig. 4). When the gene is disrupted by insertion of the plasmid sequence, the sequence is too large to be PCR amplified under the same conditions and primers. However, the recombination event can be detected by PCR amplification using the other pair of primers (HE-R and OR2), as indicated in Fig. 4.

FIG. 4.

Scheme of integration of the plasmid pMPO5-ΔhlyA into the M. pulmonis chromosomal hlyA gene by a single crossing-over. The genetic organization of pMPO5-ΔhlyA and the hlyA chromosomal region are schematically shown before and after integration. The size of the HindIII (H) fragments which overlap hlyA sequences is indicated. Arrowheads indicate the positions of the primers (HE-L, HE-R, and OR2) used to document this integration. PS, promoter of the S. citri spiralin gene; ori5′, region containing the DnaA boxes upstream from the dnaA gene; ori3′, region containing the DnaA boxes downstream from the dnaA gene.

Sequence analysis of the oriC regions of mollicutes.

DnaA boxes were searched within the oriC regions of mollicutes using the MEME/MAST software (3). A random sequence containing seven copies of the E. coli DnaA box consensus sequence (TTATCCACA) was used as a training sequence. Query sequences were examined with a window of 150 bp, sliding with an increment of 100 bp. The length of the motifs searched was restricted to 9 bp in a first query and extended to 6 to 11 bp in a second one.

RESULTS

Identification of putative DnaA boxes within oriC regions of mollicutes.

Putative DnaA boxes have been searched by others within the sequenced oriC regions of mollicutes. However, these results were neither compared nor obtained with similar methods (17, 20, 41). In order to detect these putative DnaA boxes, the MEME/MAST set of softwares for finding conserved motifs in biological sequences was used. The validity of the method was first tested with the well-documented oriC regions of S. citri and M. capricolum. By use of a random sequence containing seven copies of the consensus sequence for the E. coli DnaA boxes (TTATCCACA) as a training sequence, the seven DnaA boxes previously described within the functional oriC region of S. citri (42) were detected (Fig. 2). Using the same strategy, the 10 boxes described for the M. capricolum oriC region (17) were found. These findings indicated that our method was reliable for detecting putative DnaA boxes in the genomes of other mollicutes.

FIG. 2.

Gene order and putative DnaA boxes within the oriC regions of mollicute genomes. Orientations of the genes are indicated by arrows; intergenic regions flanking the dnaA genes are magnified with their lengths (in nucleotides) indicated. PCR, putative coding region; CHP, conserved hypothetical protein. Putative DnaA boxes are represented by headless arrows. A match with the consensus defined for mollicutes (TT(A/T)TC(C/A)ACA) is symbolized by the following: black, nine of nine; horizontal stripes, eight of nine; white, seven of nine.

In the four sequenced mollicute genomes (M. genitalium, M. pneumoniae, U. urealyticum, and M. pulmonis), the oriC regions have not been experimentally located, but a number of analyses suggest that the oriC regions are situated, as for most bacteria, in the vicinity of the dnaA gene. Indeed, there is a polarity not only in the orientation of the genes on both sides of dnaA but also in the base composition asymmetries at the level of this gene (5, 26, 31, 37). It should be noted that the asymmetry in the strand base composition which is typically found around dnaA in bacterial genomes could barely be detected in the M. pneumoniae genome (31).

Using our method based on motif detection, putative DnaA boxes were found both upstream (five boxes) and downstream (three boxes) from dnaA in the M. pulmonis oriC region (Fig. 2). The consensus sequence derived from these eight DnaA boxes is TTATC[C/A]A[C/A]A and resembled the typical TTATCCACA DnaA box motif of E. coli, with a variability at positions six and eight (Table 2). Four of the eight M. pulmonis DnaA boxes perfectly matched the consensus, whereas the other four matched at eight positions out of nine. Two AT-rich stretches of 31 and 39 bp were found within the intergenic regions upstream and downstream from dnaA, respectively.

TABLE 2.

Putative DnaA boxes within the oriC intergenic regions of M. pulmonis

DnaA box	Sequence	Strand	Matcha [cons. M. pulmonis (cons. Mollicutes)]	Positionb
recD/dnaA				1-222
Box 1	TTATCCAAA	+	9 (8)	59-67
Box 2	TTATTAACA	−	8 (8)	77-85
Box 3	TTATCAACA	−	9 (9)	101-109
Box 4	TTATCCACA	−	9 (9)	143-151
Box 5	TTATCAACT	+	8 (8)	173-181
dnaA/dnaN				1608-1740
Box 6	TTATCCAAG	+	8 (7)	1620-1628
Box 7	TTATCCAAA	+	9 (8)	1673-1681
Box 8	TTAACAACA	+	8 (8)	1684-1692

^acons., consensus. See the text.

^bPosition in reference to the first nucleotide of the recD/dnaA intergenic region.

Using the same approach, five putative DnaA boxes were found in the intergenic sequence soJ-dnaA of the M. genitalium chromosome (Fig. 2). Although these boxes were not identified in the initial paper describing the complete sequence of this bacterium (15), the origin of replication was predicted to be localized in this untranscribed region by others (26). Within the oriC region of U. urealyticum, only one putative DnaA box matching 7 of 9 positions with the E. coli consensus was found in the intergenic region (146 nt) between the genes rpL34 and dnaA (data not shown). No putative DnaA boxes could be found in the vicinity of the dnaA gene of the M. pneumoniae genome. However, it should be noted that the intergenic region cysA-dnaA is only 68 nt long and that a putative coding region has been predicted immediately downstream from dnaA. In addition, as found by others (20), a few putative DnaA boxes were located within the 747-bp-long intergenic soj-dnaN intergenic sequence, but only two of them presented fewer than three mismatches with the E. coli consensus.

From the 31 putative DnaA boxes predicted from the oriC regions of M. pulmonis, M. capricolum, M. genitalium, and S. citri, a mollicutes DnaA box consensus was proposed (TT(A/T)TC(C/A)ACA); the positions three and six presented a variability in comparison with the E. coli consensus.

Functional analysis of M. pulmonis oriC for plasmid replication.

Although no functional oriC region was yet isolated from the chromosome of species belonging to the genus Mycoplasma, the construction of oriC-based plasmids which replicated in the mollicute S. citri prompted us to investigate whether similar plasmids could replicate in M. pulmonis.

The first oriC plasmids that were assayed for transformation of M. pulmonis were those derived from the S. citri sequence. The pBOT1 plasmid containing a 2.0-kbp oriC fragment of S. citri (8) was tested. However, transformation of M. pulmonis by the PEG-mediated methodology could not be achieved with pBOT1. As a control for transformation of M. pulmonis by the PEG-mediated method, the plasmid pAM120 (10), which contains the transposon Tn916, was used. It yielded tetracycline-resistant colonies with a frequency of 10⁻⁹ transformants/CFU/μg of plasmid DNA. These results suggested that the oriC plasmids are host specific. Therefore, oriC plasmids based on the M. pulmonis oriC region were constructed.

The plasmid pMPO1 was obtained by combining the 1.9-kbp oriC fragment of M. pulmonis with the pBS vector and the tetM gene driven by the S. citri spiralin gene promoter (Fig. 1). Transformation of M. pulmonis with this plasmid succeeded with a frequency of 3 × 10⁻⁶ transformants/CFU/μg of plasmid DNA. With such an efficacy, about 2,000 transformants were obtained by a single transformation with 10 μg of pMPO1 and a starting culture of M. pulmonis titering 6 × 10⁷ CFU/ml. In agar plates, tetracycline-resistant (Tet^r) colonies were observed from the fifth day after plating. Their growth was slower than that of nontransformed cells, for which colonies were seen as early as 2 days after plating. From the third passage in liquid medium containing tetracycline, the growth rate of transformed mycoplasmas was similar to that of nontransformed cells, leading to acidification of the culture medium in less than 36 h. Transformation was confirmed by Southern blot hybridization using the tetM gene and the oriC region as probes (Fig. 3B; data not shown for the tetM probe). Both sequences could be detected in the genetic material extracted from the transformants. For at least some of the transformants, there was evidence of the plasmid autonomously replicating in the bacteria (Fig. 3B, lanes 3-2P and 3-5P).

FIG. 3.

Stability of oriC plasmids in M. pulmonis. (A) Schematic representation of the integration of the plasmid pMPO1 in the M. pulmonis chromosome within the oriC region. The size of HindIII (H) DNA fragments which overlap the oriC probe is indicated. (B) Southern blot analysis of DNA from M. pulmonis transformed with pMPO5 after 2 (2P), 5 (5P), or 20 (20P) passages in liquid medium. The purified DNAs of three M. pulmonis transformants numbered 1 to 3, and those of nontransformed M. pulmonis (NT) or of the plasmid (P) were digested with HindIII before electrophoresis. M. pulmonis oriC was used as a probe. (C) Schematic representation of the integration of the plasmid pMPO5 in the M. pulmonis chromosome within the oriC region. The intergenic regions located upstream and downstream of dnaA are shown by boxes with vertical and horizontal hatchings, respectively. The integration scheme is shown assuming a crossover within the region located upstream of dnaA. The size of EcoRI (E) DNA fragments which overlap the oriC probe is indicated. If the integration would occur within the other intergenic region, the predicted sizes of the EcoRI fragments overlapping the oriC probe are 12 and 0.65 kbp (data not shown). An integration with a double crossover would be lethal because it would result in the loss of the dnaA sequence from the chromosome. (D) Southern blot analysis of DNA from M. pulmonis transformed with pMPO5 after five passages in liquid medium. The purified DNAs of M. pulmonis transformants (D) (lanes 1 to 5), of nontransformed M. pulmonis (NT), or of the plasmid (P) were digested with EcoRI before electrophoresis. M. pulmonis oriC was used as a probe.

Since it was shown previously that the presence of a functional dnaA gene is not required for oriC plasmid replication in S. citri (35), subcloning experiments were undertaken to reduce the oriC fragment of pMPO1 to the minimal sequences required for plasmid replication (Fig. 1). In the two plasmids that were obtained, pMPO2 and pMPO3,, the oriC fragments correspond to the DnaA box region upstream (262 bp) and to the DnaA box region downstream (327 bp) from the dnaA gene, respectively (Fig. 1). None of these two plasmids replicated in M. pulmonis. Similarly, M. pulmonis could not be transformed with the plasmid pMPO4, which contains these two DnaA box regions cloned side by side at the BamHI site of the vector pSRT2 (Fig. 1). In contrast, pMPO5, in which the two DnaA box regions are spaced by the tetM gene, transformed M. pulmonis with a frequency of 4 × 10⁻⁶ transformants/CFU/μg of plasmid. Subculturing Tet^R colonies of mycoplasmas transformed with pMPO5 revealed that the growth rate of these transformants was markedly slower than for those transformed with the plasmid pMPO1, with up to 1 week being required for acidification of the culture medium. The presence of pMPO5 in the transformants was confirmed by Southern blot hybridization with an oriC probe (Fig. 3D).

Stability of oriC-based plasmids in M. pulmonis.

To evaluate the stability of oriC plasmids in M. pulmonis, genomic DNAs from transformants obtained with either pMPO1 or pMPO5 were analyzed by Southern blots after different passages in the presence of selection pressure (tetracycline). For the plasmid pMPO1 (Fig. 3A), the appearance of the 6.7- and 4.0-kpb HindIII DNA fragments hybridizing with the oriC probe indicates that integration of pMPO1 into chromosomal oriC did occur. In contrast, the presence of the 3.7- and 7.0-kbp hybridizing signals should indicate that some of the transformed cells still contained wild-type oriC and free plasmid. Three independent transformants were analyzed after 2, 5, and 20 passages in vitro (Fig. 3B). For transformant 2, there was evidence of at least partial integration of pMPO1 even after only two passages, and integration of the plasmid was complete after 20 passages. In contrast, integration of pMPO1 followed a different time course for transformant 3, with no detectable integration after five passages, indicating that the plasmid was autonomously replicating in the bacteria, and only partial integration after 20 passages (Fig. 3B). An intermediate pattern of integration was obtained with transformant 1.

The scheme of integration of the plasmid pMPO5 is depicted in Fig. 3C. In contrast to pMPO1, analysis of pMPO5 transformants revealed only two bands, corresponding to the free plasmid and intact chromosomal oriC, indicating that reduction of the oriC region led to a stable replicative plasmid. In all transformants tested, only free plasmid was detected, at least until the 15th passage (Fig. 3D).

Evidence for specific gene inactivation in M. pulmonis targeted by a recombinant oriC-based plasmid.

In an attempt to inactivate the hlyA gene through homologous recombination, we first used the classical strategy based on suicide vectors. An internal fragment (688 bp) of the M. pulmonis hlyA gene was PCR amplified with the primers HE1 and HE2 (Table 1) and inserted in the unique EcoRI site of the pSRT2 plasmid. Transformation of M. pulmonis with the recombinant plasmid pSRT2-ΔhlyA repeatedly yielded no Tet^r transformants. This suggests that the frequency of recombination events with use of nonreplicative plasmids is probably very low with M. pulmonis.

Therefore, we used the replicative, oriC plasmids pMPO1 and pMPO5 as disruption vectors according to the strategy presented for pMPO5 in Fig. 4 . The use of replicative vectors in these experiments was expected to enhance the opportunity for recombination between the plasmid and the chromosome and hence for disruption of the target gene. Similarly to the construction of pSRT2-ΔhlyA, the internal fragment of the M. pulmonis hlyA gene was inserted into the EcoRI site of the pMPO1 and pMPO5 plasmids, yielding the disruption plasmids pMPO1-ΔhlyA and pMPO5-ΔhlyA, respectively. After transformation of M. pulmonis with these constructs, Tet^R colonies were obtained with the same frequency as with the parental plasmids. Transformants were subcultured in liquid medium in which the tetracycline concentration was gradually increased from 2 to 50 μg/ml. Genomic DNA was extracted from these cultures after 5, 10, 15, and 20 passages in vitro and analyzed by Southern blot hybridization with an hlyA probe. Similarly to its parental plasmid, the plasmid pMPO5-ΔhlyA replicated as a free genetic element in the transformants, even after 20 passages (data not shown). In contrast, the plasmid pMPO1-ΔhlyA was not stably maintained as a free plasmid because it was integrated into chromosomal oriC after a few passages. For transformants obtained with both pMPO1-ΔhlyA and pMPO5-ΔhlyA, events of integration of the plasmids into the targeted hlyA gene could not be detected by analysis of the genetic material by Southern blot hybridization (data not shown). In order to increase the probability of isolating the few mycoplasma cells in which recombination might have occurred, further experiments were conducted with the pMPO5-ΔhlyA transformants.

First, the transformants were grown in the absence of tetracycline to determine whether loss of the free plasmid in the absence of selection could favor the selection of rare recombination events, such as plasmid integration at the hlyA gene. Three randomly selected M. pulmonis transformants with pMPO5-ΔhlyA were cultured for four passages in the presence of tetracycline. These transformants were then submitted to 1, 5, 10, or 15 passages without antibiotic before being plated in the absence and in the presence of tetracycline. After incubation, colonies were counted on both types of medium, allowing us to determine the percentage of the cells that kept the tet^r genetic determinant, regardless of its chromosomal or plasmidic localization (Fig. 5). When submitted to one passage without selective pressure, an average of 42% of colonies were tetracycline resistant. This percentage markedly decreased with passage number down to 4.1% after 5 passages, 0.008% after 10 passages, and 0.0005% after 15 passages without tetracycline. These data indicate that the plasmid pMPO5-ΔhlyA is not maintained by the mycoplasma cell in the absence of selective pressure, suggesting that these passages in a medium without antibiotic could be a means for isolating M. pulmonis hlyA mutants.

FIG. 5.

Stability of the plasmid pMPO5-ΔhlyA. The percentages of colonies that remained Tet^r after different passages without selective pressure (no tetracycline added to the medium) are represented.

Therefore, M. pulmonis cells transformed with pMPO5-ΔhlyA were cultured for five passages in selective medium, submitted to five additional passages without tetracycline, and then plated onto solid medium containing 2 μg of tetracycline/ml. Randomly selected clones (Tet^r colonies) were grown in liquid medium with antibiotic, and genomic DNA was analyzed by Southern blot hybridization (Fig. 6A). The hybridization patterns of all five transformants are almost identical. According to Fig. 4, the appearance of the 4.8- and 5.5-kpb HindIII DNA fragments hybridizing with the hlyA probe clearly indicates that integration of the disruption vector into the target gene did occur. However, the presence of the 6.2- and 4.0-kbp hybridizing signals indicates that some of the transformed cells still contained the wild-type hlyA gene and free plasmid. Similar patterns were obtained from all selected transformants, even after rounds of repeated filter cloning. Additional evidence of the hlyA gene disruption was obtained by using a PCR-based approach (Fig. 6B). Using the primers OR2 and HER, a 1.05-kbp DNA was amplified, showing that pMPO5-ΔhlyA had integrated into the M. pulmonis chromosome by one crossover recombination at the hlyA gene. As expected from the Southern blot hybridization patterns (Fig. 6A), PCR amplification with the primers HE-R and HE-L also yielded positive results, indicating that cells carrying the intact hlyA gene were still present in all five clones tested. In conclusion, although integration events could be documented, we were not able to isolate a clone with a disrupted hlyA gene.

FIG. 6.

Detection of M. pulmonis hlyA disruption by Southern blot and PCR. (A) Southern blot hybridization using an hlyA probe with HindIII-digested genomic DNA from M. pulmonis clones transformed with the plasmid pMPO5-ΔhlyA (lanes T1 to T5). NT, nontransformed HindIII-digested M. pulmonis DNA; P, pMPO5-ΔhlyA plasmid DNA. (B) Agarose gel electrophoresis of PCR products obtained using either the primers OR2 and HE-R (lanes a) or the primers HE-L and HE-R (lanes b). The PCR with primers OR2 and HE-R will yield a PCR product only if there is integration of the vector pMPO5-ΔhlyA into the M. pulmonis hlyA gene. The PCR with primers HE-L and HE-R will yield a product only if the chromosomal hlyA is not disrupted. Target DNAs included the purified DNA from 3 M. pulmonis clones transformed with pMPO5-ΔhlyA (lanes T1 to T3), DNA from nontransformed M. pulmonis culture (NT), DNA from pMPO5-ΔhlyA plasmid DNA alone (P), a negative control (NC) with no DNA added, and a mixture of 2 ng of the pMPO5-ΔhlyA plasmid with 50 ng of nontransformed M. pulmonis DNA (NT+P). M, 100-bp DNA ladder (Gibco-Life Technologies).

DISCUSSION

The complete sequencing of the M. pulmonis genome has opened the way to functional genomics for this rodent pathogen, which is considered a model of respiratory mycoplasmosis (4, 38). However, due to the paucity of genetic tools, genetic studies could not be conducted so far. Transformation of M. pulmonis to antibiotic resistance has been achieved with Tn916 (10) and Tn4001 (11), but there is still no suitable vector for cloning genes into this organism. With the aim of developing plasmid vectors for M. pulmonis, we have constructed oriC plasmids, a strategy which was successfully used with the plant mollicute S. citri (41). As a preliminary step in identifying M. pulmonis oriC, we have compared the putative oriC regions of mollicutes for which sequence data are available. The results indicated a lack of gene order conservation in this region. For example, the tandem of the dnaA-dnaN genes is conserved in S. citri, M. capricolum, and M. pulmonis, whereas these genes are divergent and spaced by the soJ gene in the M. pneumoniae and M. genitalium genomes. The U. urealyticum genome offers even more diversity because these two genes are 95 kpb apart.

Using software developed for finding conserved motifs, we were able to locate potential DnaA boxes in the vicinity of dnaA. The distribution of putative DnaA boxes on both sides of dnaA was found to be similar for S. citri, M. capricolum, and M. pulmonis but differed for the other three examined mollicute genomes. In M. pulmonis in particular, the two DnaA box regions flanking the dnaA gene were found to contain five and three DnaA boxes, respectively. Additional putative DnaA boxes with a maximum of two mismatches with the mollicute consensus were found within the dnaA gene sequences of M. pulmonis (eight boxes), M. capricolum (three boxes), M. genitalium (three boxes), and S. citri (six boxes).

The functional characterization of M. pulmonis oriC was achieved by transformation of the mycoplasma with pMPO1. In the mycoplasmal transformants, the plasmid was found to replicate as a free extrachromosomal element before integrating into the chromosome during passaging, similarly to the plasmid pBOT1 in S. citri (8). In spite of similarities between the oriC regions of M. pulmonis and S. citri, the plasmid pBOT1 failed to transform M. pulmonis. This result suggests that host-specific factors regulate the replication of these plasmids. This finding was not totally unexpected because it was described for other bacterial genera. As an example, Mycobacterium tuberculosis oriC-based plasmids do not replicate in rapidly growing mycobacterial species such as Mycobacterium smegmatis (32). Nevertheless, it remains to be determined if this barrier of specificity can be abolished in the case of mollicute species that are more phylogenetically related, such as S. citri and M. capricolum.

The search for the minimal oriC sequences required for plasmid replication showed that the two DnaA box regions are necessary to support replication in M. pulmonis. Although it is uncommon among gram-positive bacteria, a similar situation has been described for B. subtilis (29). With this organism, it was suggested that a DNA-loop structure involving protein-mediated interaction between the two DnaA box regions would be required for the initiation of replication (30). This hypothesis is supported by our results showing that cloning both the upstream and the downstream regions from M. pulmonis dnaA did not result in a replicative plasmid (pMPO4) unless these two regions were spaced, either by dnaA (pMPO1) or by tetM (pMPO5). For the mollicute S. citri (35) and for several gram-positive bacteria, including Streptomyces spp. (43), M. smegmatis, Mycobacterium bovis, M. tuberculosis, and Mycobacterium avium (27), the 3′ flanking region of dnaA was sufficient to support plasmid replication. Otherwise, the 5′ region can support plasmid replication for Pseudomonas putida and Pseudomonas aeruginosa (42). It is not possible to draw conclusions about the factors necessary for the function of chromosomal oriC from the analysis of cloned oriC, because it is known that they have different requirements (2).

The pMPO1 plasmid, which contains an oriC region encompassing the dnaA gene, was able to replicate in M. pulmonis but integrated, within a few in vitro passages, chromosomal oriC. Therefore, the use of this plasmid is limited to the targeted integration of genes into the oriC chromosomal region. In contrast, the pMPO5 plasmid remained free for at least 15 in vitro passages. It therefore represents the first shuttle plasmid vector functioning between E. coli and M. pulmonis.

The use of the oriC plasmid for targeting gene inactivation by homologous recombination was evaluated. The gene hlyA was chosen because its putative role in virulence suggested that it would be dispensable, at least in vitro. This gene is also not present in the genomes of M. pneumoniae and M. genitalium, which indicates that it is not essential for the life of these organisms. Although events of disruption of this gene could be detected in the cultured transformants with pMPO5-ΔhlyA, we were not able to isolate an hlyA mutant. This could be attributed to a polar effect resulting in the disruption of another gene, located downstream of hlyA and encoding an essential function. However, other attempts in our laboratory with other candidate genes, such as the mnuA gene (MYPU_6930), which encodes a putative nuclease, led to similar results, suggesting that other mechanisms may operate. The length of the homologous sequence could also be a critical factor for the crossing-over to occur between the plasmid-borne and chromosomal sequences. Although we cannot exclude this hypothesis, the construct for targeting the mnuA gene carried an 1,134-bp insert of mnuA homologous sequence, compared to 688 bp for pMPO5-ΔhlyA, and still did not allow the isolation of mnuA mutants. For these reasons, as the most probable explanation of our inability to isolate hlyA mutants using oriC-based plasmids, we favor a mechanism which would result from both the incompatibility of having two functional oriC regions on the chromosome and the resolution of the integrated genetic element. To solve the problem due to this incompatibility, we are currently trying to use a strategy based on a double crossing-over between the target chromosomal gene and the plasmid sequences. Furthermore, to select the supposedly rare events of integration, we are currently trying to counter-select the replication of the oriC plasmid, as proposed for other bacteria (36). Although a case of homologous recombination with M. pulmonis using suicide vectors has been reported (28), further studies indicated that these results were obtained with a strain that did not belong to the species M. pulmonis but belonged to the other mollicute species Acholeplasma oculi (1, 9, 28). Our results confirmed that homologous recombination for M. pulmonis using suicide vectors is difficult, if not impossible, to detect. This is in contrast with at least one other Mycoplasma species, M. genitalium, for which gene inactivation by this strategy has been obtained (6, 7).

If we except the cryptic unstable replicating plasmids previously described for M. capricolum and M. mycoides subsp. mycoides (23, 24), there are no other plasmids known to replicate in Mycoplasma species. In addition, M. pulmonis and S. citri are the only mollicutes for which oriC-based plasmids have been developed. Therefore, there is the possibility that this strategy may be applicable to other members of this bacterial class for which genetic tools have been scarce (13).