|Home | About | Journals | Submit | Contact Us | Français|
An improved genetic tool suitable for routine markerless allelic exchange in Bacillus anthracis has been constructed. Its utility was demonstrated by the introduction of insertions, deletions, and missense mutations on the chromosome and plasmid pXO1 of the Sterne strain of B. anthracis.
Bacillus anthracis, a gram-positive, spore-forming bacterium, is the causative agent of anthrax. Full virulence requires the production of a toxin and the formation of a protective capsule. The primary genes involved in the production of these two virulence factors are contained within two large nonessential plasmids, pXO1 and pXO2 (13). Distributed elsewhere on the two plasmids and on the chromosome are genes involved in regulating the expression of anthrax toxin, the capsule, and a number of other genes potentially involved in virulence (2, 6, 11). The entire genomic sequence of the Ames strain of B. anthracis has recently been determined, enabling the identification of interesting, potentially virulence-related genes by reverse genetics (19).
The ideal mutation to initially introduce in such “genome-mining” studies is an in-frame deletion of the candidate gene. Such a mutation avoids problems with polarity and other effects on the expression of surrounding genes, which accompany either insertion or less-precise deletion mutations and which can complicate interpretation. In addition, the ability to readily introduce missense mutations, which enables a determination of the effects of single-amino-acid substitutions, is necessary for more-sophisticated genetic analyses of structure-function relationships. Both of these types of desirable mutations are “markerless” in that they are not necessarily associated with a phenotype that can be selected or screened for during genetic manipulation (e.g., antibiotic resistance in the case of an insertion mutation). For B. anthracis, markerless gene replacements for some loci have been reported (3), but the methods by which these mutations have been isolated can be time and labor-intensive. This is due to the lack of a counterselection scheme, in which, by selecting for the loss of a plasmid vector, one can select for the second of two successive crossovers between such a vector and the chromosome in order to achieve gene replacement.
An alternative to counterselection schemes involves the use of the intron-encoded homing restriction enzyme I-SceI. The ability of this enzyme, which recognizes an 18-bp sequence, to cleave an introduced site that is essentially unique in a genome has been exploited in the promotion of homologous recombination in organisms as diverse as bacteria, Drosophila, and other higher eukaryotes (4, 20, 22). In one case, the use of I-SceI in promoting allelic exchange in Escherichia coli has been reported (17). In such a scheme, the integration of a suicide plasmid by a cloned region of homology containing the desired genetic change results in one of the two crossovers required to effect allelic exchange. To promote the second, the synthesis of the I-SceI enzyme results in cleavage at the unique I-SceI site within the vector. This double-stranded break is a potent substrate for host recombination systems that can repair the break by homologous recombination of the regions of sequence homology that flank the ends of the break as a result of the initial plasmid cross-in. As in allelic-exchange schemes driven by counterselectable markers, the loss of the plasmid sequences by homologous recombination leads to a population in which approximately 50% will have undergone the desired gene replacement. We have adapted this procedure for use with B. anthracis.
The method, described as follows, uses two plasmids, pBKJ236 and pBKJ223. These are illustrated schematically in Fig. Fig.1,1, and the steps in the method are depicted in Fig. Fig.2.2. Gene replacement constructs are first cloned into plasmid pBKJ236 for integration into the B. anthracis chromosome by homologous recombination. This vector was constructed by modification of pJRS233, which contains an erythromycin resistance gene, a replication origin for stable maintenance in Escherichia coli, and a temperature-sensitive replication origin for conditional maintenance in gram-positive organisms (16). In addition to these features, we added the oriT from RP4 to facilitate conjugative transfer from E. coli to B. anthracis (26) and the 18-bp recognition site for I-SceI. It should be noted that any suicide vector can be modified for use in this method simply by inserting an I-SceI site. We chose to separate the steps of plasmid transfer and recombination with the chromosome in order to overcome the relatively low efficiencies of genetic transfer from E. coli to B. anthracis. Thus, plasmid integrants are isolated by a shift to the replication-nonpermissive temperature after conjugative transfer and growth at the permissive temperature. The second plasmid, pBKJ233, is then introduced by electroporation and selection for tetracycline resistance. A derivative of pUTE29 (8), this plasmid contains the gene for the I-SceI enzyme under the control of a hybrid amylase promoter and gram-positive ribosome-binding site. Transformants are streaked twice on solid medium containing tetracycline, and then single colonies are scored for loss of erythromycin resistance. Following screening by PCR for the incorporation of the desired mutation, the pBKJ233 plasmid is lost spontaneously by streaking the cells twice on medium lacking tetracycline and scoring a small number of colonies for tetracycline sensitivity. The replicational instability of the pUTE29 vector has been previously described (21).
In order to demonstrate the utility of this approach, a number of mutations were introduced into B. anthracis 7702 (Sterne). We have also used this method with success with a similar strain, 34F2 (data not shown). The structures of the lesions are presented in Table Table1,1, and the efficiency of the pBKJ233-encoded I-SceI nuclease in stimulating the second crossover event and allelic exchange is presented in Table Table2.2. In separate control experiments with the strain harboring the pBKJ240 plasmid integrant, the pUTE29 vector showed no such stimulation (data not shown). Figure Figure33 shows the results of the PCR analysis demonstrating the incorporation of the altered allele in the resulting mutant strains. The plcR locus was used as an initial test of this method because it has previously been documented that the plcR gene is nonfunctional in B. anthracis due to a frameshift mutation (1). Both a clean, in-frame deletion of the plcR gene (ΔplcR240) and a deletion marked with a spectinomycin resistance cassette (plcR241::spc) were successfully introduced. The scrB gene was chosen since it presented a target with a potentially scorable colonial phenotype. This gene is predicted to encode sucrose-6-phosphate hydrolase, which is essential for the metabolism of sucrose by the sucrose phosphoenolpyruvate-dependent phosphotransferase system (7, 9). Analysis of the B. anthracis (Ames) genome suggested that no other pathway for catabolizing sucrose existed. Indeed, both the insertion mutation (scrB239::spc) and a missense mutation (scrB237) (G214Q) conferred a sucrose utilization defect visible as a change of colony color from yellow (wild type) to pink on nutrient agar supplemented with 1% sucrose and 0.0025% phenol red. This result validates the prediction that scrB represents the only pathway for sucrose utilization in B. anthracis. Since sporulation is also an easily scorable phenotype, we introduced an in-frame deletion (spo0A245) into the spo0A gene, a transcription factor required in the early steps of sporulation (14). As expected, the resultant mutant strain, BA722, yielded no detectable spores (CFU after treatment at 65°C for 30 min) under conditions (growth in Difco sporulation medium) which, for the B. anthracis 7702 parent strain, resulted in nearly 50% of the CFU being spores (data not shown). The identification of a deletion encompassing the spo0A gene in B. anthracis has been previously reported (27). However, this spontaneous deletion had endpoints in flanking DNA and thus affects other open reading frames (ORF) as well. To ascertain whether the method described here could be used for the replacement of genes on the large virulence plasmids of B. anthracis, the genes encoding the three components of anthrax toxin, pagA, lef, and cya, were targeted. Each of the three single mutant strains which resulted demonstrated a lack of production of the corresponding toxin component, but not of the other two toxin components, when tested by Western blotting (data not shown). Finally, in order to demonstrate the use of this tool to perform sequential mutageneses, a double lef cya mutant (BA721) was constructed. Synthesis of protective antigen, the product of the pagA gene, was normal for this strain, while lethal factor and edema factor were undetectable (data not shown).
In summary, the gene replacement procedure described here has been demonstrated to perform efficiently with B. anthracis and has been used to achieve allelic exchange of both marked and unmarked mutations, including deletions, insertions, and gene replacements as large as 1 kb or as small as 1 bp. This procedure can be used to target both chromosomal and plasmid-borne genes and is relatively fast. Once the initial crossover event has occurred, candidates for final screening can be generated in less than a week with little effort. An additional virtue is that genetic modification of the B. anthracis strain of interest is not required as it is in counterselection schemes such as that based on rpsL mutation and streptomycin sensitivity. Indeed, the natural resistance of B. anthracis to the antibiotic polymyxin B renders unnecessary even the introduction of a chromosomal marker for use in selection against the donor E. coli in conjugation experiments. As a result, any B. anthracis strain can be used, and engineered strains will be isogenic with the parent strain, with the exception of the introduced mutation(s). The ability to routinely introduce engineered alleles which are not marked with antibiotic resistance allows for the serial introduction of an unlimited number of such alterations. It also expands the repertoire of mutations that can be introduced to include single-base-pair changes, thus allowing detailed genetic analyses of the structure-function relationships of proteins and DNA sites in a B. anthracis genetic background.
Future work is required for the refinement of this system to accommodate select agents in cases where the use of some of the antibiotic markers implemented here would be prohibited. However, much experimental work is being performed with the non-human-pathogenic strain (Sterne) used in this study. The speed and efficiency of the techniques reported here allow the construction of multiple mutations in parallel and thus will enable types of comprehensive genetic analysis that have not been feasible for B. anthracis to date. It is our hope that these methods will facilitate the genetic study and manipulation of this fascinating and currently all-too-important bacterial pathogen.
We are indebted to Theresa Koehler for helpful discussions; to Gyorgi Posfai, June Scott, Robert Hartley, and Tod Merkel for providing strains; and to Drusilla Burns and Michael Schmitt for their critical reading of the manuscript.
Editor: J. T. Barbieri