3.1Construction and characteristics of the temperature-sensitive vector with kanamycin resistance cassette
We tested several temperature-sensitive plasmids for their ability to replicate in B. anthracis and be cured from the cells at nonpermissive temperature. Plasmids based on the widely used pE194ts replicon (e.g., pLTV1, pKSV7) were able to replicate in B. anthracis but curing was inefficient even at the highest temperature that this bacterium can survive (44 °C), with more than 50% of cells retaining the plasmid in a free form. In contrast, vectors pG-host9  and pTV1-OK , that are based on the mutant replicon  of the Lactococcus cremoris plasmid pWV01 , were stable in B. anthracis at 30 °C but completely disappeared after an overnight incubation at 37 °C.
Plasmids based on this replicon have rather unique properties. They can replicate in both E. coli and Gram-positive organisms such as Bacillus subtilis, group B Streptococci and Streptococcus mutans. Furthermore, their replication is highly temperature-sensitive and bacteria can be effectively cured from the plasmids at 37 °C [10–12]. The only exception is E. coli strain TG1 which for an unknown reason maintains pWV01-based plasmids at 37 °C, thus allowing for quick genetic manipulations unimpeded by slow bacterial growth .
On the basis of pG-host9 that carries EmR determinant we constructed a new plasmid pKS1, which contains a KmRaphA3 cassette [17,18] in the middle of a large multicloning site (Fig. 1). This plasmid replicates in both E. coli and B. anthracis and confers Km resistance and Em resistance to both of them. The E. coli carrying the plasmid are resistant to 300 μg ml−1 of Em and 50 μg ml−1 of Km. Transformants of B. anthracis harboring pKS1 after electroporation can be selected with Km (100 μg ml−1) and Em (3 μg ml−1).
The pKS1 plasmids with a variety of insertions into multicloning site were found to be structurally stable in E. coli as well as in B. anthracis. The plasmid is easily maintained at 30 °C in the Gram-positive host and was readily lost at temperatures of ?37 °C. Therefore, pKS1 is well suited to serve as a delivery vector to aid insertional inactivation of B. anthracis genes.
3.2Inactivation of comEC, racE1 and racE2 genes of B. anthracis
The procedure schematically shown in Fig. 2 has been tested on three B. anthracis genes: the close homolog (GenBank record NP658356) of B. subtilis comEC (P39695), which is involved in natural competence [19–21] but was found nonessential for B. subtilis growth , and two homologs of B. subtilis racE that encodes glutamate racemase . Glutamate racemase catalyzes formation of d-glutamate required for peptidoglycan biosynthesis and is an essential enzyme in B. subtilis and other bacteria. The RacE1 (NP654789) and RacE2 (NP658511) of B. anthracis are almost equally homologous to RacE (NP390717) of B. subtilis with 53% and 60% sequence identity, respectively. For the purposes of drug development, it was interesting to find out which of these racemase genes, if any, is essential when knocked out individually.
Each of these three B. anthracis genes were cloned into pKS1 as two DNA segments corresponding to the upstream and downstream halves of the gene. The corresponding regions of B. anthracis DNA were amplified by PCR and cloned into multicloning sites flanking the KmR determinant of pKS1. Their mutual orientation was preserved in the process. In the case of comEC the sizes of cloned DNA fragments were about 1 kb each, while both racE1 and racE2 genes were split into two ?400 bp fragments.
Since transformation efficiency of B. anthracis is strongly affected by adenine methylation of the plasmid , the resulting pKS1/comEC, pKS1/racE1 and pKS1/racE2 plasmids were passed, at 30 °C, through the E. coli strain GM2163 that lacks dam and dcm DNA methylases, and used to transform B. anthracis by electroporation. Transformants were selected on LB agar plate with Km (100 μg ml−1) and Em (3 μg ml−1) at permissive temperature (30 °C). Approximately 2–4 × 105 transformants per μg of DNA were obtained for all three plasmids.
One KmREmR transformant for each plasmid was randomly chosen for further analysis. It was grown without antibiotics for about 20 generations at a nonpermissive temperature (37 °C). At this temperature, replication of pKS1 derivatives ceases and the only bacteria that would retain antibiotic resistance should have the plasmid integrated into the chromosome. The cultures were plated on plates containing Km and on nonselective plates; the frequency of integration (KmR cells/total cells) was estimated to be about 5 × 10−2 in all three cases. Of ?100 tested KmR colonies all retained erythromycin resistance suggesting that KmR cells resulted from integration of the entire plasmid into the genome.
PCR analysis has shown that this integration occurs via a single cross-over event between the resident gene of the host strain and one of the homologous DNA fragments flanking the aphA3 gene (Fig. 2). Indeed, PCR fragments of expected sizes were obtained when PCR reactions were performed using DNA isolated from 6 KmR colonies as a template, and a combination of plasmid primers and primers located just outside of the homology regions. Furthermore, long PCRs with two outside primers revealed an insertion of a DNA fragment corresponding in size to the entire plasmid into chromosomal DNA isolated from three out of three analyzed colonies.
The final step in constructing gene knock-outs was to induce the second cross- over event between the homologous chromosomal DNA and the second region flanking the KmR gene as depicted in Fig. 2. One KmR ErR colony was chosen for each gene and grown in LB supplemented with Km for approximately 20 generations. The cultures were appropriately diluted and plated on Km plates at 37 °C. The colonies that lost EmR were then identified by replica-patching 150–300 colonies obtained on Km plates onto Em-containing LB agar plates (3 μg ml−1). Interestingly, the results differed dramatically depending on whether the growth of the KmREmR colonies in Km containing medium was performed at 30 or 37 °C. If it was done at 37 °C, all the KmR colonies preserved erythromycin resistance, indicating that homologous recombination in B. anthracis is a relatively rare event. However, in case of pKS1/comEC, if the growth of the culture was performed at 30 °C, then about 50–60% of the colonies showed KmR EmS phenotype, suggesting that intrachromosomal recombination shown in Fig. 2 did indeed occur.
This was confirmed by PCR that indicated that the plasmid disappeared from the chromosome leaving behind KmR gene inserted in the middle of comEC gene (not shown). The reason why incubation at 30 °C but not at 37 °C promotes intrachromosomal recombination is not entirely clear. It appears likely that replication from the plasmid replication origin, which becomes possible at 30 °C, increases the chances for this recombination to occur.
The situation was different for racE1 and racE2. Regardless of the temperature at which incubation of KmREmR cultures was performed, all tested KmR colonies retained erythromycin resistance. This was due to essentiality of both these genes. Indeed, when we supplemented the medium of racE1 recombinants with 2 mM d-glutamate and performed incubation at 30 °C, approximately 50% of KmR colonies showed no erythromycin resistance. This number was significantly smaller for racE2: only 0.5% of KmR colonies lost the integrated plasmid. PCR analysis confirmed the correctness of the final constructs: in both cases the KmR determinant inserted into the middle of the gene and the rest of the plasmid disappeared from the chromosome.
Further testing of racE1::KmR and racE2::KmR strains revealed that racE1 knock-out leads to a very moderate growth defect that can be fully alleviated by d-glutamate. In contrast, racE2 knock-out severely inhibits bacterial growth, and external d-glutamate restores it only partially (data not shown). These results indicate that of the two glutamate racemases of B. anthracis, RacE2 promises to be a much better drug target.
In summary, we have demonstrated the utility of pWV01-based vectors for genetic manipulations in B. anthracis, specifically for gene inactivation. The desired genetic rearrangements occur quickly and with high frequency. Although in this particular instance we simply disrupted the target genes, the same technique can be used for replacing the target gene with KmR cassette. Multiple other potential applications of the described genetic manipulations can be easily imagined. Plasmid pKS1 is available upon request.