|Home | About | Journals | Submit | Contact Us | Français|
We have made significant improvements to a broad-host-range system for the cloning and manipulation of large bacterial genomic regions based on site-specific recombination between directly repeated oriT sites during conjugation. Using two suicide capture vectors carrying flanking homology regions, oriT sites are recombined on either side of the target region. Using a broad-host-range conjugation helper plasmid, the region between the oriT sites is conjugated into an Escherichia coli recipient strain, where it is circularized and maintained as a chimeric mini-F vector. The cloned target region is functionalized in multiple ways to accommodate downstream manipulation. The target region is flanked with Gateway attB sites for recombination into other vectors and by rare 18-bp I-SceI restriction sites for subcloning. The Tn7-functionalized target can also be inserted at a naturally occurring chromosomal attTn7 site(s) or maintained as a broad-host-range plasmid for complementation or heterologous expression studies. We have used the oriTn7 capture technique to clone and complement Burkholderia pseudomallei genomic regions up to 140 kb in size and have created isogenic Burkholderia strains with various combinations of genomic islands. We believe this system will greatly aid the cloning and genetic analysis of genomic islands, biosynthetic gene clusters, and large open reading frames.
To aid the genetic characterization of strain-to-strain genomic variation and the heterologous production of natural products, the genome-scale manipulations required by synthetic biology necessitate the specific cloning and manipulation of DNA regions greater than 10 kb (1–4). However, PCR amplification- and restriction fragment-based cloning techniques are inefficient, inaccurate, and not always possible for regions of >10 kb. Several techniques have been developed to bridge this technical gap. The oriT-directed capture system developed by Chain et al. has advantages for the specific cloning of large genomic regions compared with site-specific recombinase-catalyzed “pop-out” and short-flank recombineering-based strategies (2, 5–10). The system is based on the principle that recombination will occur between two directly oriented RP4 origin of transfer (oriT) sites during conjugative transfer (11). The oriT-directed capture system utilizes two suicide vectors, one of which is mini-F replicon based, each with different selectable markers. Segments flanking the target region are cloned into the suicide vectors, which are then homologously recombined into the donor strain genome, resulting in directly oriented oriT sites flanking the target region. After transformation of this target donor strain with a conjugation helper plasmid, the conjugation machinery encoded by this plasmid recognizes one of the oriT sites integrated into the genome and begins transfer in a directional manner of downstream genomic DNA, including the target region and the mini-F replicon, into an Escherichia coli recipient. Transfer is terminated upon reaching the second oriT site, which is then recombined with the initiating oriT in the E. coli recipient, thus generating a circular mini-F-based plasmid carrying the target genome region. The system developed by Chain et al. (5) was used to clone specific regions in excess of 200 kb and does not rely on the location of flanking restriction sites required by RecE-based recombineering strategies (2, 5). The oriT-directed capture system does not generate deletions in the target chromosome during capture, because only one strand of the genome is mobilized. This prevents any fitness effects that might be caused by large deletions seen in site-specific recombinase-based pop-out systems (10). The circularized capture vector is simultaneously generated and recovered as an E. coli transconjugant. This eliminates the need for organism-specific conditional replicons, additional plasmid rescue steps, and the potential difficulties associated with the recovery of large plasmids. The captured DNA is recovered directly from the donor genome, which eliminates the introduction of point mutations. Finally, all steps involving the handling of large DNA are conducted in vivo, which reduces issues of large fragment DNA instability and large vector transformation inefficiency. This system should be functional in any organism that can be established as a conjugation donor.
We have engineered significant improvements into the oriT-directed capture system for increased functionality and the downstream manipulation of captured target regions. Most notably, we created new broad-host-range conjugation helpers and have incorporated mini-Tn7 functionalization for single-copy genomic integration to create the oriTn7 capture system. The transposon Tn7 is capable of integrating site-specifically into a naturally evolved attTn7 genomic site located downstream of highly conserved glmS genes (12–14). The final pFTarget vector incorporates portions from each capture vector so that the large target region is carried by a complete mini-Tn7 transposon. We have built the oriTn7 capture system for use in Burkholderia pseudomallei, the causative agent of melioidosis and a category B select agent. There is extensive strain-to-strain variation within the B. pseudomallei pangenome, much of this in genomic regions of difference greater than 10 kb, and many of these regions are genomic islands (15–21). The characterization of natural gene loss, gene acquisition, and strain-to-strain variability is likely to be important in understanding B. pseudomallei pathogenicity and phenotypic diversity. This characterization requires the creation of isogenic strains carrying specific deletions or insertions of the genomic regions of difference. We have successfully used the oriTn7 capture system to specifically clone genomic regions from 12 kb to 140 kb corresponding to a region absent from the human clinical B. pseudomallei isolate 708a and have used Tn7 integration to complement the matching 140-kb deletion of this region. In addition, we have used oriTn7 capture to create a set of isogenic Burkholderia strains carrying various combinations of two B. pseudomallei gene clusters whose presence is geographically biased in strains from different regions where this pathogen is endemic. Although in its current form, the oriTn7 capture system was built specifically for use in B. pseudomallei and related species, it was designed with portability in mind and could be readily adapted for use in other bacteria.
Strains used in this study are listed in Table S1 in the supplemental material. Bacteria were cultured in liquid or on agar-solidified Lennox LB (Mo Bio Laboratories, Carlsbad, CA) at 37°C with aeration. Burkholderia strains carrying temperature-sensitive plasmids were maintained at 30°C, and plasmid curing was achieved by incubation at 42°C in the absence of selection. All procedures involving B. pseudomallei were performed in approved select agent biosafety level 3 (BSL3) facilities in the Rocky Mountain Regional Biosafety Laboratory (CSU) using approved procedures and protocols compliant with regulations regarding select agents. For E. coli cultures, media were supplemented with antibiotics at the following final concentrations: ampicillin (Amp), 100 μg/ml; kanamycin (Km), 40 μg/ml; zeocin (Zeo), 25 μg/ml; trimethoprim (Tmp), 100 μg/ml; chloramphenicol (Cm), 25 μg/ml; streptomycin (Sm), 50 μg/ml; gentamicin (Gm), 10 μg/ml; rifampin (Rf), 50 μg/ml; polymyxin B (PmB), 15 μg/ml; 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) and 5-bromo-4-chloro-3-indolyl glucuronide (X-Gluc), 40 μg/ml; and diaminopimelic acid (DAP), 200 to 400 μg/ml. For Pseudomonas syringae cultures, media were supplemented with Rf at a final concentration of 50 μg/ml. For culture of Burkholderia thailandensis, media were supplemented with antibiotics at the following final concentrations: Km, 40 μg/ml; Zeo, 25 μg/ml; Gm, 10 μg/ml; and trimethoprim (Tmp), 100 μg/ml. For B. thailandensis AmrAB-OprA efflux pump-expressing strains, antibiotic concentrations were adjusted to 500 to 1,000 μg/ml Km and 1,000 to 2,000 μg/ml Zeo. For B. pseudomallei, concentrations were as follows: Km, 40 μg/ml; Zeo, 25 μg/ml, Gm, 10 μg/ml; and PmB, 15 μg/ml. For B. pseudomallei AmrAB-OprA efflux pump-expressing strains, antibiotic concentrations were adjusted to 500 to 1,000 μg/ml Km and 1,000 to 2,000 μg/ml Zeo. Counterselection of B. pseudomallei was done with Tmp at 100 μg/ml and Rf at 50 μg/ml. Arabinose (ARA) or rhamnose (RHA) was used at a final concentration of 0.2% to induce gene expression from plasmids contained in Burkholderia spp.
Plasmid DNA was purified using the GeneJET plasmid miniprep kit (Fermentas, Glen Burnie, MD). Cleanup of DNA enzymatic reactions and gel extractions were conducted using the GenElute gel extraction kit (Sigma-Aldrich, St. Louis, MO). Bacterial genomic DNA was isolated using the Puregene genomic DNA purification kit (Gentra Systems, Qiagen, Valencia, CA). DNA boiling preparations for PCR analysis were obtained by combining 1 μl of overnight culture with 30 μl distilled water (dH2O) or picking a bacterial colony or patch with a sterile toothpick followed by suspension in 30 μl dH2O and incubation at 100°C for 10 min. Site-directed mutagenesis was conducted with the QuikChange site-directed mutagenesis kit according to the manufacturer's recommendations (Stratagene, Agilent Technologies, Santa Clara, CA). DNA restriction endonucleases, T4 ligase, calf alkaline intestinal phosphatase, and Taq polymerase were obtained from New England BioLabs (Ipswich, MA) and used according to the manufacturer's recommendations. HiLo and λ DNA monocut mix DNA ladders were obtained from Minnesota Molecular (Minneapolis, MN) and New England BioLabs, respectively. Oligonucleotides were obtained from Integrated DNA Technologies (Coralville, IA) and are listed in Table S2 in the supplemental material. DNA sequencing was conducted using an ABI 3130×L genetic analyzer (Applied Biosystems, Carlsbad, CA) at the Colorado State University Proteomics and Metabolomics Facility. For multiplex PCR, oligonucleotide primers for each probe were designed to have annealing temperature differences of 1°C and amplicon size differences of 100 bp. Multiplex oligonucleotides were combined into a master mix and used at a 4 to 12 μM final concentration. Southern blot analysis was performed using the NEBlot Phototope and Phototope-Star chemiluminescent labeling and detection kits from New England BioLabs following the manufacturer's recommendations and using standard capillary transfer and blotting procedures (22).
Plasmid transformation of E. coli was done by using either standard electroporation or chemical transformation procedures (22). Electrotransformation of B. pseudomallei was conducted as described previously (23). Briefly, overnight cultures were harvested by centrifugation, washed three times with 300 mM sterile sucrose, and concentrated 10-fold prior to electroporation with 0.2 to 1.0 μg plasmid DNA.
Bacterial conjugations were conducted as either biparental or multiparental matings with RHO3 (or derivative) conjugation donors or pRK2013 (or derivative) conjugation helper plasmids. One to three ml overnight cultures was harvested by centrifugation, washed twice with fresh LB, and concentrated, typically 5-fold, to 109 to 1010 CFU/ml. Equal parts of each parent strain were combined and applied to sterile cellulose acetate or nitrocellulose filters placed on LB plates (augmented with DAP for conjugations including RHO3 donor strains) with each parent strain spotted individually as a control. Spots were allowed to dry, and plates were incubated overnight at 37°C unless otherwise noted. Cells were recovered from the filters by centrifugation in 1.5-ml microcentrifuge tubes with 1 ml of LB, washed with 1 ml LB, and spread onto LB plates with appropriate selection and counterselection to recover the desired exconjugants.
Plasmids used in this study are listed in Table S3 in the supplemental material. Plasmid and strain construction details are provided in the supplemental methods.
Left and right flanking regions of homology to target regions were amplified by PCR and cloned into directional pENTR vectors (see the supplemental methods). LR recombination between the right flanking fragment containing pENTR vectors and the pUCTCAPR-GW right capture vectors was facilitated by the compatible Gmr and Zeor markers carried by pUCTCAPR. The pENTR vectors carrying various right flanking fragments were LR recombined with pUCTCAPR vectors using either LR Clonase or LR Clonase II from Invitrogen Life Technologies according to the manufacturer's recommendations. LR recombinants were recovered on LB+Gm plates, screened for Zeor, and confirmed by NotI and/or BsrGI restriction enzyme digests.
The design of the left capture vector pFTCAPL-GW required the incorporation of a Kmr resistance marker to comply with select agent regulations for B. pseudomallei. Unfortunately, this conflicts with the Kmr marker carried by directional pENTR/D-TOPO vectors. Two strategies were used to overcome the Kmr marker conflict between the pENTR vectors and pFTCAPL vector. In the first strategy, pENTR vectors were digested with NsiI, and their Kmr markers were replaced with a Tmpr marker by ligation with the 0.9-kb PstI Tmpr marker fragment from pFTP2 to create pENTRTmp vectors. This allowed the recovery E. coli EPI300 clones on LB+Km plates with correct LR recombinants without recovering a Kmr pENTR background. In the second strategy, pENTR inserts were PCR amplified with primers M13F+M13R, which appends attL1 and attL2 sites to the cloned inserts. The linear PCR products were then gel purified and used in LR reactions with pFTCAPL-GW before transformation into E. coli EPI3000 and recovery on LB+Km plates. This PCR-based strategy is the one that we currently favor for the LR recombination of both left and right capture vectors, as it is both faster and less cumbersome. EPI300 pFTCAPL LR recombinant clones were induced to high copy numbers with 0.2% arabinose prior to plasmid purification and confirmed by BsrGI and other restriction enzyme digests.
Right and left capture vectors were transformed into the conjugation strain RHO3. Using RHO3 donors, capture vectors were conjugally transferred and homologously recombined into the B. pseudomallei strains at the right and left flanks by biparental mating on LB+DAP plates followed by recovery of merodiploid exconjugants with zeocin, gentamicin, or kanamycin selection as appropriate. Due to the occasional difficulty in recovering Zeor exconjugants with AmrAB-OprA efflux pump-proficient strains, the Zeor-conferring right capture vectors were always recombined into the B. pseudomallei strains first to simplify screening. Isolated exconjugant right capture vector merodiploid colonies were screened by PCR with the primer pair 536+537 for the presence of an oriT. Confirmed TCAPR B. pseudomallei merodiploids were then used as recipients in a second biparental mating to homologously recombine left capture vectors. Recovered Kmr exconjugant TCAPR TCAPL double merodiploids were then pooled and saved as oriTn7 capture donors.
One ml each of LB+Km- or LB+Km+Zeo-grown cultures of double merodiploid B. pseudomallei donor strains, LB+Rf- or LB+Rf+Tmp-grown cultures of E. coli EPI300R1, and LB+Km- or LB+Km+Gm-grown cultures of DH5α/pUCPRK2013 or DH5α/pBBRK2013 helper strains was harvested by centrifugation and washed twice with 1 ml fresh LB medium prior to suspension in 100 to 200 μl LB. Portions of 20 μl each of donor, recipient, and helper strains were combined in a multiparental mating on a filter placed on an LB plate along with individual parental controls and incubated overnight at 37°C. E. coli exconjugants were recovered by selection on LB+Rf+Tmp+Km+Zeo plates. Candidate pFTarget-Tn7 clones were streaked for single-colony isolation on LB+Zeo+X-Gluc plates, and blue colonies were screened for Kmr, for Gms, and often for PmBs.
pFTarget-Tn7 vectors were confirmed by multiplex PCR, Eckhardt gel electrophoresis, and I-SceI restriction digestion. E. coli pFTarget-Tn7 candidate clones were inoculated into LB+Km or LB+Zeo medium with and without arabinose and grown overnight. Plasmid DNA was isolated and used as a template for multiplex PCR to confirm pFTarget-Tn7 insert identity. The multiplex primer set consisting of primers 2506 through 2519 was used to screen for the presence of the AMR, MBA, and 140 regions. Multiplex primers 2502 through 2505 were used to screen for the B. pseudomallei YLF and BTFC regions (24). Purified pFTarget-Tn7 plasmid DNA was analyzed by I-SceI restriction digestion to release the captured Tn7 target. The Eckhardt gel electrophoresis protocol as modified by the Griffitts lab (25) was used to resolve and visualize large pFTarget-Tn7 vectors. Pseudomonas syringae pv. phaseolicola 1448A 131.9-kb and 51.7 kb plasmids were used for size comparison (26). Briefly, overnight 30°C LB cultures of P. syringae pv. phaseolicola were subinoculated 2:7 in fresh LB and incubated for 4 h at 30°C with aeration to an optical density at 600 nm (OD600) between 0.5 and 0.7. Overnight 37°C LB Km E. coli pFTarget-Tn7 cultures were subinoculated 1:10 in fresh LB+Km medium and incubated for 5 h at 37°C with aeration to an OD600 between 0.5 and 0.7. Cell samples were chilled on ice, washed with 0.3% Sarkosyl, and lysed in the wells. The Eckhardt gel (1× SBE [10 mM NaOH, 1 mM EDTA, 29 mM boric acid, pH 8.0], 0.9% agarose, 0.5% SDS) was run in 1× SBE buffer at 100 V for 2.5 h total with 30 min current-on periods and 15-min current-off rest periods to prevent melting/warping of the gel. The gel was stained with SYBR Safe DNA gel stain from Invitrogen Life Technologies according to the manufacturer's recommendations.
To perform pFTarget-Tn7 transposition, pFTarget-Tn7 vectors were transferred into derivatives of the RHO3 conjugation strain, which can be metabolically counterselected by plating on LB medium lacking DAP. For pFTarget-Tn7 vectors with inserts of less than 15 kb, copy control-induced plasmid DNA was electroporated into RHO5, and transformants were recovered on LB+Km+DAP. To avoid deletion of insert sequences that routinely occurred when pFTarget-Tn7 vectors were retransformed with the 140-kb insert into RHO3, these plasmids were transferred into RHO3cm via conjugation. EPI300R1 pFTarget-Tn7 was used as the donor, DH5α/pBBRK2013 as the conjugation helper, and RHO3cm/pBBRSac3 as the recipient. The pBBRSac3 vector acts as an exclusion plasmid via plasmid incompatibility to prevent RHO3cm exconjugants from receiving the pBBRK2013 conjugation helper plasmid. Exconjugants were recovered on LB+DAP+Km+Zeo+Cm+Amp plates and then streaked for single-colony isolation on LB+DAP+Km+Zeo+Cm+10% sucrose to cure pBBRSac3. Recovered clones were screened for Kmr, Amps, Gms, and DAP auxotrophy. Plasmid DNA was isolated from RHO3cm/pFTarget-Tn7 exconjugants and reconfirmed by multiplex PCR. RHO5 or RHO3cm/pFTarget-Tn7 donors and RHO3/pTNS3 transposition helpers were conjugated with B. pseudomallei recipients and recovered on LB+Zeo or LB+Gm plates as appropriate. pFTarget-Tn7 transposition candidates were screened for Kms to eliminate single crossover events carrying the pFTarget-Tn7 backbone. Candidate Δ(amrRAB-oprA)-complemented strains were additionally screened for Gmr. B. pseudomallei pFTarget-Tn7 integrant candidates were checked for Tn7 insertion at attTn7-1, -2, or -3 using primer 479 in combination with primer 1509, 1510, or 1511, respectively (23). B. thailandensis candidates were checked for Tn7 insertion at glmS1 or glmS2 with primer 479 and primer 618 or 619, respectively (27). The appropriate multiplex PCR described above was used to confirm the pFTarget-Tn7 insert. Siderophore production phenotypes of Tn7 transposition candidates and control strains were determined by a quantitative chrome azurol S assay as described previously (28) from overnight cultures grown at 37°C in CAA medium (29).
B. thailandensis strain BT36 was electroporated with pARAtrfA, and transformants were recovered on LB+Zeo plates at 30°C. BT36 pARAtrfA and BT36 plasmid-negative control recipients were combined with the RHO5/pFTarget2-Tn7-YLF donor in biparental matings on an LB plate augmented with DAP and ARA and incubated overnight at 30°C, and Kmr exconjugants were recovered. BT36 pARAtrfA pFTarget2-Tn7-YLF maintenance was done at 30°C with ARA augmentation. To cure pARAtrfA, BT36 pARAtrfA pFTarget2-Tn7-YLF exconjugants were streaked for single-colony isolation on LB in the absence of selection at 42°C. Plasmid DNA was isolated from BT36/pARAtrfA/pFTarget2-Tn7-YLF overnight cultures by boiling for PCR analysis and plasmid kit isolation for retransformation into E. coli EPI300R1.
To improve upon the oriT-directed capture system described by Chain et al., we constructed new left and right capture vectors and broad-host-range conjugation helpers, which facilitate the cloning, manipulation, and complementation of large specific genomic regions. The capture vectors pFTCAPL-GW and pUCTCAPR-GW were built to comply with B. pseudomallei select agent regulations, which restrict antibiotic marker use to ble (Zeor), accCI (Gmr), and nptI and nptII (Kmr), However, they were also designed to allow easy swapping of the Kmr uidA and Flp recombinase target (FRT) Zeor markers in the pFTCAPL-GW and pUCTCAPR-GW vectors by EcoRI and XbaI digestion, respectively. Rather than use the ColE1-based pRK2013 conjugation helper plasmid to establish non-E. coli strains as conjugation donors, we created two new broad-host-range conjugation helper plasmids, pUCPRK2013 and pBBRK2013, by replacing the ColEI replicon of pRK2013 with either the pRO1600 replicon-containing vector pUCP24 or the Bordetella BBR replicon-containing vector pBBR1MCS-5. These vectors both used the accCI and nptI resistance markers, which are compliant with specific regulations regarding B. pseudomallei.
The steps and features of the oriTn7 system are summarized in Fig. 1. In brief, segments of DNA flanking the target region to the left and right are cloned into pFTCAPL-GW and pUCTCAPR-GW via Gateway LR recombination. The capture vectors are then homologously recombined with their matching genomic regions, creating a merodiploid strain in which the target region is flanked by oriT sites (Fig. 1A). By using either the pUCPRK2013 or pBBRK2013 conjugation helper plasmid, the merodiploid strain is then used as a conjugation donor so that the region between the oriT sites will be conjugated into an E. coli recipient strain (Fig. 1B). This results in the creation of a chimeric vector in which the target region is carried as cargo in a mini-Tn7 transposon and maintained by a mini-F replicon (Fig. 1C). The target region may then be inserted in a single copy into the attTn7 site via Tn7 site-specific transposition for complementation or heterologous expression (Fig. 1D).
To test the efficacy of the oriTn7 capture system, two B. pseudomallei genomic regions of difference that cover a range of target sizes were selected for cloning.
The first region is an ~140-kb region encompassing the genes encoding the aminoglycoside and macrolide antibiotic AmrAB-OprA efflux pump and the malleobactin siderophore synthesis and uptake cluster, which is absent from the clinical isolate 708a. Absence of the ~140-kb region has a minimal effect on virulence, as strains lacking this region remain fully lethal in a murine melioidosis model (30). For analysis of the ~140-kb deletion region, three targets corresponding to different functional portions of the region were captured from B. pseudomallei 1710b (Fig. 2). Although DNA sequence analysis using next-generation sequencing would have been an option for assessing the integrity of the large cloned DNA fragments, we opted for more traditional technologies, such as PCR and phenotypic analyses, to assess integrity and functionality of cloned genes and operons.
A set of seven multiplex PCR probes, one probe approximately every 20 kb along the 140-kb region length, was used to confirm the presence of the captured targets (Fig. 2A). Eckhardt gel electrophoresis was used to resolve the large circular pFTarget-Tn7 plasmids to confirm their sizes and to confirm that a single plasmid was present in the capture recipients (Fig. 2B). Multiplex PCR using plasmid DNA templates confirmed the presence of the appropriate targets (Fig. 2C to toE).E). The 140 target is 142.8 kb in size and captures the entire region absent from 708a. All seven multiplex probe sites are present in this target. The MBA region is 64.9 kb in size and captures the amrAB-oprA efflux pump operon and its associated regulatory gene amrR, as well as the malleobactin siderophore synthesis and uptake gene clusters. Multiplex probes 1 through 4 are located within this region. The AMR target is 12.1 kb in size and captures only the amrAB-oprA efflux pump operon and its amrR regulatory gene. Only multiplex probe site 1 is carried on this target. For each target region, the right capture flank was varied while the left capture flank was maintained. As it is occasionally difficult to cleanly recover Zeor exconjugants in AmrAB-OprA efflux pump-expressing strains, the Zeor-conferring right capture vector was always recombined into the chromosome first so that the merodiploid could be confirmed by checking for the presence of the oriT site by PCR. For 108 recipient cells, the number of recovered exconjugants from capture matings generally decreased with the size of the target from 102 to 103 for the AMR target to 101 for the 140 target.
The capacity of the system for complementation was tested in Bp338, a B. pseudomallei 1710b strain with a laboratory-generated deletion that mimics the ~140-kb natural deletion in B. pseudomallei 708a. We were able to successfully recover Tn7 integrants of all three pFTarget-Tn7 regions into Bp338 with frequencies ranging from 102 to 103 exconjugants per ~108 recipient cells. Kms Tn7 integrants were identifiable from the background of Kmr single-crossover exconjugants at approximately 1:10 for the MBA and 140 target regions and 1:3 for the AMR target region. The integrants were checked for attTn7 insertions by PCR, and target presence was confirmed by multiplex PCR (Fig. 2C to toE).E). Similar attempts were made to recover Tn7 integrants in B. pseudomallei 708a, but Tn7 integrants were successfully recovered only with the AMR target region. For the MBA and 140 target regions, only Kmr capture flank region crossover exconjugants of B. pseudomallei 708a were recovered (Fig. 2A). Other than acknowledging that strain 1710b is easier to genetically manipulate than other B. pseudomallei strains, we do not understand the reason(s) for our inability to recover site-specific mini-Tn7 integrants carrying the MBA and 140 regions. pFTarget-Tn7 integrants were also phenotypically confirmed. Resistance to Gm was restored in Tn7 integrants with any of the three target regions, but siderophore production was restored only in Tn7 integrants carrying either the MBA or 140 target (Fig. 2F). These experiments confirmed the malleobactin siderophore phenotypes previously associated with the 140-kb region deletion either naturally present in strain 708a or genetically engineered in the strain 1710b background (31). More importantly, the newly developed large fragment capture technology allowed us to do what we could not experimentally achieve previously, that is, to successfully complement the naturally present or engineered large genetic lesions causing the observed malleobactin siderophore deficiency and AmrAB-OprA efflux pump susceptibility phenotypes (31).
The second test region of 11 kb is the Yersinia-like fimbrial (YLF) gene cluster, which is one of a pair of B. pseudomallei gene clusters whose presence is geographically biased (24). The YLF and B. thailandensis-like flagellum and chemotaxis (BTFC) gene clusters are found at the same chromosomal location in different strains and appear to be mutually exclusive in nature (24). The YLF cluster is more common in Thai B. pseudomallei isolates, while Australian B. pseudomallei strains more commonly carry the BTFC cluster (24). B. thailandensis strains carry the BTFC cluster and provided the cluster's namesake. Phenotypes associated with the YLF cluster are unknown, but the BTFC cluster includes the gene motA2, which has been shown to be essential for intracellular motility and intercellular spread by B. thailandensis in the absence of the BimA actin recruitment-based motility system (32). Both YLF- and BTFC-carrying strains cause human disease and are virulent in animal models. Thus, neither of these regions carries genes that would be expected to enhance virulence of a strain from which either the YLF or BTFC cluster is absent (24, 33).
As we had already created B. thailandensis and Australian B. pseudomallei strain derivatives with deletions of their native BTFC gene clusters, we sought to capture the 11.4-kb YLF gene cluster from the Thai B. pseudomallei isolate 1026b and use Tn7 integration to create sets of strains carrying various combinations of the two gene clusters. The YLF target insert was sufficiently small that the I-SceI digest fragments of pFTarget2-Tn7-YLF were resolved and confirmed by standard gel electrophoresis (Fig. 3A). pFTarget-Tn7 integration was conducted and confirmed by multiplex PCR in the Australian B. pseudomallei isolate MSHR305 and its ΔBTFC derivative as well as by PCR in B. thailandensis E264 and its ΔBTFC derivative (Fig. 3B and andC)C) (32). Thus, we have generated sets of isogenic B. pseudomallei and B. thailandensis strains carrying neither gene cluster, their native BTFC cluster, their nonnative YLF cluster, or both gene clusters in combination. The ability to clone the YLF and BTFC gene clusters into defined B. pseudomallei genetic backgrounds will facilitate future characterization of the roles that the respective clusters play in this bacterium's biology.
As an alternative to Tn7 integration, we sought to determine if the pFTarget-Tn7 vectors could be maintained as broad-host-range replicating plasmids by utilizing the RK2 replication origin on the ccFOS vector backbone. Plasmid replication from the RK2 origin requires expression of the TrfA protein, which can be provided in trans. To this end, we created pARAtrfA, a temperature-sensitive Burkholderia PBAD-trfA expression vector, as a RK2 replication helper. We chose this strategy to aid in our analyses, as it would result in temperature-sensitive replication of pFTarget-Tn7 (Fig. 4A). For stable pFTarget-Tn7 plasmid maintenance, the use of a non-temperature-sensitive constitutive TrfA expression replication helper would be preferable. B. thailandensis BT36 was transformed with pARAtrfA and used as a conjugation recipient with a RHO5/pFTarget2-Tn7-YLF donor. Exconjugants were recovered at 30°C on LB+Km+ARA plates to maintain the plasmids. Candidates were purified at 42°C to cure both plasmids. All 12 single colonies tested were sensitive to both Zeo and Km. Control conjugations with BT36 lacking pARAtrfA resulted in the recovery of a small number of presumed crossover exconjugants whose Zeo and Km resistance could not be cured at 42°C. PCR confirmed both the presence of the YLF region in BT36/pFTarget2-Tn7-YLF plasmid-replicating exconjugants and the loss of YLF in temperature-cured strains (Fig. 4B). Plasmid DNA purified from pFTarget2-Tn7-YLF-replicating BT36 was transformed into E. coli EPI300R1. Comparison between I-SceI digests of plasmid DNA from the original pFTarget2-Tn7-YLF and recovered E. coli clones indicated that pFTarget-Tn7-YLF was recovered intact from BT36 (Fig. 4C).
Based on our initial success, we sought to determine if pARAtrfA would allow large-insert pFTarget-Tn7 vectors to be maintained as replicating plasmids in B. pseudomallei 708a or if pARAtrfA might enhance Tn7 integration of large target inserts into 708a by allowing even transient pFTarget-Tn7 replication. Unfortunately, while pARAtrfA enhanced the recovery of Kmr 708a flank region crossover exconjugants, it did not allow the temperature-sensitive replication of large target MBA or 140-kb insert plasmids or the recovery of large-insert Tn7 integrants.
The oriTn7 capture system offers some key advantages for the specific cloning and downstream manipulation of genomic fragments over other large-fragment cloning systems. The oriTn7 capture system can easily accommodate fragment sizes from 10 to over 100 kb and is mediated by reliable bacterial conjugation. All manipulation of large DNA is conducted in vivo, which eliminates the issue of DNA fragmentation. Finally, the captured target region is functionalized with Gateway attB sites and I-SceI recognition sites for transfer into other vectors and is also functionalized as a mini-Tn7 for single-copy chromosomal integration or maintenance as a multicopy plasmid by using the RK2 origin of replication.
Details of specific methodology considerations for the steps of the oriTn7 capture system are discussed below. First, to create target region-specific capture vectors, left and right capture regions flanking the desired target region are PCR amplified and TOPO cloned into directional pENTR vectors. Left and right flanking region primers are designed to amplify flanking regions of sufficient size to facilitate single homologous crossover events. In addition, flanking regions corresponding to central fragments of genes may be undesirable, as they will create corresponding gene interruptions after homologous recombination. The forward primers for each flanking region have the 5′ CACC directional TOPO sequence appended such that the correct flanking region orientations are maintained with respect to the targets genomic orientation throughout the procedure. The left and right flanking regions carried in the pENTR vectors are then LR recombined with the left (pFTCAPL-GW) and right (pUCTCAPR2-GW or pUCTCAPR3-GW) capture vectors (Fig. 1A).
The left and right capture vectors are sequentially recombined into the target genome via homologous recombination, resulting in a double merodiploid donor strain with two directly repeated oriT sites flanking the target region (Fig. 1B). The mini-F and pUC plasmid replicons in the left and right capture vectors are incapable of replication in Burkholderia spp. and many other bacteria. This allows antibiotic selection to identify RecA-mediated homologous recombination events. Exconjugant double merodiploid recombinants with both capture vectors recombined into the chromosome are pooled for use in subsequent capture matings.
Using one of the newly constructed broad-host-range conjugation helpers, pUCPRK2013 or pBBRK2013, conjugation is initiated from one integrated oriT site of the double merodiploid donor strain, and a single strand is transferred into the E. coli recipient strain. Conjugation continues until the second integrated oriT is reached. This halts the conjugation process, resulting in the creation of a chimeric circularized, Kmr Zeor, and Gms pFTarget-Tn7 vector carrying the desired target region (Fig. 1C). Although both broad-host-range conjugation helpers has proved effective for capture matings, in general, the use of pBBRK2013 results in 2-fold-greater numbers of recovered exconjugants from B. pseudomallei. Conjugation initiated from the second oriT is nonproductive, as it would require conjugation of the entire chromosome.
Recovered pFTarget-Tn7 candidate exconjugants are purified and screened for Gms. This screen checks for donor breakthrough, recovery of both the evicted capture vectors in combination, and the recovery of the pFTarget-Tn7 vector in combination with a conjugation helper plasmid. This last option was the most common cause of Gmr exconjugants. In general, Gms exconjugant clones appear to be recovered at a higher percentage as the size of the target insert increases. For example, for the 140-kb fragment, ~75% of the exconjugants were Gms, whereas for the 12.1-kb AMR fragment, ~25% were Gms. However, if no Gms clones are recovered from a capture mating, small insert pFTarget-Tn7 vectors (15 kb or less) can be isolated and retransformed by electroporation. Additionally, for large-insert pFTarget-Tn7 vectors, conjugation from an EPI300R1/pFTarget-Tn7 pBBRK2013 donor into a RHO3cm/pBBRsac3 BBR exclusion recipient can be used to recover clones lacking the pBBRK2013 helper plasmid.
In the final step of constructing a Tn7 integrant expression or complementation strain, the pFTarget-Tn7 vector is mobilized by conjugation into the desired recipient strain in combination with a Tn7 transposase-encoding helper plasmid. The expression of the site-specific Tn7 transposition pathway proteins is necessary to integrate the target Tn7 at glmS-associated attTn7 sites (Fig. 1D) (27). Putative Tn7 integrants are screened for single-crossover events by checking for Kms. For inserts under 15 kb, Tn7 integrants are easily recovered, presumably because the rate of Tn7 transposition for inserts of this size is higher than the rate of RecA-mediated recombination with the genome. To successfully recover attTn7-located integrants for large inserts, RecA-mediated crossover between the target insert and the genome must be restricted. This can be accomplished by limiting the homology between the target insert and genome by using a strain carrying a matching deletion of the target insert, by using a strain with limited genome homology to the target, or by integrating into a recA-deficient strain.
Although the left and right capture vectors by necessity share homology at their oriT sites, efforts were made to keep intervector homology to a minimum to prevent RecA-mediated homologous recombination between the two capture vectors during the creation of double merodiploid donor strains. Cloning larger target-flanking regions will help bias against inter-oriT homologous recombination, and it would also typically be favorable to homologously recombine the capture vector carrying the smaller of the two flanking regions first rather than second. However, a capture mating from a donor with inter-oriT recombined capture vectors would only result in the recreation of the original misrecombined capture vector after conjugation into the E. coli recipient. As the original capture vectors do not individually carry the required combination of antibiotic markers to be recovered, pooling the double merodiploid exconjugants is a desirable strategy for this step of the system.
The final captured target regions carried in pFTarget-Tn7 vectors are usefully functionalized in various ways. The pFTarget-Tn7 vectors carry both a mini-F-based replicon, which can easily accommodate inserts of over 100 kb, and an RK2 oriV, which allows broad-host-range replication of pFTarget-Tn7 vectors if TrfA is provided in trans or induction to high copy number in CopyControl strains that express a mutant, copy-up variant of the TrfA protein (34). The single oriT site on the pFTarget-Tn7 vectors both is intrinsic to the system and facilitates conjugation-based mobilization of large pFTarget-Tn7 vectors between strains. The incorporation of Gateway cassettes into the left and right capture vectors in combination with directional pENTR/D-TOPO vectors simplifies work flow and primer design but also results in the captured target region being flanked with Gateway attB sites. The attB sites can be used to recombine the target region into attP carrying Gateway pDONR vectors, assuming that the particular pDONR vector is capable of supporting replication of the target insert. The target Tn7 is flanked by 18-bp, rare cutting I-SceI endonuclease recognition sites, which both aid in restriction-based screening of candidates and allow I-SceI-based restriction subcloning of targets inserts. Lastly, the pFTarget-Tn7 element carries a Flp recombinase-excisable FRT-Zeor-FRT cassette. This allows the target insert to be site-specifically transposed in a single copy into the chromosomes of most Gram-negative bacteria, followed by Flp-mediated excision of the Zeor selection marker. Some bacteria, for instance, B. thailandensis and B. pseudomallei, contain more than one Tn7 insertion site, but simultaneous insertion into multiple sites occurs less frequently than single-site insertion, and single and multiple insertions can easily be differentiated by PCR.
In addition, if both capture flanking regions have homology to the borders of a matching chromosomal deletion, the pFTarget-Tn7 vector may also be used as an allelic exchange vector for deletion restoration. The I-SceI recognition sites in the vector backbone can serve as counterselection markers in combination with an I-SceI endonuclease expression plasmid (35). Colorimetric screening for the uidA gene would aid the identification of double-crossover events resolving the merodiploid state. Unfortunately, this strategy cannot be used in B. pseudomallei strains expressing AmrAB-OprA, as all three currently permissible antibiotic resistance markers—Gmr, Kmr, and Zeor—are already in use, leaving none to select for introduction of an I-SceI expression plasmid. However, one could envision the use of other permissible non-antibiotic resistance markers, such as glyphosate resistance, for maintenance of the I-SceI expression vector (36).
Currently, the oriTn7 capture system is limited to use in genetically tractable Gram-negative bacteria that can be established, at least transiently, as conjugation donors. The F replicon and the pUC replicon must also be nonreplicative in the target organism. Even with these limitations, the genera Burkholderia, Pseudomonas, Ralstonia, and Sinorhizobium should all be compatible, and likely many other genera as well.
Funding was provided by indirect cost recovery returns made possible by several extramural National Institute of Allergy and Infectious Diseases (National Institutes of Health) research grants and the Ruth L. Kirschstein National Research Service Award F32 (AI088884) to B.H.K.
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
We thank Bryan Swingle for providing pBBR1-MCS5, Alan Collmer for providing pCPP5250, Carolina Lopez for providing pEXpheS*Gen, and Lily Trunck for providing B. thailandensis BT36.
Published ahead of print 7 June 2013
Supplemental material for this article may be found at http://dx.doi.org/10.1128/AEM.00994-13.