|Home | About | Journals | Submit | Contact Us | Français|
An 8,883-bp mini-pXO1 plasmid containing a replicon from Bacillus anthracis pXO1 (181.6 kb) was identified by making large deletions in the original plasmid using a newly developed Cre-loxP system. Portions of the truncated mini-pXO1 were cloned into an Escherichia coli vector unable to replicate in B. anthracis. A 5.95-kb region encompassing three putative genes was identified as the minimal pXO1 fragment required for replication of the resulting recombinant shuttle plasmid (named pMR) in B. anthracis. Deletion analysis showed that the only genes essential for replication were the pXO1-14 and pXO1-16 genes, which are transcribed in opposite directions and encode predicted proteins of 66.5 and 67.1 kDa, respectively. The ORF14 protein contains a helix-turn-helix motif, while the ORF16 upstream region contains attributes of a theta-replicating plasmid origin of replication (Ori), namely, an exclusively A+T-containing segment, five 9-bp direct repeats, an inverted repeat, and a σA-dependent promoter for the putative replication initiator Rep protein (ORF16). Spontaneous mutations generated in the ORF14, ORF16, and Ori regions of pMR during PCR amplification produced a temperature-sensitive plasmid that is unable to replicate in B. anthracis at 37°C. The efficacy of transformation of plasmid-free B. anthracis Ames and Sterne strains by the original pMR was ~103 CFU/μg, while Bacillus cereus strains 569 and ATCC 10987 were transformed with efficiencies of 104 and 102 CFU/μg, respectively. Around 95% of B. anthracis cells retained pMR after one round of sporulation and germination.
Bacillus anthracis, the etiological agent of anthrax, is a gram-positive, rod-shaped, spore-forming bacterium. A key feature distinguishing B. anthracis from its nearest neighbors is the presence of two large plasmids that encode key virulence factors. Plasmid pXO2 encodes enzymes that synthesize the poly-γ-d-glutamic acid capsule, whereas pXO1 encodes the anthrax toxins, consisting of protective antigen, lethal factor, and edema factor (10, 15). The complete sequence of pXO1 was first determined by Okinaka et al. (12). However, the sequence was not found to contain regions having homology to any of the rep genes associated with origins of replication in large plasmids of gram-positive organisms. Subsequently, two groups sought to characterize the structure of the pXO1 replicon. Robertson et al. (19) cloned pXO1 DNA fragments into an E. coli plasmid and obtained shuttle plasmids that could then replicate in Bacillus species. This strategy produced several clones containing an 11-kb region between coordinates 86249 and 97209 of the pXO1 sequence (12). This region contains nine open reading frames (ORFs) (ORF72 to ORF80 in the annotation scheme used; GenBank accession no. NC_001496), but the predicted products were again not recognizable as Rep proteins associated with theta-replicating plasmids. (The numbering of nucleotides and ORFs as annotated by Okinaka et al.  will be used throughout this report.)
In a more recent effort to explain pXO1 replication, Tinsley and Khan cloned a 5-kb region from B. anthracis Sterne plasmid pXO1 into an E. coli vector and showed that this plasmid could replicate when introduced into B. anthracis (22). This fragment included ORF43 to ORF46 rather than those previously implicated in replication (19). Tinsley and Khan demonstrated that one particular coding sequence, renamed repX (ORF45), was required for the initiation of replication. They also identified a 158-bp region (bp 55726 to 55883) that appeared to be required for the initiation of replication. This putative origin is contained within the highly conserved core replication region identified in all pXO1-like plasmids of the Bacillus cereus group (17). In this region, B. anthracis Sterne pXO1 contains a perfect 24-bp palindrome (22). However, in the several other pXO1 plasmid sequences determined up to now, as well as in the B. cereus pXO1-like plasmids, the palindrome is imperfect, having a 2-bp mismatch. The effect, if any, of this difference on plasmid replication is unclear (17). Anand et al. (2) showed that the RepX protein can undergo GTP-dependent dynamic polymerization and found that RepX binds to DNA in a GTP-dependent manner, albeit weakly and nonspecifically. RepX shares homology at its amino terminus with the bacterial cell division protein FtsZ. Both the FtsZ and RepX proteins belong to the tubulin family, members of which act as cytoskeletal proteins involved in cell division and DNA segregation in prokaryotic and eukaryotic cells (2). Later the same group showed that RepX forms polymers in vivo in B. anthracis (1). On the basis of these data they speculated that RepX may be a hybrid protein capable of participating in both plasmid replication and segregation. It was also suggested that the megaplasmids found in members of the B. cereus group of organisms may have a conserved mechanism which utilizes members of the FtsZ/tubulin family of proteins for plasmid maintenance (1).
We recently described a method for mutating or deleting large DNA segments in B. anthracis (15). A plasmid containing two fragments homologous to the target sequence flanking a spectinomycin resistance cassette that is in turn flanked by bacteriophage P1 loxP sites is inserted by a single crossover event. After selection with spectinomycin and then identification of a subsequent double-crossover event, temporary expression of Cre recombinase excises the spectinomycin marker, leaving a single loxP site within the targeted gene or genomic segment. The procedure could then be repeated to inactivate other genes, but in each case a loxP sequence remained. An advantage of this method is that the 34-bp loxP sequence does not necessarily produce polar effects when introduced into a gene or operon. For example, insertion of the loxP-Ω-sp-loxP cassette into pXO2 between the promoter and the capB start codon prevented capBCAD operon expression and capsule synthesis because the Ω-sp cassette contains transcriptional terminators. However, Cre-mediated excision of the Ω-sp cassette restored capsule production even though a single loxP site remained between the promoter site and the capB start codon (15). In the course of that work, we found that large segments (of at least 30 kb) could be deleted from chromosomal or pXO2 DNA (15). Here we describe a modified Cre-loxP system that does not depend on a low-frequency double-crossover event for insertion of the spectinomycin resistance cassette into a selected region of the B. anthracis genome. Instead, in the new approach, we used two sequential single-crossover events to introduce loxP sites flanking the DNA region targeted for deletion. Cre recombinase action at the loxP sites excised the intervening DNA, leaving a single loxP site within the targeted gene or selected genomic segment. This strategy allows deletion of DNA fragments from any genome (or plasmid), thereby facilitating functional analysis of specific genes or genome regions. Using this approach, we identified and characterized a new pXO1 minimal replicon different from those reported previously (19, 22).
E. coli strains were grown in Luria-Bertani (LB) broth and used as hosts for cloning. LB agar was used for selection of transformants (20). B. anthracis as well as B. cereus strains were also grown in LB medium. Antibiotics (Sigma-Aldrich) were added to the medium when appropriate to give the following final concentrations: ampicillin (Ap), 100 μg/ml (only for E. coli); erythromycin (Em), 400 μg/ml for E. coli and 5 μg/ml for B. anthracis; and spectinomycin (Sp), 150 μg/ml for both E. coli and B. anthracis. SOC medium (Quality Biologicals) was used for outgrowth of transformation mixtures prior to plating on selective medium to isolate transformants. B. anthracis spores were prepared by growth on NBY-Mn agar (nutrient broth, 8 g/liter; yeast extract, 3 g/liter; MnSO4·H2O, 25 mg/liter; agar, 15 g/liter) at 30°C for 5 days (21). Spores and vegetative cells of B. anthracis were visualized with a Nikon Eclipse E600W light microscope.
Preparation of plasmid DNA from E. coli, transformation of E. coli, and recombinant DNA techniques were carried out by standard procedures (20). E. coli XL2-Blue and SCS110 competent cells were purchased from Stratagene and E. coli TOP10 competent cells from Invitrogen. Recombinant plasmid construction was carried out in E. coli XL2-Blue or TOP10. Plasmid DNA from B. anthracis was isolated according to the protocol for the purification of plasmid DNA from Bacillus subtilis (Qiagen). Chromosomal DNA from B. anthracis was isolated with the Wizard genomic purification kit (Promega) in accordance with the protocol for isolation of genomic DNA from gram-positive bacteria. B. anthracis was electroporated with unmethylated plasmid DNA isolated from E. coli SCS110 (Dam− Dcm−). Electroporation-competent B. anthracis cells were prepared and transformed as previously described (13) with slight modifications. B. anthracis cells grown in 25 ml of LB broth to an optical density at 600 nm of 0.15 to 0.20 were washed three times with 1 ml of ice-cold electroporation buffer (10% sucrose, 15% glycerol, 2 mM phosphate buffer, pH 8.4), resuspended in 200 μl electroporation buffer, mixed with 0.1 to 1.0 μg of plasmid in an ice-chilled electroporation cuvette (0.2 cm), and electroporated using the Bio-Rad Gene Pulser II (voltage, 1.77 kV; capacitance, 25 μF; resistance, 200 Ω). Bacteria were recovered from the cuvette with 1 ml of SOC medium, shaken at 225 rpm and 30 or 37°C for 1 h, and spread on plates with the appropriate antibiotic. Depending on the plasmid and the temperature of incubation used, the plates were kept for 1 to 2 days before further analysis. The same procedure was used for B. cereus electroporation except that the voltage used was 1.35 kV.
Restriction enzymes, T4 ligase, T4 DNA polymerase, and alkaline phosphatase were purchased from MBI Fermentas or New England Biolabs. Taq polymerase and Platinum PCR SuperMix High Fidelity kits were purchased from Invitrogen. The TOPO TA cloning kit (Invitrogen) and pGEM-T Easy Vector system (Promega) were used for PCR fragment cloning. Ready-To-Go PCR Beads (Amersham Biosciences) were used for DNA rearrangement analysis. For routine PCR analysis, a single colony was suspended in 200 μl of TE buffer (20) (pH 8.0), heated to 95°C for 45 s, and then cooled to room temperature. Cellular debris was removed by centrifugation at 15,000 × g for 10 min. Two microliters of the lysate contained sufficient template to support a PCR with the PCR beads. The GeneRuler DNA Ladder Mix (MBI Fermentas) was used for determination of DNA fragment length. All constructs were verified by DNA sequencing and/or restriction enzyme digestion. The oligonucleotides, plasmids, and strains used in this study are listed in Tables Tables11 to to3,3, respectively.
The general scheme for producing B. anthracis pXO1 deletions using the new Cre-loxP system, presented in Fig. Fig.1,1, employs plasmids we designate generically as pSC, for single-crossover plasmid. These were derived from the highly temperature-sensitive plasmid pHY304, which has permissive and restrictive temperatures of 30°C and 37°C, respectively. To produce pSC we cut pHY304 with Ecl136II/BstZ17I and ligated this with a BstZ17I-restricted fragment from a 181-bp DNA sequence (direct loxP repeat [DLR]) synthesized by BlueHeron Biotechnology, WA. The DLR sequence contains a multiple-cloning site between two directly repeated loxP sequences. The resulting pDR1 plasmid was cut with BstZ17I and ligated with a pUC18 PvuII fragment, producing the shuttle plasmid pSC (Fig. (Fig.1,1, step I). All the endonuclease restriction sites indicated on the pSC map between two direct loxP sequences are unique restriction sites. The plasmid pCrePA, previously used for expression of Cre recombinase, was modified by replacing the Emr gene with the Spr gene from pJRS312, to create pCrePAS. For this purpose, pCrePA was cut with ScaI/FspI and ligated with the FspI/MslI fragment of pJRS312. The resulting pCrePAS plasmid, shown in Fig. Fig.1,1, step II, has permissive and restrictive temperatures of 30°C and 37°C, respectively. Both pSC and pCrePAS were used in accordance with the schemes presented in Fig. Fig.11 and Fig. Fig.22.
Fragments of 5,950 and 5,110 bp were amplified from plasmid Δ2 using Platinum PCR SuperMix High Fidelity kit (Invitrogen) and the MRF/MRR and MRF1/MRR1 primer pairs, respectively. The PCR products had 3′ adenines added in accordance with the TOPO TA cloning protocol (Invitrogen) and were cloned into pCR2.1-TOPO, producing the TOPO-MR and TOPO-MR1 plasmids. These were cut with EcoRI and ligated into the pΩL plasmid at its EcoRI site, producing plasmids pMR and pMR1. Plasmid pMR was obtained as variants having both orientations of the EcoRI fragment containing pXO1 ORF14 to ORF16. Both variants were used for deletional analysis (see Fig. Fig.4).4). From the variants having the first orientation, the NruI/PvuI fragment extending into ORF14 was deleted, producing pMRΔ14. The other variant had the Ecl136II/PvuI fragment that included all sequences beyond ORF16 deleted, producing pMR16Δ. Deletion of the C-terminal region of ORF16 from the Ecl136II to the PvuII site produced pMRΔ16. To produce pMRΔ15, we cut pMR with XcmI and treated it with T4 DNA polymerase according to the New England Biolabs protocol for blunting ends by 3′ overhang removal. However, the pMRΔ15 obtained had an unexpected larger deletion of ORF15 and flanking regions (as detailed in Table Table2).2). To produce pMRMut, we repeated the entire cloning scheme used for pMR with the exception that regular Taq polymerase was used instead of the Platinum PCR SuperMix High Fidelity polymerase. Point mutations introduced during this PCR amplification to produce pMRMut were subsequently identified by complete sequencing of the pXO1 fragment with the same primers used for Δ2 plasmid sequencing (see Fig. Fig.44).
To allow convenient assessment of the stability of pXO1, a spectinomycin resistance cassette was inserted downstream of the pag gene by a double-crossover event. A 2,265-bp region located immediately downstream of pag (protective antigen gene) was amplified with the SPF/SPR primers and cloned into pGEM-T Easy, producing pGEM1. The 2.3-kb BamHI fragment from pΩL containing the Ω-sp cassette was inserted into the unique BamHI site of the amplified B. anthracis sequence in pGEM1, producing pGEM1Ω. The EcoRI fragment containing the pXO1 fragment with the internal Ω-sp from pGEM1Ω was then inserted into the EcoRI site of pHY304, producing pHY304S. The Ω-sp from pHY304S was inserted into the homologous region downstream of pag by the double-crossover method previously described (15).
The stabilities of plasmids pXO1S, pMR, and pMRMut in B. anthracis Ames 35 (pXO1S) and Ames 33 (pMR and pMRMut) were determined using methods described previously (11), with slight modifications. Bacterial strains were inoculated into LB broth with spectinomycin and incubated at 37°C. The cultures then were diluted 106-fold into LB broth without antibiotic and incubated overnight at 37°C. Sixteen passages in nonselective LB medium were performed at daily intervals. At each passage, aliquots were plated on agar without spectinomycin and about 100 colonies were transferred to agar containing spectinomycin to determine the percentage of cells retaining the plasmid. To determine plasmid stability during sporulation, overnight cultures from LB medium with spectinomycin were spread on NBY-Mn agar without the antibiotic and incubated for 5 days at 30°C. Cultures were suspended in water and heated at 70°C for 30 min to kill the nonsporulated cells, and the percentage of plasmid-containing spores was then determined as described above.
Plasmid DNAs were sequenced using primers listed in Table Table1.1. All primers for sequencing were synthesized by Operon Biotechnologies, Inc., AL, or the FDA core facility, Bethesda, MD. Sequences were determined using a primer walking strategy (Macrogen). Sequence data were assembled using either the Clustal module of the GCG-Lite suite of sequence analysis program (http://molbio.info.nih.gov/gcglite) or Vector NTI software (Invitrogen). The National Center for Biotechnology Information (NCBI) BLAST and FASTA programs (http://www.ncbi.nlm.nih.gov) were used for homology searches in GenBank and nonredundant protein sequence databases. Repeat Finder software (http://www.proweb.org/proweb/Tools/selfblast.html) was used to locate direct and inverted repeat (IRs). Protein motifs were identified using SignalP software version 3.0 (http://www.cbs.dtu.dk/services/SignalP) for signal sequences, TMpred software for prediction of transmembrane regions and orientation, and NPS@ software (http://npsa-pbil.ibcp.fr/cgi-bin/primanal_hth.pl) for helix-turn-helix prediction.
Although two regions of pXO1 have been implicated in its replication, the mechanism(s) of their action is not understood. To identify regions that are either necessary or sufficient for pXO1 maintenance, we modified a Cre recombinase-based system (15) that is able to generate large deletions in pXO1 (Fig. (Fig.1).1). Unlike the prior method, this one does not require or involve low-frequency double-crossover events. The scheme is illustrated in Fig. Fig.1,1, where, for illustration of the method, the Δ1 region is defined as containing an independent replicon. The scheme begins by insertion of a single loxP site into the target region using two consequent steps: step I, insertion as a single crossover of a plasmid (pSCL1) having the (left) L1 region of homology bracketed by loxP sites, followed by, step II, excision of the pSC vector backbone from the target by means of Cre recombinase treatment along with elimination of the plasmid by growth at a restrictive temperature (Fig. (Fig.1).1). This leaves a single loxP site flanked by duplicate copies of the L1 sequence. The process continues with a second single-crossover event (step III) that inserts two additional loxP sites flanking pSC in a manner similar to step I. Cre recombinase treatment of that transformant excises sequences between the three loxP sites to produce several possible circular DNAs, most simply those designated Δ1 and pXO1Δ1 (Fig. (Fig.1).1). In the present case, where Δ1 is suggested to be the only replicon present in pXO1, the minireplicon plasmid Δ1 survives and the larger pXO1Δ1does not. Repetition of steps III and IV performed on the Δ1 plasmid using pSC containing fragments F2 andF3 could further reduce its size, until a Cre recombinase treatment removed an essential part of the replicon, in which case no plasmids would survive; this is the case illustrated in step VI (Fig. (Fig.11).
Results from application of this strategy to pXO1 are shown in Fig. Fig.2.2. At first, to validate the new Cre-loxP strategy, we eliminated the region described as a pathogenicity island (PAI). The PAI region is a 44.8-kb region defined by terminal inverted IS1627 elements and containing ORF97 to ORF126 (12). In addition to the toxin genes, the PAI contains a set of regulatory elements that control the expression of the toxin genes, a spore germination response operon, and 19 other ORFs, including several transposition-related elements (23). This PAI is inverted in some strains (21). To eliminate the PAI, we inserted a single loxP site into the PAI's counterclockwise (or left) end IS1627 element using an L1 fragment corresponding to bp 117241 to 117872. The insertion of the loxP sites at the clockwise end of the PAI using a fragment homologous to bp 160861 to 161664 was followed by Cre recombinase treatment to generate the D1 deletion, which was confirmed by PCR analysis (Fig. 2A and B, left panel). The independent (i.e., not sequential) D2 deletion used again the loxP site introduced at the counterclockwise end of the PAI and a fragment amplified with primers D2F/D2R, corresponding to bp 67668 to 68049. This deletion was intended in part to test the requirement for the replicon reported to be within bp 86249 to 97209 (R replicon in Fig. Fig.2A)2A) (19). The resulting pXO1 deletant lacking (only) the D2 region (as confirmed by PCR analysis) was retained (Fig. 2A and B, left panel). Deletions were then continued extending clockwise from the PAI. The pSC plasmid having a fragment of DNA homologous to the clockwise end of the PAI that was used above to delete the PAI was now used to insert a single loxP site. PCR with primers D3F/D3R generated a fragment homologous to bp 18001 to 18672 that was used to insert loxP sites which, after Cre treatment, defined the clockwise end of the D3 deletion (Fig. 2A and B, left panel). Again, deletion of (only) D3 did not prevent pXO1 replication.
None of the three large deletions described above prevented retention of the resulting pXO1 variant plasmids by B. anthracis cells, indicating that other regions contained one or more functional replicons. Taking into account the location of the putative TK replicon at bp 54863 to 60166 (22), we designed a fourth independent deletion (provisionally designated D4) that began with insertion of a single loxP site using a fragment homologous to bp 18001 to 18672 that was produced with primers D3F′/D3R′. This was followed by insertion of two loxP sites using a fragment homologous to bp 54639 to 55390 that was produced with primers D4F′/D4R′. Expression of Cre recombinase in bacteria having the resulting three loxP sites did not cause deletion and loss of the targeted (D4) region but instead led to its independent replication. This DNA fragment (bp 18001 to 55390) replicated independently, and the remaining larger portion of pXO1 encompassing bp 54639 to 181677 and continuing through bp 1 to 18672 bp was lost. This result defines a new minireplicon of pXO1 and also calls into question the essentiality of the putative TK replicon. In accordance with the scheme in Fig. Fig.1,1, we designated the replicative pXO1-fragment Δ1 (Fig. 2A and B, left panel).
To further define the replicon within Δ1, we inserted additional loxP sites at bp 25803 and 26801 bp (primers D4FC′/D4RC′ were used for the fragment amplification). The PCR analysis performed following Cre treatment revealed a shortened self-replicating part of pXO1 that we designated Δ2 (Fig. (Fig.2).2). The Δ2 plasmid contains ORF13 to ORF19 of pXO1 (Fig. (Fig.3,3, middle image). To confirm all of the deletions described above, we synthesized clockwise- and counterclockwise-facing primers (Table (Table1)1) to amplify fragments spanning the junction sites of the deletions. The fragments obtained were of the expected size, as seen in Fig. Fig.2B,2B, right panel. All these junction fragments were sequenced and found to have the predicted sequences flanking a single loxP site. The loxP site, primer locations, and sequence of the junction in the Δ2 plasmid are shown in Fig. Fig.2C2C.
A further attempt to truncate the 8883-bp Δ2 plasmid using the Cre-loxP system did not yield smaller replicative plasmids. Genetic analysis of Δ2 plasmid ORFs indicated that the final recombination event (Fig. (Fig.3,3, top two images, corresponding to step VI in Fig. Fig.1)1) separated ORF14 and ORF15 (in plasmid Δ3) from ORF16 to ORF19 (plasmid Δ2-3). This separation was fatal for both Δ3 and Δ2-3. This information and the manageable size of the replicon defined to this point allowed the use of alternative strategies to identify the regions essential for replication of the Δ2 plasmid.
Based on indications that ORF14 to ORF16 might constitute the replicon, the region defined by the MRF/MRR primer pair was amplified and cloned into pΩL via pCR 2.1-TOPO cloning according to the scheme presented in Fig. Fig.3,3, bottom. The resulting plasmid, pMR, equally transformed both B. anthracis Ames 33 and SdT strains. The efficacy of transformation was around 103 CFU per μg of the plasmid DNA. No transformants were expected or obtained when the parental pΩL plasmid was similarly electroporated into Ames 33 or SdT. These results showed that the 5,950-bp region at bp 19032 to 24981 contains a functional minireplicon of the pXO1 plasmid. A parallel cloning of the shorter region defined by primer pair MRF1/MRR1 was done, and no replicative plasmid was obtained, suggesting that the region between ORF16 and ORF17 is also essential (see also results for pMR16Δ, below). The pMR plasmid was isolated from B. anthracis, and the region corresponding to the MRF/MRR fragment was sequenced. The sequence was identical to that of the original pXO1 plasmid (GenBank accession no. NC_001496). The sequence was seen to contain two stem-loop structures (Fig. (Fig.4A)4A) that are potential transcription terminators for the pXO1-15 or pXO1-16 genes (see below).
To determine which of the ORFs and adjacent regions of the pMR plasmid are required, various truncations and deletions of the ORFs were constructed (Fig. (Fig.4A).4A). The ORF15 deletion (in pMRΔ15) also encompassed the left stem-loop structure (bp 21717 to 21741). The right stem-loop structure (bp 24889 to 24935) was deleted along with the region downstream of ORF16 in pMR16Δ (Fig. (Fig.4A).4A). Electroporation of the resulting plasmids into B. anthracis showed that both pMR and pMRΔ15 gave transformants at efficacies of ~ 103 CFU/μg, while pMR16Δ, pMRΔ14, and pMRΔ16 gave no transformants at either 30 or 37°C. We also found that the pMRMut plasmid had a drastically reduced ability to transform B. anthracis at 37°C, yielding small colonies at a 100-fold-lower frequency than for pMR. However, the ability of this plasmid to transform B. anthracis was restored at 30°C, although the colonies remained smaller than for pMR transformants (Fig. (Fig.5A).5A). Both the pMR and pMRMut plasmids were tested for their ability to replicate in B. cereus strains 569 and ATCC 10987. The plasmids transformed these strains in the same temperature-sensitive manner as found for B. anthracis but with efficiencies that were higher for B. cereus 569 (~104 CFU/μg) and lower for B. cereus ATCC 10987 (~102 CFU/μg). The presence of the plasmids in the B. cereus transformants was confirmed by PCR analysis and sequencing (data not shown).
Like a majority of large plasmids, pXO1 probably replicates by the theta-type mechanism (6). In many cases, replicons of plasmids using the theta-type mechanism of replication contain an origin of replication (Ori) and an adjacent gene that codes for a replication initiator protein (Rep). Replication origins often consist of multiple cis-acting elements, including direct repeats (iterons), AT-rich regions, and IRs (5). Analysis of the regions adjacent to ORF14 and ORF16 showed that only the latter has features characteristic of an origin of replication: a 24-bp element with 100% AT content preceding one slightly divergent and four perfect repeats of a core 9-bp iteron sequence (TTTCCCAAG) not found elsewhere in pXO1 and an IR located between the iterons and the putative promoter of the adjacent pXO1-16 gene (Fig. (Fig.4B).4B). It is notable that one of the mutations that singly or in combination make pMRMut temperature sensitive for replication is a single-nucleotide deletion in this region (Fig. (Fig.4A4A).
BLAST analyses revealed that both the ORF14 and ORF16 amino acid sequences are well conserved among B. cereus group plasmids (Table (Table4).4). The ORF14 protein has 39 to 99% identity with conserved hypothetical proteins in a number of these plasmids, while the ORF16 protein is more strongly conserved among the same plasmids, having at least 64% identity. It is interesting that the ORF14 protein contains a helix-turn-helix motif at amino acids 516 to 537, in the C-terminal region of the protein.
The ORF15 protein is much less conserved. It is 100% identical to a hypothetical protein of pBCXO1 (8) but less than 40% identical to proteins of pBc239 (24) and pG9842 from B. cereus G9842 (NC_011775). This protein is predicted to have a signal sequence of 32 amino acids, suggesting that it may be secreted or surface associated.
The segregational stabilities of the pXO1 derivative plasmids pXO1S, pMR, and pMRMut differed greatly when the host strains were grown in the absence of positive selection by spectinomycin (Fig. (Fig.5B).5B). The pXO1S plasmid, which is simply a spectinomycin-marked, full-size pXO1, was very stable, with less than 5% plasmid loss from the B. anthracis Ames 35S population after 16 passages at 37°C. The pMR plasmid, which contains only the pXO1 minireplicon, was found in 80% of the B. anthracis Ames 33(pMR) cells after the 5th passage but disappeared by the 16th passage at 37°C. Only two passages at 37°C were required to eliminate pMRMut from the population of B. anthracis Ames 33(pMRMut). We also found that the pMR plasmid was retained during sporulation, since 95% ± 5% of both B. anthracis Ames 33(pMR) and SdT(pMR) bacteria produced by germination of spores had retained the plasmid.
In the present study, we demonstrated that a 5.95-kb fragment spanning bp 19032 to 24981 in pXO1 (12) is able to replicate independently and therefore constitutes a minireplicon. This region contains genes potentially encoding three proteins, ORF14 to ORF16, two of which, ORF14 and ORF16, are required for replication. Our sequence analyses, together with those of Okinaka et al. (12) and Rasko et al. (17), identified canonical features of plasmid replication origins, including an AT-rich region, direct repeat and IR sequences, and an adjacent putative replication (Rep) protein, ORF16. This protein is very similar to a hypothetical protein encoded by a number of other plasmids from the B. cereus group. The level of amino acid identity for ORF16 is >64% for 10 different plasmids and >90% for 8 of them. One of these proteins, pBt007 (from pBtoxis) has 96% identity with pXO1-16 and was recognized as widespread in B. thuringiensis isolates containing different virulence plasmids (3). Although the pXO1-16 protein shares no homology with Rep proteins of other large plasmids, the structural features of its putative Ori, which includes iteron sequences, an AT-rich region, and an IR, suggest that this protein could induce replication of pXO1 using the theta mode of replication.
However, the inability of ORF16 (alone) to maintain the Δ2-3 plasmid in B. anthracis cells (Fig. (Fig.3,3, upper panel) indicates that ORF14 is also essential for maintenance of the pMR minireplicon. Moreover, both proteins are functional only when the corresponding genes are located within the same plasmid, as in plasmid Δ2. Separation of these two genes to different plasmids, Δ3 and Δ2-3, resulted in the loss of both the plasmids. Although we showed that both ORF14 and ORF16 are absolutely necessary for pXO1 minireplicon maintenance, the precise roles of the proteins remain unknown. Further studies will be needed to fully understand the roles of both ORF14 and ORF16 in initiation of pXO1 replication and its maintenance in B. anthracis cells.
The ORF15 protein was not found to be essential for replication of pMR and therefore is probably dispensable for pXO1 replication. Proteins having some homology to ORF15 were found only in pBCXO1 (8) and (with lower homology) in pBc239 (24) and pG9842 (NC_011775). It is not evident from an evolutionary point of view how and for what purpose a gene encoding a putative secreted protein would be inserted between two genes encoding intracellular proteins essential for pXO1 replication.
Decreased segregational stability of pMR in comparison with pXO1S was observed in B. anthracis during repeated passage at 37°C. This could result from the absence of a functional partition system within pMR. Partition is an active process needed by low-copy-number plasmids to ensure accurate distribution of plasmids to both daughter cells. Partition modules typically contain three components: a cis-acting site consisting of a set of sequence repeats (parS) and two genes, one coding for an ATPase (ParA) and the other for a parS-binding protein (ParB) (4). No components of a partition system have been identified on the pXO1 genome, and we could not identify motifs characteristic of parA genes in pMR. This suggests that RepX (ORF45) and its DNA target may participate in pXO1 partitioning, as an alternative partition module (1, 2, 22).
We found that the pXO1 minireplicon can replicate in B. cereus strains 569 and ATCC 10987. Both strains contain other plasmids: three in B. cereus 569 (14, 18) and one in B. cereus ATCC 10987 (16). The ability of pMR to transform these B. cereus strains regardless of their endogenous plasmid content suggests that plasmids derived from pMR may have value in genetic manipulation of many members of the B. cereus cluster. The differences in transformation efficiencies observed between the two B. cereus strains could be due to differences in restriction/modification systems or possibly could result from plasmid incompatibilities. These issues merit further study.
The inability of pMRMut to replicate efficiently at 37°C indicates that this mutant plasmid contains a temperature-sensitive replicon. The four mutations in this plasmid include a single-nucleotide deletion in the putative Ori region and three relatively conservative amino acid substitutions in the essential ORFs, in each case introducing a Thr residue (Fig. (Fig.4A).4A). Further analysis would be needed to determine which of these mutations, individually or in combination, accounts for the temperature-sensitive phenotype. In any case, the very low stability of this plasmid at 37°C in the absence of positive selection with spectinomycin suggests that pMRMut-based cloning vectors will be useful in situations where it is advantageous to conditionally eliminate a vector (9, 15).
The minireplicon identified here differs from those previously described. The first report on a pXO1 origin of replication (19) did not actually demonstrate replication in B. anthracis but instead involving cloning of the region containing ORF72 to ORF80 as a shuttle vector into B. subtilis. The work was reported in a publication of a type that did not include details about the plasmid used for the cloning or the methods used for B. subtilis transformation or mini-pXO1 isolation. For these reasons and because the result has not been confirmed by others, the conclusion that this region contains a functional replicon must be viewed with skepticism. In our limited analysis of the region containing ORF72 to ORF80, we did not obtain a self-maintained replicon.
Surprisingly, the other reported pXO1 minireplicon containing the repX gene and adjacent ori region along with surrounding ORFs (22) also did not emerge from the Cre-loxP deletional analysis. This minireplicon did not maintain the remainder of pXO1 after deletion of the pXO1 region encompassing ORF14 to ORF16. Although the authors of the publication clearly demonstrated the presence of the repX-containing DNA in B. anthracis cells by Southern hybridization, no data on the stability of the putative minireplicon in B. anthracis were presented. The very low efficacy of B. anthracis transformation (1 to 2 CFU/μg) also raises concerns. The data presented do not appear to rigorously exclude the possibility that the plasmid DNA was spontaneously integrated into the B. anthracis genome in a recE-independent manner. Subsequent publications on RepX and the adjacent region have studied its biochemical properties in detail but have not directly shown that it plays a role in initiating pXO1 replication. Instead, evidence suggesting that this protein may play a role in plasmid segregation and partitioning has been presented (1, 2).
Although the two prior analyses of pXO1 discussed above identified regions that could replicate independently (19, 22), we did not observe deleted variants of pXO1 that contained those regions and lacked the ORF14 to -16 region (which we show can replicate independently). However, we did not use other approaches to rigorously examine or directly test the ability of the other two putative replicons to support independent replication. Thus, our analysis does not preclude the existence of other independent replicons within pXO1.
The success of the Cre-loxP strategy used here to identify the minireplicon results in part from its lack of bias. Deletions of large portions of pXO1 were achieved in an “in vivo” environment that does not require or involve cloning, transformation, or passage through alternative host organisms (e.g., E. coli). Furthermore, no assumptions need to be made about the structure of the replicon. All regions of the targeted plasmid are retained following Cre recombinase action, which creates two (or more) circular DNAs, each of which has the potential to replicate provided that it contains the necessary genes for replication and partition between dividing progeny bacteria. The method is therefore applicable to analysis of both simple and compound replicons. In the case of simple replicons, the essential genes will be isolated as a small plasmid, as occurred here for pXO1. For a compound replicon, any distantly located, separated, essential genes will be revealed when deletions of each required gene are made.
This research was supported by the Intramural Research Program of the NIH, National Institute of Allergy and Infectious Diseases.
Published ahead of print on 5 June 2009.