The acrystalliferous strain B. thuringiensis 407 Cry- belonging to serotype 1 (Lereclus et al., 1989) was used throughout this study and designated as Bt 407. Escherichia coli K-12 strain TG1 was used for the construction of plasmids and cloning experiments, and E. coli ET 12567 Dam- Dcm- (Stratagene, La Jolla, CA, United States) was used to generate unmethylated DNA for the electrotransformation of Bt 407. E. coli and Bt 407 cells were transformed by electroporation as previously described (Dower et al., 1988; Lereclus et al., 1989). Bacillus strains were grown at 30°C in Luria Broth (LB) or in HCT, a sporulation-specific medium (Lecadet et al., 1980). E. coli strains were grown at 37°C in LB. For electrotransformation experiments, Bt 407 was grown in brain heart infusion broth (Difco). Media for bacterial selection were supplemented with 50 or 100 μg/mL X-Gal (5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside) for E. coli and B. cereus, respectively, and with 100 μg/mL ampicillin (Amp) for E. coli, and 200 μg/mL kanamycin (Km) or 10 μg/mL erythromycin for B. cereus as needed.

Chromosomal DNA was extracted from Bacillus cells harvested in the mid-exponential growth phase, with the GENTRA Puregene DNA Purification bacteria Kit (QIAgen, France). Plasmid DNA was extracted from E. coli by a standard alkaline lysis procedure using QIAprep spin columns (QIAgen, France). Restriction enzymes and T4 DNA ligase (New England Biolabs, United States) were used in accordance with the manufacturer’s recommendations. Oligonucleotide primers used in this study are listed in Table 1 and were synthesized by Sigma-Proligo (Paris, France). PCR was performed in a thermocycler, Applied Biosystems 2720 Thermal cycler (Applied Biosystems, United States). Amplified fragments were purified with the QIAquick PCR purification Kit (QIAgen). Digested DNA fragments were separated on 1% agarose gels and were extracted from gels with the QIAquick gel extraction Kit (QIAgen). Nucleotide sequences were determined by Beckman Coulter Genomics (Takeley, United Kingdom). Electroporation to transform Bt 407 was carried out as previously described (Lereclus et al., 1989).

A B. thuringiensis strain containing a dltX deletion was generated by precise, in frame allelic exchange and deletion replacement without antibiotic resistance cassettes. The thermosensitive plasmid MAD (pMAD) was used in these experiments. The 846 bp sequence immediately upstream from dltX was amplified with the primers dltX-a, dltX-b, and a 1064 bp sequence immediately downstream from dltX was amplified with the primers dltX-c and dltX-d. The primers dltX-b and dltX-c introduce overlapping PCR products. The two amplicons were then subjected to another PCR cycle with the primers dlt-a and dlt-d., such that a modified dlt operon, from which 159 bp had been deleted, was amplified. This amplicon was digested with NcoI and BglII and was introduced between the corresponding cloning sites of pMAD. Bt 407 was then transformed with 10 μg of the recombinant plasmid by electroporation as previously described (Lereclus et al., 1989). Transformants were subjected to allelic exchange by homologous recombination and bacteria sensitive to erythromycin, resulting from double crossing over event in which the chromosomal dltX copy was replaced with the overlapping sequences, were selected. The procedure for selection of mutants by allelic exchange via double crossover has been described previously (Bravo et al., 1996). Chromosomal allelic replacement in the dltX mutant was confirmed by DNA sequencing of the PCR fragments generated from the primer pairs dltX-a and dltX-d, and the resulting dltX-deficient strain was designated 407ΔdltX.

For the complemented strain, the entire dltX open reading frame of Bt 407 strain with its promotor region was amplified from genomic DNA of Bt 407 by PCR with Taq DNA polymerase (Extensor High-Fidelity; Thermo Scientific). The forward primer Comp F included a restriction site for XbaI and the reverse primer Comp R included a restriction site for XmaI. The resulting 611 kb fragment was digested with XbaI and XmaI, gel-purified, and ligated to the pHT304-18 shuttle vector previously digested with the same enzymes. An aliquot of ligation mixture (∼50 ng DNA) was used to transform E. coli TG1 by electroporation. The resulting construct, pHT304-18ΩdltX, was verified by restriction mapping and transferred into E. coli ET 12567 by electroporation. Unmethylated plasmid pHT304-18ΩdltX from E. coli ET 12567 was then introduced into strain Bt 407ΔdltX by electroporation.

The upstream promoter region of the dlt operon (Pdlt) was amplified by PCR with Taq DNA polymerase (New England Biolabs) and with the primer pair XbaI-dlt-gfp and EcoRI-dlt-gfp (Table 1). The 239 bp PCR product was digested with XbaI and EcoRI, and ligated into the same sites upstream of the promoterless gfp gene carried on pHT315Ωgfp (Daou et al., 2009) resulting in the creation of the plasmid pHT315ΩPdlt-gfp. The plasmid was verified by restriction analysis, PCR, and sequencing and was transferred into E. coli ET 12567 by electroporation. Unmethylated plasmids from E. coli ET 12567 were then introduced into Bt 407 and Bt 407ΔdltX mutant strains by electroporation.

The amounts of dltA transcripts in Bt407 and ΔdltX strains were measured by real time reverse transcription. RNA extraction and cDNA synthesis were performed as described previously (Réjasse et al., 2012). Primers BCF1372.Q3 and BCR1372.Q3 located inside the dltA gene were designed with Primer Express software from Applied Biosystems. Real time PCR was carried out with Sybr green PCR master mix (Applied Biosystems) as recommended by the supplier. Mean values were calculated from two separate experiments in which qPCR reactions were performed in triplicate. The cycle threshold was used to determine the relative dltA gene expression levels in the two genetic backgrounds. Data were analyzed by the comparative threshold cycle (ΔΔCt) method with the Relative Expression Software Tool (REST 2009 V2.0.13, QIAgen). Expression ratios were normalized to two Bt 407 endogenous reference housekeeping genes, pur and tpi.

Scanning electron microscopy (SEM) was performed at the Microscopy and Imaging Platform MIMA2 (Micalis, B2HM, Massy, France) of the INRA research center of Jouy-en-Josas (France). Samples were fixed with 3% glutaraldehyde in 0.1 M sodium cacodylate buffer at pH 7.4 at room temperature (RT) for 1 h. They were then washed three times for 5 min with a solution of 0.1 M sodium cacodylate. Samples were progressively dehydrated with increasing concentrations of ethanol (50%–70%–90%–2 × 100%) at RT for 10 min in each bath, except for 70% (35 min). Samples were critical point dried at 75 bar and 37°C in a Quorum Technologies K850 device (Elexience, France). Liquid CO₂ was used as the transition fluid. Coupons were mounted on aluminum stubs with double-sided sticky tape. Samples were sputter coated in Ar with Pt (30 nm of thickness) in a Polaron SC7640 device (Elexience, France) at 10 mA and 0.8 kV for 200 s. Observations were performed on a FE-SEM S4500 (Hitachi, Japan) with a sample holder tilted at 45° and a low SE detector, at 2 kV and 16 mm WD.

Strains were grown in LB medium to an OD of 2. A volume of 100 mL of cells was pelleted by centrifugation, washed twice with ammonium acetate buffer (20 mM; pH 4.7), and resuspended in 5 ml of the same buffer. Cells were heat-inactivated (10 min, 100°C) and lyophilized. D-alanine was released from whole cells by mild alkaline hydrolysis, as reported previously (Kristian et al., 2005). Mild alkaline hydrolysis was carried out with 10 mg of dried cells at 37°C for 1 h in 150 μL NaOH at 0.1 N. After neutralization with 0.1 N HCl, the cells were removed by centrifugation and the supernatant was dried under vacuum, and used for precolumn derivatization with Marfey’s reagent (1-fluoro-2,4-dinitrophenyl-5-L-alanine amide; Sigma) as described previously (Kochhar and Christen, 1989). Marfey’s reagent reacts with the optical isomers of amino acids to form diastereomeric N-aryl derivatives that can be separated by high-performance liquid chromatography (HPLC). Amino acid derivatives were separated on a C18 reversed-phase column (Hypersil 100, 250 mm × 4.6 mm, 5 μm, Thermo) at 30°C with a Waters HPLC system. Elution was performed with a linear gradient of acetonitrile in sodium acetate buffer (20 mM, pH 4) as described previously (Kristian et al., 2005). Absorbance of the eluate was monitored at 340 nm. D-Ala derivatives were identified by their retention time and quantified using an external standard calibration curve.

Bacteria were suspended in 1.5 × 10^-3 M sodium chloride at a concentration of 10⁸ CFU to measure electrophoretic mobility. The pH of the suspension was adjusted to pH 7 with nitric acid (HNO₃). Electrophoretic mobility was measured with a 50-V electric field and a Laser Zetameter (Zetaphoremetre II; Societé d’Etude Physico-Chimiques, Limours, France). The results were assessed from an automated video of about 200 particles for each measurement. Each experiment was performed in duplicate with two independently prepared cultures. The typical standard deviation for the electrophoretic mobility mean was 0.25 × 10^-8 m²s^-1V^-1.

Polymyxin B has long been used to define the mechanisms by which AMPs kill bacteria [46]. Susceptibility to polymyxin B was evaluated by determining the half inhibitory concentration (IC₅₀) from dose-response curves obtained with various concentrations of polymyxin B. The tests were performed in 96-well microplates containing 7–9 concentrations (from 200 to 800 μg/ml) of polymyxin B (Sigma) for WT 407 and 407ΔdltX complemented strains, and from 3 to 25 μg/ml for 407ΔdltX. Bacterial growth was scored after inoculation of strains at an initial OD₆₀₀ of 0.1 and incubation at 30°C for 6 h. IC₅₀ corresponds to the concentration of polymyxin B required to inhibit inoculum viability by half and was determined as the concentration required to bring the curve down to point half way between its top and bottom plateau. The results shown are the means of at least three independent experiments performed in duplicate.

Results concerning D-alanine quantification of TAs, electrophoretic mobility and in vivo pathogenicity assays were analyzed by the two-tailed Student’s t-test. A p-value of 0.05 was considered to be significant.

In silico BLAST searches indicate that the dltX gene is present in 809 species of which 805 are Firmicutes, including Bacillus, Staphylococcus, Listeria, Lactobacillus, Streptococcus, and Enterococcus. In most cases, dltX is located immediately upstream from dltA. In B. thuringiensis this operon consists of five genes dltXABCD. In strain Bt 407, transcripts of dltX have been detected by mapping RNA-seq datasets, obtained in vitro 2 and 5 h after the end of exponential phase and in vivo in infected G. mellonella cadavers, 36 h post mortem, with the Bt 407 reference genome (Sébastien Gélis-Jeanvoine, personal communication). These data prompted us to carry out a detailed analysis of the role of this protein in the process of D-alanylation. DltX belongs to the DUF3687 superfamily of proteins. To date, 811 sequences with this domain are listed in Pfam. Proteins in this family are approximately 50 amino acids in length and their protein sequences are highly conserved among the firmicutes. There are two completely conserved residues (L and Y, at positions 23 and 45 of the protein, respectively) that may be functionally important. A number of entries are annotated as D-Ala-teichoic acid biosynthesis protein; however, there is no direct evidence to support this annotation. In Bt 407, DltX contains 11 positively charged amino groups clustered near the N-terminus followed by a hydrophobic region of 22 amino acids and by a putative non-cytoplasmic domain of 15 amino acids in the C-terminus. The 22 amino acids hydrophobic region is predicted to form an α-helix, which suggests that DltX is a transmembrane protein. The topology of DltX present in the Gram negative species is the same, although there is no sequence similarity. In addition, DltX does not have a predicted signal peptide, which suggests that it is probably not exported (Figure 1).

We sought to study the influence of dltX on teichoic acid D-alanylation; therefore, we constructed a mutant harboring a precise in frame allelic deletion of the B. thuringiensis dltX gene, to avoid impact on the transcription of the downstream located dlt genes. We used the overlap extension (OE-PCR) method, which introduces a deletion without the use of antibiotic resistance cassettes (Figure 2). As a result of this deletion, the other genes of the operon are placed directly under the control of the upstream promoter region and are not affected in transcription. We first investigated WT and ΔdltX mutant growth dynamics by following optical density (at 600 nm) in liquid cultures and did not detect any significant differences in bacterial growth and cell doubling time. We then used a reporter construct in which the dlt upstream promoter region was fused to gfp, to investigate whether the deletion of dltX affects the transcriptional regulation of the dlt operon. We introduced the construct into WT and ΔdltX mutant backgrounds and quantified gfp expression in both strains. We found no differences in the expression of PdltΩgfp fusions (data not shown) between strains. We also performed RT-qPCR to examine the abundance of dltA mRNA, in WT and ΔdltX strains, during the mid-exponential growth phase in standard LB medium. We observed no differences in dltA expression between WT and dltX mutant cells. Together, these findings indicate that DltX does not affect the transcription of the operon.

We used SEM to analyze cells collected in mid-log phase to examine phenotypical differences between WT and dltX mutant strains. Deletion of dltX strongly affected cell surface morphology (Figure 3). WT cells had a typical rod shape with a regular and smooth surface (Figure 3A), whereas the dltX mutant cells had an irregular and wrinkled shape, and showed rib-like protrusions on their surface (Figure 3B). Complemented mutant cells (Figure 3C) showed an appearance similar to that of WT cells.

We compared the growth of WT Bt 407, Bt 407ΔdltX mutant and complemented mutant strain Bt 407ΔdltX (pHT304-18Ωdlt) exposed to polymyxin B or Nisin, to determine if dltX has a direct role in resistance to CAMPs. Addition of 150 μg polymyxin B ml^-1 or 75 μg Nisin ml^-1 to a growing culture caused immediate growth interruption of the mutant whereas WT and complemented strains were not affected (data not shown). Therefore, the dltX mutant is highly susceptible to CAMPs and its complementation with dltX completely restored the resistant parental phenotype. Moreover, the IC50 value of polymyxin B was 35 fold lower in the ΔdltX mutant than in either the WT or complemented strains. Precisely, the IC₅₀ value of the ΔdltX mutant was 13.9 μg ml^-1, that of the WT was 484 μg ml^-1, and that of the complemented strain was 503 μg ml^-1 (Figure 4). These results indicate that dltX is necessary for B. thuringiensis resistance to polymyxin B and that transcription of dltABCD alone under the control of the upstream promoter region is not sufficient to confer resistance.

The finding that dltX is essential for the resistance of Bt407 to CAMPs led us to investigate the effect of its deletion on bacterial virulence in two insect models. We injected 10⁴ vegetative cells of the WT, mutant, or complemented strains into fifth instar G. mellonella larvae and compared their virulence by monitoring the mortality level of Galleria infected with the different strains (Figure 5). The WT and complemented strains were significantly more virulent in insects (p < 0.0001) than the dltX mutant strain (Figure 5). In fact, virulence in G. mellonella was almost completely abolished in the dltX mutant (13% mortality) whereas WT and complemented strains were highly virulent and gave almost 100% mortality 48 h post-infection.

The immune system in insects comprises cellular and humoral responses that have been largely investigated in the model organism Drosophila melanogaster (Lemaitre and Hoffmann, 2007). The humoral response involves the infection-induced activation of NF-κB transcription factors, Dif and Relish, which in turn activate AMP encoding genes. Dif is downstream from the Toll pathway and Relish is downstream from the IMD pathway (Hedengren et al., 1999; Rutschmann et al., 2000). These signaling cascades are specifically elicited by microbial associated molecular patterns (MAMPs) that are detected by cognate pattern recognition receptors (PRRs), which are members of the Peptidoglycan Recognition Receptor and Glucan Binding Receptor families (Ferrandon et al., 2007). Notably, the Toll pathway is triggered upon sensing of Lysine (Lys)-type peptidoglycan, which is a common component of most Gram-positive bacteria, whereas the IMD pathway is activated by the mesodiaminopilmelic acid (DAP)-type peptidoglycan common to Gram-negative bacteria (Leulier et al., 2003; Ferrandon et al., 2007). The wall of all B. cereus group species also consists of a DAP-type peptidoglycan (Vollmer et al., 2008). We therefore, compared the virulence of WT, ΔdltX, and the complemented strain in WT oregonR and relish mutant flies to confirm that the loss of virulence phenotype of the ΔdltX mutant is associated with its susceptibility to CAMPs in vivo. Bt407 was resistant to the fly humoral response because WT and relish immunodeficient insects exhibited similar survival curves with high lethality of the infected flies at 15 h post-infection (Figure 6). Similar to findings obtained with Galleria larvae, survival curves of WT flies infected with the ΔdltX mutant showed that loss of dltX is sufficient to significantly impair virulence. Indeed, 75% of the flies infected with the ΔdltX mutant survived 24 h post-infection. Interestingly however, the virulence of the ΔdltX mutant was fully restored in relish mutant flies which do not produce CAMPs in response to the infection. This phenotype is associated with the loss of DltX because the complemented strain was as virulent as the WT strain in both WT oregon R and relish mutant flies. These findings clearly demonstrate that DltX is required for the resistance of Bt 407 to the humoral antimicrobial response of Drosophila infected by injection into the hemocoel.

We sought to determine whether the sensitivity of the dltX mutant to CAMPs was due to impairment in the incorporation of D-alanyl esters into TAs of the cell wall. D-Alanine was released by mild alkaline hydrolysis from whole heat-inactivated bacterial cells and quantified by HPLC analysis. Almost no D-alanine was released from the cell walls of the ΔdltX mutant, contrary to what is observed for the WT Bt 407, indicating that D-alanylation of TA was significantly impaired in the mutant. D-alanine amount released from the complemented strain was significantly higher than the one released from WT strain. This result can be explained by the copy number of the plasmid used for complementation (Figure 7).

The loss of D-alanylation in the dltX mutant may result in a net change in electrical charge of the bacterial surface, which may explain the higher sensitivity of the dltX mutant to CAMPs both in vitro and in vivo. We measured the electrophoretic mobility (EM) of each strain to see if the deletion of dltX was associated with a change of the global charge at the bacterial surface. EM was significantly higher in the ΔdltX mutant than in the WT strain (p = 0.0007) (Figure 8), showing that deletion of dltX alters electrical surface charge. Thus, the ΔdltX strain has a high negative charge at its surface.