Development and Host Compatibility of Plasmids for Two Important Ruminant Pathogens, Mycoplasma bovis and Mycoplasma agalactiae

Mycoplasma bovis is a cause of pneumonia, mastitis, arthritis and otitis media in cattle throughout the world. However, despite its clinical significance, there is a paucity of tools to genetically manipulate it, impeding our capacity to further explore the molecular basis of its virulence. To address this limitation, we developed a series of homologous and heterologous replicable plasmids from M. bovis and M. agalactiae. The shortest replicable oriC plasmid based on the region downstream of dnaA in M. bovis was 247 bp and contained two DnaA boxes, while oriC plasmids based on the region downstream of dnaA in M. agalactiae strains 5632 and PG2 were 219 bp and 217 bp in length, respectively, and contained only a single DnaA box. The efficiency of transformation in M. bovis and M. agalactiae was inversely correlated with the size of the oriC region in the construct, and, in general, homologous oriC plasmids had a higher transformation efficiency than heterologous oriC plasmids. The larger pWholeoriC45 and pMM21-7 plasmids integrated into the genomic oriC region of M. bovis, while the smaller oriC plasmids remained extrachromosomal for up to 20 serial passages in selective media. Although specific gene disruptions were not be achieved in M. bovis in this study, the oriC plasmids developed here could still be useful as tools in complementation studies and for expression of exogenous genes in both M. bovis and M. agalactiae.


Introduction
Mycoplasma bovis is a major cause of chronic caseonecrotic bronchopneumonia in calves [1,2]. Its contribution to bovine respiratory tract disease is commonly complicated by concurrent infection with a number of viral and bacterial pathogens, including bovine herpesvirus type 1, parainfluenza virus type 3, bovine respiratory syncytial virus, bovine pestivirus, bovine adenovirus, Mannheimia haemolytica, Pasteurella multocida and Histophilus somni. M. bovis is also an important cause of mastitis and lowered milk production, and can cause arthritis, otitis media, keratoconjunctivitis, meningitis, abortion and infertility [3,4]. It has been reported to be an emerging global pathogen [3][4][5][6], with estimated annual losses of €144-192 million attributed to its contribution to bovine pneumonia in Europe [4] and of $108 million to its contribution to bovine mastitis in the USA [7].
The genome sequence of the M. bovis type strain PG45, as well as the Hubei-1 and HB0801 strains, have been completed and published recently [8][9][10], but the paucity of genetic tools for generating targeted gene knockouts and for complementation studies impedes our ability to examine gene function in M. bovis. Transposon mutagenesis using Tn916 and Tn4001 and their derivatives has been used to study the genetics of several human and animal mycoplasmas, but the random insertion of a transposon in the genome of an organism does not allow specific targeting of a gene of interest [11,12].
Although homologous recombination with linear DNA has been achieved occasionally in some mycoplasmas, targeted gene disruption through homologous recombination remains the main alternative to transposons for mutagenesis studies. The classical method of using non-replicable plasmids to disrupt gene targets through homologous recombination has been successful only in a few mollicutes, including Acholeplasma laidlawii [13], M. gallisepticum [14][15][16] and M. genitalium [17,18], probably because the likelihood of allelic exchange is reduced by the low transformation efficiencies and low recombination frequencies seen in mycoplasmas.
In the present study, a series of plasmids containing the oriC regions of M. bovis and M. agalactiae fused with a tetracycline resistance marker were developed. Using these constructs functional analyses of homologous and heterologous oriC plasmids were conducted in M. bovis strain PG45 and M. agalactiae strain PG2 to determine the optimal oriC regions to include in plasmid constructs for these two species.

Bacterial strains and culture conditions
The M. bovis type strain PG45 was grown at 37°C in modified Frey's broth or on mycoplasma agar plates [26], while the M. agalactiae type strain PG2 was grown at 37°C in Aluotto broth or on Aluotto agar [27]. M. bovis and M. agalactiae transformants were selected on plates containing tetracycline at 5 μg/ml and 2 μg/ml (Sigma Aldrich), respectively. E. coli DH5α cells (Life Technologies) were used for subcloning and for amplifying different oriC plasmids.

Construction of oriC plasmids from M. bovis
Several regions of the predicted oriC region of M. bovis strain PG45 were amplified by PCR to generate the plasmids pWholeoriC45, pIRR45 and pIRL45 using appropriate oligonucleotide primers (Invitrogen), as listed in Table 1. Initially an oriC region including the dnaA gene and both intergenic regions, designated whole oriC (1.8 kbp), was amplified using the primers OMBIRL for/OMBIRR rev. The left (IRL) and right intergenic (IRR) regions were amplified separately, without the intervening dnaA gene, using the primer pairs OMBIRL for/OMBIRL rev and OMBIRR for/OMBIRR rev, respectively. Each 50 μl PCR contained 1 x Hi Fi reaction buffer, 200 μM of each deoxynucleotide triphosphate, 2 mM MgSO 4 , 250 nM of each primer, 2.5 U Platinum Taq High Fidelity DNA polymerase (Invitrogen), template DNA, and water to make the final volume up to 50 μl.
For amplification of the whole oriC region, 20 ng of genomic DNA was used as template and the reaction was incubated in a thermocycler (iCycler, Bio-Rad) at 94°C for 5 min, then through 40 cycles of 94°C for 30 sec, 50°C for 2 min and 68°C for 4 min, and finally at 68°C for 7 min. For amplification of the IRL and IRR regions, the DNA template was prepared by pelleting 1 ml of culture of M. bovis PG45 and resuspending the cells in 100 μl distilled water, then heating the suspension at 95°C for 5 min. A 2 μl volume of this was used as template in the PCR reaction. For amplification of the intergenic regions, the conditions were: 95°C for 2 min, followed by 35 cycles of 95°C for 30 sec, 50°C for 30 sec and 68°C for 20 sec, with a final extension at 68°C for 5 min. PCR products were separated by agarose gel electrophoresis, the expected bands excised and DNA extracted from the gel slices. All the products were ligated separately into the T/A-cloning site ( Fig. 1) of pGEM-T (Promega), the clones selected on the basis of blue-white screening and the sequence of the insert confirmed by DNA sequencing using ABI PRISM Big Dye 3.1 Terminator chemistry (Applied Biosystems). The tetracycline resistance gene (tetM) with its own promoter and terminator was digested from the plasmid pMlori [15] using the restriction endonuclease SpeI (New England Biolabs) and ligated into the SpeI cleavage site of the pGEM-T construct.

Construction of oriC plasmids from M. agalactiae
The heterologous oriC plasmid pMM21-7 containing the 1.3 kbp oriC region of M. agalactiae strain 5632 [28], developed for transformation of M. agalactiae, was initially used to transform M. bovis. Subsequently, plasmid p20-1miniO/T [20] with a reduced oriC region (700 bp) from M. agalactiae strain 5632 was used to transform M. bovis and M. agalactiae. To determine the minimal replicable oriC region of M. agalactiae, pIRR5632 and pIRRPG2 constructs were developed based on the IRR regions of M. agalactiae strains 5632 and PG2, respectively, using the primers OMAIRR for/OMAIRR rev ( Table 1). The PCR composition and thermocycling conditions were similar to those used for amplifying the intergenic regions of M. bovis. The amplified IRR regions were cloned separately into the T/A-cloning site ( Fig. 1) of pGEM-T and the tetM gene was then ligated into the SpeI cleavage site of the vector. To examine whether the size of the oriC region capable of replication could be reduced further, several oriC regions were amplified (S1 Fig.) using plasmid pIRR5632 as template. These included the 133 bp intergenic region, as well as regions with only a single DnaA box and nearby AT rich clusters. The amplification of these 133, 90 and 66 bp regions was performed with the primers listed in Table 1 and reactions were incubated at 95°C for 3 min, followed by 35 cycles of 95°C for 30 sec, 45°C for 20 sec and 68°C for 20 sec, with a final extension at 68°C for 5 min, while a 38 bp region was generated by oligonucleotide annealing, starting at 95°C for 3 min then slowly cooling to 25°C over a 45 min period. These regions were cloned separately into pGEM-T and the tetM gene was then ligated into the construct as described above.

Transformation of M. bovis and M. agalactiae
Approximately 5 μg of DNA of each plasmid was introduced into M. bovis strain PG45 cells at late log phase by electroporation [26,29], with 5 μg tetracycline/ml used for selection in broth cultures and on agar plates.
Competent M. agalactiae strain PG2 cells were prepared from late log phase cultures. For transformation, 3 μg of DNA of each plasmid was mixed with 20 μl electrocompetent cells (approximately 10 8 cells) and 20 μl ice-cold HEPES-sucrose buffer (8 mM HEPES, 272 mM sucrose, pH 7.4) in a Gene Pulser cuvette (Bio-Rad) with a 0.2 cm electrode gap. The remainder of the transformation protocol was similar to that used for M. bovis strain PG45, with the exception that tetracycline was used at a concentration of 2 μg/ml in Aluotto broth and plates.

Determination of transformation efficiency
To calculate the efficiency of transformation with different oriC plasmids, a sample of the transformed culture was taken after 2 h incubation at 37°C in non-selective broth and inoculated onto an agar plate to assess the total number of colony forming units (CFU) in the absence of antibiotic selection, and the remaining cells were plated onto selective agar plates containing an appropriate amount of tetracycline. To confirm the presence of the plasmid, a tetM screening PCR [26] was performed with the primers LAtetM for and LBtetM rev (Table 1). In the presence of free or integrated oriC plasmids, this PCR generated a 339 bp tetM product that could be detected following agarose gel electrophoresis.

Design of constructs to engineer gene knockouts
To construct an oriC plasmid that could integrate into a target gene by homologous recombination, internal fragments of xer1, and the p48 and the type II restriction endonuclease genes were amplified separately from M. bovis (Table 1), with NotI and PstI restriction endonuclease cleavage sites incorporated into the forward and reverse primers, respectively, and inserted into the IRR-based oriC plasmids (S2 Fig.). For amplification of 392 and 462 bp internal fragments of the p48 and type II restriction endonuclease genes, the PCR was incubated through one cycle of 94°C for 3 min, followed by 35 cycles of 94°C for 45 sec, 55°C for 45 sec and 68°C for 30 sec, with a final extension at 68°C of 5 min, while the amplification of the 251 bp xer1 internal fragment was carried out by incubation through one cycle of 94°C for 2 min, followed by 35 cycles of 94°C for 30 sec, 50°C for 30 sec and 68°C for 20 sec, with a final extension of 68°C for 5 min. A heterologous recA gene, along with its promoter and terminator, was amplified from M. gallisepticum strain S6 with the primer pair GKoriCrecA for/GKoriCrecA rev (Table 1) and ligated between the PstI and SalI cleavage sites (S2 Fig.) in oriC plasmids that already contained engineered gene fragments. The amplification reaction was incubated through one cycle of 94°C for 5 min, followed by 40 cycles of 94°C for 30 sec, 50°C for 30 sec and 68°C for 1.4 min, with a final extension at 68°C for 7 min. The xer1 and p48 gene fragments were inserted into pIRR45, while the type II restriction endonuclease gene fragment was introduced into pIRRPG2.

DNA isolation, digestion and Southern hybridisation
To evaluate chromosomal integration, the presence of the free replicating form of the plasmids and gene target disruption in M. bovis, independent transformants were selected and passaged in selective media containing tetracycline at 5 μg/ml. DNA from M. bovis strain PG45 was isolated by phenol-chloroform extraction [30]. Control plasmid DNA was isolated from E. coli DH5α cells and total genomic DNA from transformant cells was extracted using the High Pure PCR template preparation kit (Roche).
Approximately 3 μg of genomic DNA from the wild type and M. bovis strain PG45 transformants was digested at 37°C for 16 h. After separation in a 0.8% agarose gel containing 0.1 μg ethidium bromide/ml, the fragments were transferred to Hybond-N + nylon membrane (GE Healthcare) by capillary action [30]. The membrane was washed with 6 x SSC (1 x SSC consists of 150 mM NaCl, 15 mM sodium citrate, pH 7.0) for 10 min, air-dried and the transferred DNA fixed to the membrane by exposure to UV light for 4 min. The membrane was prehybridised with 15 ml of DIG Easy Hyb hybridisation buffer (Roche) in a roller bottle at 42°C for 30 min in a hybridisation oven (Hybaid). Approximately 250 ng of the digoxigenin (DIG)-labelled probe was denatured at 95°C for 5 min, immediately cooled on ice, then added to the hybridisation buffer and incubated with the membrane overnight. The membrane was washed twice in low stringency wash buffer (2 x SSC, 0.1% SDS) at RT for 5 min each and then twice in higher stringency buffer (0.5 x SSC, 0.1% SDS) at 65°C for 15 min each. After cooling to RT, the membrane was washed briefly in washing buffer (0.1 M maleic acid, 0.15 M NaCl, pH 7.5, and 0.3% v/v Tween-20), incubated for 30 min in 1 x blocking buffer (1% w/v blocking reagent in maleic acid buffer) at RT, then a dilution of 1:5000 of an anti-DIG antibody coupled to alkaline phosphatase (Roche) in 1 x blocking buffer added and incubated with the blot for 30 min at RT. The membrane was washed twice in washing buffer for 30 min each at RT and then equilibrated in detection buffer (0.1 M Tris-HCl, 0.1 M NaCl, pH 9.5) for about 5 min. Hybridisation was detected by addition of the chemiluminescent substrate disodium 3-(4-methoxyspiro {1,2dioxetane-3,2'-(5'-chloro)tricyclo[3.3.1.13,7]decan}-4-yl) phenyl phosphate (CSPD, Roche) as recommended by the manufacturer and exposing the membrane to X-ray film (Kodak), which was developed using an Agfa automated film processor.
Southern blot hybridisation analysis was performed with DIG-labelled probes targeting the 247 bp IRR oriC region of M. bovis strain PG45 (oriC probe), the 339 bp tetM region or the 862 bp ampicillin resistance gene (ampR) region, and overnight prehybridisation was performed either at 40°C (oriC probe) or 46°C (tetM and ampR probe) and a high stringency wash step of 15 min each was performed twice at 62°C (oriC probe) or 65°C (tetM and ampR probes).

Replicability of homologous and heterologous oriC plasmids
M. bovis strain PG45 was transformed with the homologous oriC plasmid pWholeoriC45, containing the dnaA gene and its upstream and downstream intergenic regions along with the tetM determinant, which confers tetracycline resistance in Mollicutes. As a control, M. agalactiae strain PG2 was transformed with the homologous oriC plasmid pMM21-7 [28], containing the 1.3 kbp oriC region of M. agalactiae strain 5632. To examine the replicability of heterologous oriC plasmids in M. bovis and M. agalactiae, the pMM21-7 plasmid or its derivative p20-1miniO/T were used to transform M. bovis strain PG45 cells, while M. agalactiae strain PG2 cells were transformed with the pWholeoriC45 plasmid.
Mycoplasma cells transformed with the homologous and heterologous constructs generated colonies on selective agar plates containing an appropriate concentration of tetracycline. The individual clones were selected and grown in appropriate selective broth. PCR analyses confirmed that tetracycline resistance was correlated with the presence of the tetM determinant in all the transformants, as detectable by amplification of a predicted 339 bp PCR product. These results indicated that the oriC region capable of replicating in M. bovis is located within the dnaA gene or its flanking regions.

Minimal functional oriC region
In order to determine the smallest oriC region of M. bovis required for plasmid replication, M. bovis and M. agalactiae cells were transformed with the pIRL45 and pIRR45 plasmids, both of which lacked the dnaA gene (Fig. 1). The pIRR45 plasmid, which contained the IRR region downstream of dnaA, was consistently able to transform both species, whilst the pIRL45 plasmid, which contained the region upstream of the dnaA gene, was repeatedly unable to transform M. bovis ( Table 2). The oriC region of pIRR45 consisted of a 247 bp region with two AT rich regions of 17 bp each and two putative DnaA boxes that differ from the consensus sequence [TTATCCACA] by one or two bases. The region was amplified using primers OMBIRR for/OMBIRR rev, which complement sequences in the dnaA and dnaN genes located upstream and downstream of the right intergenic region. Consequently, this 247 bp oriC region included the 133 bp IRR, as well as 24 bp of the dnaA gene and 90 bp of the dnaN gene at either end. The pIRL45 plasmid (Fig. 1), which was unable to transform M. bovis, contained the intergenic region between the recD and dnaA genes and contained a 236 bp region with two AT rich Plasmids pIRR5632 and pIRRPG2, which were developed from M. agalactiae strains 5632 and PG2, respectively, were able to replicate in both M. bovis and M. agalactiae. The minimal oriC regions of M. agalactiae strains 5632 and PG2 were 219 bp and 217 bp, respectively, with each oriC region containing two AT rich regions of 17 bp each and a single DnaA box with a sequence similar to that of the consensus sequence (Fig. 1). As plasmid pIRR5632 replicated efficiently in M. bovis and M. agalactiae and had only a single DnaA box, an attempt was made to further reduce the size of the oriC region by generating a series of plasmids containing the DnaA box and the nearby AT rich regions (S1 Fig.), but all these plasmids failed to replicate in either species. This indicated that the minimal oriC region of M. agalactiae capable of initiating replication in these ruminant pathogens was 217 bp and included the 133 bp IRR together with 30 bp from the dnaA gene and 54 bp from the dnaN gene at either end.

Efficiency of transformation with oriC plasmids in M. bovis and M. agalactiae
The pseudoresistant clones previously reported in studies using transposon mutagenesis in M. bovis [29] were not detected in the studies reported here. Therefore, it was possible to calculate the transformation efficiency, as there was no need to grow cultures in selective broth overnight before subculturing onto agar plates, which would have resulted in a biased calculation. There was a marked difference in transformation efficiency (Table 2) depending on the size and origin of the oriC plasmids. The transformation efficiency increased markedly with a decrease in the size of the oriC region in the construct, and, with the exception of pIRRPG2, homologous oriC plasmids were able to transform more efficiently than heterologous oriC plasmids. Plasmid pIRR5632 had a higher transformation efficiency in M. bovis (1.68×10 −6 ) and in M. agalactiae (2.05×10 −6 ). The larger pMM21-7 and pWholeoriC45 plasmids resulted in higher transformation efficiencies in homologous cultures, but with a decrease in size of the oriC region even heterologous plasmids were able to generate a considerable number of tetracycline resistant clones. Therefore, transformation efficiency was higher with homologous and smaller plasmids, although it was not possible to reduce the size of replicable oriC region beyond that of the smallest plasmids assessed here.

Homologous recombination of the longer oriC plasmids in M. bovis
The presence of free, as well as integrated, plasmid in the transformants was demonstrated by Southern blot hybridisation using DIG-labelled tetM, ampR or M. bovis specific oriC probes. Two independent transformants generated using the pWholeoriC45 and the pMM21-7 plasmids in M. bovis strain PG45 were passaged by 1:5 dilution at each passage for up to 20 times in the presence of tetracycline and sampled at the 5 th , 10 th , 15 th and 20 th passage.
The pWholeoriC45 plasmid had no EcoRI site and only a single PstI site in the cloning site of pGEM-T, so EcoRI-PstI digestion of the extrachromosomal form of the plasmid should yield a product of 7.1 kbp, while the chromosomal oriC in untransformed cells lies in an EcoRI fragment of 6.6 kbp. As pWholeoriC45 contained a unique PstI site, plasmid integration via a single homologous recombination event at the chromosomal oriC locus was expected to result in two oriC copies that would be separated into two fragments upon digestion with EcoRI and PstI, and their location could be further confirmed using ampR and tetM probes. In pWholeoriC45 transformants digested with a combination of EcoRI and PstI (Fig. 2), the detection of 8.2 and 5.5 kbp fragments by the oriC probe indicated that integration of pWholeoriC45 into the chromosomal DNA had occurred. In contrast, the detection of 6.6 and 7.1 kbp fragments indicated that some of the transformed cells still contained the wild-type chromosomal oriC and free plasmid. At passage 5, the presence of four bands hybridising to the oriC probe suggested that a mixed population of cells was present. There was evidence of partial integration of the plasmid within the oriC region even at the 5 th passage in both of the transformants, and extrachromosomal plasmid was present till the 10 th passage in transformant 1 (Fig. 2, panel A) and the 15 th passage in transformant 2 (results not shown), after which no free plasmid could be detected. The blot was further probed with either the ampR or tetM probes. Hybridising fragments of 7.1 and 5.5 kbp were detected with the ampR probe (Fig. 2, panel B), and 7.1 and 8.2 kbp fragments with the tetM probe (Fig. 2, panel C), demonstrating the presence of both free and integrated forms of pWholeoriC45.
The DNA of M. bovis pMM21-7 transformants was digested with either ClaI or EcoRI and hybridised with the ampR probe. In clone 1, plasmid was maintained extrachromosomally till the 20 th passage, while in clone 2 both free and integrated forms of plasmids were visible at the 5 th passage, and by the 10 th passage the plasmid appeared to have completely integrated into the genome (Fig. 3, panels A and B).

Stability of smaller oriC plasmids in M. bovis
To evaluate the stability of the smaller homologous and heterologous plasmids in M. bovis, three or more independent transformants obtained with each of the plasmids p20-1miniO/T, pIRR45, pIRRPG2 and pIRR5632 were passaged. The DNA of p20-1miniO/T transformants was digested with ClaI, while that of transformants with plasmids based on the IRR region were digested with EcoRI and PstI. In contrast to the transformants obtained with the larger oriC plasmids, transformants containing the p20-1miniO/T (Fig. 3, panel C) and smaller IRR based oriC plasmids had only a single hybridising band (Fig. 4, panel A), corresponding to the size expected for the free plasmid, when probed with either tetM or ampR gene fragments, whilst two bands, corresponding to free IRR based plasmid and the intact chromosomal oriC region, were detected when they were hybridised with the oriC probe (Fig. 4, panel B). All the randomly selected clones stably maintained the oriC plasmids till the 20 th passage. The oriC probe generated strong hybridisation signals with the chromosomal oriC region as well as the extrachromosomal plasmid, while the tetM and ampR probes bound very weakly to the chromosomal oriC region in some clones, in addition to the intense signal from bands corresponding to extrachromosomal plasmid. This may suggest a low level of integration of the plasmids at the oriC region, but could not be explained by a single crossover event, as this would yield predicted hybridising bands of 4.0 or 8.8 kbp with the ampR probe.

Assessment of gene disruption by homologous recombination
In an attempt to inactivate gene targets, engineered constructs containing internal fragments of the xer1, p48 or type II restriction endonuclease genes were generated (S2 Fig.). The recA gene of M. gallisepticum strain S6, along with its promoter, was incorporated into the plasmid in order to promote homologous recombination. The oriC plasmids based on the IRR region were chosen for gene disruption as these were found to be present extrachromosomally till the 20 th passage. Following electroporation of these plasmids into M. bovis, 3 to 5 clones were selected from each transformation and passaged 10 times at a 1:9 dilution each passage. The M. gallisepticum recA gene contained an EcoRI site that was unique within the plasmids, so DNA of transformants obtained using the engineered constructs were digested with EcoRI. Southern blot analysis was performed using the ampR probe. The oriC plasmids containing the internal gene fragments and the heterologous recA replicated as free plasmids in the transformants. The ampR probe bound to a 7.4 kbp band, indicating the presence of extrachromosomal plasmid, and a 9.4 kbp band (results not shown), corresponding to the predicted size of the fragment following integration of the plasmids into the chromosomal oriC region of M. bovis. However, no bands indicative of homologous recombination with the target genes were seen.

Discussion
In most bacteria, oriC is located in the vicinity of dnaA (S3 Fig.), and DNA replication starts with the specific interaction of DnaA with the 9 bp DnaA boxes in the oriC region. Consensus sequences (TTATCCACA; DnaA boxes) have been identified in the intergenic regions flanking the dnaA gene in mollicutes [31]. Replicable plasmids based on oriC regions have been successfully developed, including pMM21-7, which contains a 1.3 kbp oriC region, and its shorter derivative p20-1miniO/T, in M. agalactiae [20,28]. In the present study, M. bovis was initially transformed with the pWholeoriC45 plasmid, containing dnaA and upstream and downstream intergenic regions, making it the first homologous oriC plasmid capable of replicating in M. bovis. Moreover, M. bovis PG45 could be successfully transformed with the heterologous pMM21-7 oriC plasmid originally developed for M. agalactiae and its derivative p20-1miniO/  bovis. Schematic representation of genomic oriC region, IRR plasmids and predicted organisation of the oriC region following integration of different IRR plasmids. Southern blot analysis of EcoRI-PstI digested DNA probed with DIG-labelled probes targeting ampR (panel A) and oriC region (panel B). Smaller oriC plasmids were stable in M. bovis till p15-p20, but some integration appeared to have occurred in the oriC region. Lane 1, untransformed M. bovis; lanes, 2, 6 and 8, pIRR5632, pIRRPG2 and pIRR45, respectively. Lanes 3-5: pIRR5632 M. bovis transformants at p10, p15 and p20; lane 7, pIRRPG2 M. bovis transformants at p15; lanes 9-11, pIRR45 M. bovis transformants at p10, p15 and p20. Host specificity studies in mycoplasmas have suggested that the host range of oriC plasmids is restricted to closely related species [31]. The oriC plasmids of M. gallisepticum and M. imitans have been shown to be able to replicate in both species [15]. Similarly, oriC plasmids derived from M. mycoides subspecies mycoides and M. mycoides subspecies capri are able to replicate in each other, as well as in M. capricolum [19]. Conversely, attempts to transform M. agalactiae and members of the mycoides cluster with oriC plasmids derived from M. pulmonis were unsuccessful [19,28]. The nucleotide sequence of the replicable oriC region and the DnaA protein sequences of M. pulmonis are distinct from those of M. agalactiae and members of the mycoides cluster. A phylogenetic tree based on DnaA protein sequences of mycoplasmas (S4 Fig.) shows that there are close relationships between the DnaA proteins of M. bovis and M. agalactiae, between those of M. gallisepticum and M. imitans, and between those of various species in the mycoides cluster, including M. mycoides subspecies mycoides, M. mycoides subspecies capri and M. capricolum. The nucleotide sequences of the oriC regions in each of these groups of species exhibit high levels of similarity, which is probably one of the reasons for the compatibility of oriC plasmids derived from one species with other closely related species. The present study supports this hypothesis. The replicable oriC regions of M. bovis and M. agalactiae have more than 80% nucleotide sequence identity (S5 Fig.) and more than 95% DnaA protein identity. This could explain why M. bovis and M. agalactiae can support replication of their heterologous oriC plasmids.
To identify the smallest oriC region capable of replicating in M. bovis and M. agalactiae, a series of oriC plasmid constructs were developed. Only the plasmids containing the region downstream of dnaA could be detected in M. bovis and M. agalactiae, indicating that the AT rich sequences found in this region were essential for plasmid replication, even though the region upstream from the dnaA gene of M. bovis included two DnaA boxes with similar consensus sequences. This suggests that, at least in M. bovis and M. agalactiae, dnaA and the region upstream of it are not essential or sufficient for replication of oriC plasmids.
The minimal functional oriC region of M. bovis determined in this study was 247 bp in size, while those in M. agalactiae strains 5632 and PG2 were 219 bp and 217 bp, respectively; these regions are all located downstream of the dnaA gene. In two other members of the hominis phylogenetic group, M. pulmonis [22] and M. hyopneumoniae [32], the minimal region necessary for replication of an oriC plasmid devoid of dnaA contained both the upstream and downstream dnaA flanking intergenic regions.
The regions downstream of dnaA in M. bovis and M. agalactiae contain two DnaA boxes and a single DnaA box, respectively. It has been reported previously that only a few DnaA boxes and an AT rich region were sufficient for replication in mycoplasmas. For instance, a functional oriC vector for M. gallisepticum strain S6 contains two DnaA boxes and two AT rich clusters [15], and that for S. citri contains three DnaA boxes and two AT rich regions [33].
As the intergenic region downstream of dnaA in M. agalactiae strain 5632 contains a single DnaA box, the size of the oriC construct was reduced further. A number of smaller versions of oriC plasmids based on the replicable pIRR5632 plasmid were produced, but all failed to replicate in M. bovis or M. agalactiae. This showed that the intergenic region downstream of dnaA alone is not sufficient for optimal replication of oriC plasmids. In Mycobacterium smegmatis, it was found that replication of oriC plasmids lacking regions from either the 5' end or 3' end of adjacent genes was greatly reduced. It has been concluded that the 3' end of dnaA participates in chromosomal partitioning and is therefore essential for optimal activity of oriC plasmids in M. smegmatis, as well as in B. subtilis [34,35]. In S. citri, an oriC based plasmid containing three DnaA boxes but only one AT rich region failed to replicate, indicating that the AT cluster downstream of the DnaA boxes is essential for plasmid replication [33]. All oriC plasmids in the present study were generated with a pGEM-T backbone, and further examination of pGEM-T revealed at least 4 DnaA boxes, with the nucleotide sequence of one box exactly the same as the consensus sequence, further confirming that DnaA boxes alone are not sufficient for initiation of replication and that the location and size of the AT rich clusters, relative to DnaA boxes, is critical.
The transformation efficiency was found to be higher for homologous and smaller oriC plasmids. These results were in agreement with earlier studies that showed that a decrease in the size of the oriC region in a homologous plasmid resulted in higher transformation efficiencies in M. agalactiae [28]. The higher transformation efficiency observed with homologous plasmids compared to heterologous plasmids could be related to differences between the AT rich regions. It has been demonstrated in M. smegmatis that an AT rich sequence of a particular length is important for the function of an oriC, and that variation in the length of this AT rich region influenced transformation efficiency, and that even an increase in the length of the AT rich sequence was inhibitory [35].
In this study, the longer oriC regions of both homologous and heterologous plasmids readily integrated into the M. bovis chromosome. This has also been described in M. gallisepticum [15], S. citri [36] and M. pulmonis [22]. In order to produce oriC plasmids that would not integrate readily into the chromosome and could be used for targeted gene disruption, smaller constructs based on the IRR regions of M. bovis PG2 and M. agalactiae 5632 were produced and, following transformation, were found to be stable and remain in an extrachromosomal location up to the 20 th passage. Targeted gene disruption was attempted in M. bovis using IRR based oriC plasmids. Recombination events in mollicutes are dependent on the recA gene [13,37]. The recA gene from E. coli has been used in M. mycoides subspecies capri to generate knockout mutants [38]. The heterologous recA gene of M. gallisepticum strain S6 incorporated into our constructs was expected to promote homologous recombination, while having sufficient nucleotide sequence difference (with only 67% sequence identity between nucleotide 933 and 1342 bp) to reduce the likelihood of recombination with the M. bovis recA. However, our attempts to disrupt gene targets using stable oriC plasmids in M. bovis failed, with most plasmids integrating into the oriC region of the chromosome by the 10 th passage. While we were unable to confirm expression of the heterologous RecA as we lacked reagents to specifically detect its expression, the inclusion of the heterologous recA probably promoted homologous recombination within the oriC region, as plasmids devoid of recA were maintained in an extrachromosomal form up to the 20 th passage. Although replicable oriC plasmids have been successfully used to inactivate several gene targets in mollicutes, recombination at the target gene site is never straightforward. In most cases, the disruption vector integrates into the host chromosome in the oriC region, rather than at the target gene site, as was the case in our study. In some cases, oriC plasmids have been shown to disrupt the targeted gene, but the mutant containing the disrupted gene could not be isolated [22], and, in a few other cases, the plasmids have integrated randomly through illegitimate recombination, rather than through homologous recombination in the oriC region or in the gene targets [15]. Internal fragments of gene targets varying in size from 251 to 462 bp nucleotides were included in the plasmids. The size of these fragments was comparable to that of the oriC region, so there appeared to be a preference for the plasmid oriC region to recombine with the oriC region of the M. bovis chromosome rather than for the internal gene fragment to recombine with the target gene. It may be that binding of DnaA to oriC results in more stable single stranded regions for initiation of recombination than can be generated at the targeted genes by the heterologous RecA we included in the plasmids. In addition, the GC content of the oriC regions in the plasmids were 19 and 20%, while that of the internal fragment of p48 was 35%, and those of type II restriction endonuclease gene and xer1 were 27%. The lower GC content of the oriC region may thus have also contributed to the preferential recombination at the oriC region. It is possible that use of a larger internal fragment of the target genes may have enhanced recombination in the intended target.
Although specific gene targets were unable to be disrupted, the stable extrachromosomal plasmids developed in this study could also be used in M. bovis for complementation of disrupted genes function in transposon mutants and for expression of foreign genes.

Summary
This study has developed and demonstrated the potential of oriC vectors as genetic tools in two important ruminant mycoplasmas, M. bovis and M. agalactiae. The resistance marker, tetM, was found to be functional in both free and integrated forms of the plasmids.