The strains and plasmids used in this study were showed in Table 1. Escherichia coli strains Top10, ET12567, and BL21 (DE3) were used for molecular cloning, extracting non-methylated vectors, and protein expression, respectively (Wang et al., 2006, 2022). Bti strain Bt-59 (CGMCC strain number 16821) and Btk strain HD73 (BGSC strain number BGSC 4D4) were used to clone the target gene, for transformations and monitor promoter activity. E. coli strains were cultured at 37°C in Luria-Bertani (LB) broth medium with 100 μg/mL ampicillin (Amp) if needed. Bt strains were cultured at 30°C in Schaeffer’s sporulation Medium (SSM, 8 g L⁻¹ nutrient broth, 0.12 g L⁻¹ MgSO₄, 1 g L⁻¹ KCl, 0.5 mmol L⁻¹ NaOH) with 25 μg/mL erythromycin (Ery) if needed (Schaeffer et al., 1965).

Primers used in this study were synthesized in GENERAL BIOL (Hefei, China) and showed in Supplementary Table S1. PCR was performed using PrimeSTAR HS DNA polymerase (TaKaRa, Beijing, China) with corresponding primers. DNA fragments were purified with the TIANgel Midi Purification Kit (TIANGEN, Beijing, China) and plasmids were purified with the TIANprep Mini Plasmid Kit (TIANGEN, Beijing, China). The ClonExpress II One Step Cloning Kit (Vazyme, Nanjing, China) was used to clone the DNA fragments into the linearized vectors. Plasmids were introduced into the Bt cells by electroporation as previously described (Huang et al., 2022).

The temperature-sensitive suicide plasmid pMAD with the cwlE mutation sequence was constructed to delete the cwlE gene in the Bt-59 genome (Supplementary Figure S1). A 908-bp fragment upstream of cwlE (fragment A) was amplified by PCR, and cwlE-1 and cwlE-2 were used as primers. A 1013-bp fragment downstream of cwlE (fragment B) was amplified by PCR, and cwlE-3 and cwlE-4 were used as primers. The cwlE mutation box was amplified using fragments A and B as templates and cwlE-1 and cwlE-4 as primers by overlapping PCR. The resulting fragment was cloned into the pMAD plasmid to generate pMADΔcwlE plasmid. Then, the recombinant plasmid pMADΔcwlE was transformed into Bt-59 and Bt-59 (ΔsigK) to generate the cwlE deletion mutant Bt-59 (ΔcwlE) and the sigK and cwlE double mutant Bt-59 (ΔcwlE-sigK).

A 2693-bp fragment containing the promoter, ORF, and terminator sequences of cwlE was amplified by PCR using HFcwlE-1 and HFcwlE-2 as primers and cloned into the pHT315 shuttle vector to generate pHTHFcwlE plasmid. Then, the recombinant plasmid pHTHFcwlE was transformed into Bt-59 (ΔcwlE), Bt-59 (ΔcwlE-sigK), and HD (ΔcwlC) to generate complemented strains Bt-59 (ΔcwlE::cwlE), Bt-59 (ΔcwlE-sigK::cwlE), and HD (ΔcwlC::cwlE), respectively.

To further verify the hydrolytic activity of CwlE, an 822-bp fragment of cwlE gene was amplified by PCR using pET-cwlE-F and pET-cwlE-R as primers and cloned into the pET28a plasmid. Site-directed mutagenesis of cwlE (E24A and E151A) was performed with the Fast Mutagenesis System kit (TransGen, Beijing, China). Then, the recombinant plasmid pETcwlE, pETcwlE-E24A, and pETcwlE-E151A were transformed into BL21 to generate strains BL21(pETcwlC), BL21(pETcwlC-E24A), and BL21(pETcwlC-E151A) for protein expression. Besides, the recombinant plasmid pHTHFcwlE was transformed into HD (ΔcwlC) generating complemented strain HD (ΔcwlC::cwlE) to analyze the degradation activity of CwlE on Btk cell wall.

A 767-bp promoter region upstream of the cwlE gene (P _cwlE ) was cloned into the linearized vector pHT304-18Z to construct a cwlE reporter vector. Then, the recombinant plasmids pHTP _cwlE was transformed into cells of Bt-59, HD73, and HD (ΔsigE), generating strains Bt-59 (P _cwlE -lacZ), HD (P _cwlE -lacZ), and HD (ΔsigE) (P _cwlE -lacZ), respectively. Besides, the recombinant plasmid pHTHFcwlE was transformed into HD73 and HD (ΔsigE) generating strains HD (cwlE) and HD (ΔsigE) (cwlE) to analyze the effect of sigE deletion on the expression level of cwlE.

After strains BL21(pETcwlE), BL21(pETcwlC-E24A), and BL21(pETcwlC-E151A) were grown in LB liquid medium (Amp, 100 μg/mL) at 37°C for 4 h with shaking at 220 rpm, 0.5 mM isopropyl-beta-D-thiogalactoside (Sangon, Shanghai, China) was added. The bacteria were cultured at 16°C for another 14 h with shaking at 160 rpm to induce protein expression. Then, the recombinant CwlE proteins with a His tag were purified according to the method described previously (Zhang et al., 2020). After grown in LB liquid medium at 30°C for 6 h with shaking at 220 rpm, Bt-59 was collected by centrifugation at 10000 rpm. Then, the bacteria were resuspended with TK buffer (0.05 M Tris–HCl, 0.05 M KCl, pH 7.0) and disrupted with a FastPrep-24 (MPBiomedicals). Unbroken cells were removed by low-speed centrifugation (3,500 rpm, 5 min). The crude cell walls were pelleted at 13500 rpm for 5 min at 4°C. Finally, the pellets were resuspended in TK buffer. The purified CwlE proteins were added to the prepared cell wall of Bt-59 for hydrolytic activity analysis (Huang et al., 2022). After incubated at 37°C, the OD₅₄₀ of the mixtures were measured at the designated time point (Nugroho et al., 1999).

Total RNA was extracted from Bt-59 cultured in SSM at T₁₅ with TRIzol reagent (Invitrogen, USA). The transcription start site of the cwlE gene was determined with 5’ RACE according to the manufacturer’s instruction of HiScript-TS 5′/3’ RACE Kit (Vazyme, Nanjing, China). The primers used for 5’ RACE are listed in Supplementary Table S1.

Bt strains HD (P _cwlE -lacZ) and HD (ΔsigE) (P _cwlE -lacZ) were cultured in SSM (Ery, 25 μg/mL) at 30°C with shaking at 220 rpm. Two-milliliter bacterial suspensions of HD (P _cwlE -lacZ) and HD (ΔsigE) (P _cwlE -lacZ) were collected from T₀ to T₂₀ at 1-h intervals (T₀ is the end of the exponential phase and T_n is n hours after the end of the exponential phase). The cells were harvested and the samples were prepared as previously described (Chen et al., 2018). Then, the specific β-galactosidase activity was measured and expressed as Miller units per milligram of protein (Chen et al., 2018). The values are the means for three independent experiments.

Total RNA was extracted from Bt strains cultured in SSM at T₀, T₈ and T₁₆ with TRIzol reagent (Invitrogen, USA). T₀ is the end of the exponential phase and T_n is n hours after the end of the exponential phase. The purified RNA was reverse transcribed into cDNA using HiScript IV RT SuperMix for qPCR (gDNA wiper) (Vazyme, Nanjing, China). RpsU and gatB were chosen as reference genes (Reiter et al., 2011). qPCR was performed using specific primers (Supplementary Table S1) and ChamQ Universal SYBR qPCR Master Mix (Vazyme, Nanjing, China) with a 7,500 Fluorescence Quantitative PCR Instrument (ABI, California, USA). The specificity and qualified amplification efficiency of the specific primers were verified. Three biological replicates were designed. The relative expression level of each target gene was calculated according to the 2^-ΔCT method (Huggett et al., 2013).

The wild-type strain Bt-59, cwlE deletion mutant Bt-59 (ΔcwlE), and double mutant Bt-59 (ΔcwlE-sigK) were cultured in SSM medium at 30°C with shaking at 220 rpm. The growth curves of Bti strains were drew by measuring the OD₆₀₀ values of the culture solutions every hour. After 12 h of cultivation, the number of colonies of Bt-59 and its mutants were calculated by the coating plate method. SDS-PAGE was used for the crystal protein production assay as described previously (Xu et al., 2020).

Bt strains were cultured in SSM medium at 30°C with shaking at 220 rpm to the certain time points. Samples were collected and centrifuged, and then sediments were suspended with the appropriate volume of deionized water. Bacterial suspension was placed on a glass slide, and then cell morphology was observed by optical microscopy (ZEISS, Germany).

Bt-59 and Bt-59 (ΔcwlE-sigK) were cultured in a fermentation medium at 30°C with shaking at 220 rpm for 44 h as described previously (Huang et al., 2022). The fermentation media were serially diluted with water to the appropriate concentrations and used for the bioassay. Twenty larvae of Culex pipiens and Aedes albopictus in the early fourth instar were collected for bioassay in each treatment (Huang et al., 2022). Then the fermentation media were exposure to 365 nm UV light for 30 min used for the bioassay (Patel et al., 1996). The experiment was performed with three replicates at 28°C in a biochemical incubator for 24 h. The distilled water was used as a control.

Vector NTI 11 and GENEDOC were used for sequence analysis. The primer design was performed by Primer Premier 5.0. The phylogentic tree was constructed with MEGA Microsoft Office Excel and GraphPad Prism 6 were used for data analysis. PoloPlus statistical software was used for estimating the LC₅₀ values and 95% confidence intervals (CIs) (Straus, 2008). Lethal dose ratios (LC₅₀ ratio, LCR) were calculated to compare natural variation between different treatments, and the LCR value was considered to be significant (p < 0.05) when the confidence intervals did not encompass the value 1 (Straus, 2008).

To identify the essential cell wall hydrolase in Bti mother cell lysis, a novel hydrolase gene named cwlE was identified from a 128 kb sized plasmid (CP039725.1) encoding multiple insecticidal protein genes of the strain Bt-59. cwlE is 816 base pairs in length and encodes a putative cell wall hydrolase with the MurNAc-LAA domain and two highly conserved catalytic residues E26 and E151 (Figure 1). The amino acid sequence identity showed CwlE shares 32% of its amino acid sequence identity with CwlM of B. licheniformis, 28% with CwlC of Bt-59, and 33% with CwlC of B. subtilis. The absence of a homologous sequence of cwlE in the genomes of bacteria other than Bti strains, as determined by the NCBI homology analysis, suggests that cwlE is a cell wall hydrolase gene uniquely expressed in Bti.

To investigate the function of CwlE in Bti, the recombinant CwlE-His protein was purified using nickel column affinity chromatography. To determine whether the conserved glutamic acids are essential for enzymatic catalytic activity, two distinct single amino acid substitutions (E26A and E151A) were separately introduced into the CwlE-expressing strains using site-directed mutagenesis. SDS-PAGE showed that the molecular weight of the recombinant CwlE and its mutant proteins is approximately 35 kDa (Figure 2A). Then, mixtures of the recombinant proteins and the Bti cell wall were incubated at 37°C to determine the cell wall lytic ability of the CwlE and its mutant proteins. After incubation of CwlE with the cell wall for 60 min, the optical density (OD) of the mixture decreased by about 24% (Figure 2B). In contrast to the wild-type CwlE protein, the two mutant CwlE proteins (E26A and E151A) exhibited no catalytic activity against the Bti cell wall (Figure 2B).

After 15 days of growth in SSM medium the mother cells of Btk mutant strain HD (ΔcwlC) did not autolyze in a previous study (Chen et al., 2018). The cwlE gene containing its promoter region sequence was transferred into HD (ΔcwlC) strain to further determine the cell wall lytic ability of CwlE. Cells of HD (ΔcwlC::cwlE) began to autolyze at T₂₂ and completely autolyzed after 8 h of continuous culture (Figure 2C). The result indicated that the promoter of cwlE gene showed transcriptional activity in Btk, and CwlE could hydrolyze cell wall of HD73 and restore the phenotype of HD (ΔcwlC).

The result of 5’ RACE-PCR experiment showed that the transcription start site of cwlE gene is a T located 97 nucleotides upstream of its translational start codon ATG (Figure 3A). The putative cwlE promoter contains the sequences TAATATCT and AATAAAAT, which are located −35 and − 10 nucleotides upstream of the cwlE transcription start site, respectively (Figure 3A). These sequences are similar to those in the transcriptional regulatory factor SigE dependent promoters (KHATANHT and MATANNHT; H is A, C, or T; K is G or T; M is A or C; N is A, C, G, or T.) (Du et al., 2012), suggested that the cwlE promoter might be recognized and the expression of the cwlE gene was regulated by SigE.

The results of qPCR showed that the cwlE gene containing its promoter could be expressed in HD73, while it was almost undetectable in HD (ΔsigE) (Figure 3B). To further determine whether cwlE is regulated in a SigE-dependent manner, the Bt-59 (P _cwlE -lacZ), HD (P _cwlE -lacZ), and HD (ΔsigE) (P _cwlE -lacZ) strains were constructed. β-Galactosidase activity assays showed that reporter vector was expressed both in Bt-59 and HD73, but not in sigE mutant strain of HD73 (Figure 3C). The result showed that the cwlE promoter could be recognized by SigE in Bti and Btk, and the expression of cwlE was controlled by SigE. In Bt-59, the expression of the reporter vector started at T₀, reached a maximum at T₁₃, and then decreased in the following time (Figure 3C). Furthermore, the transcriptional activity of the cwlE promoter did not differ significantly between Bt-59 and HD73 (Figure 3C).

Both the cwlE deletion mutant Bt-59 (ΔcwlE) and the cwlE-sigK double deletion mutant Bt-59 (ΔcwlE-sigK) have been constructed and verified (Figure 4A; Supplementary Figure S2). The growth curve of Bti strains during the nutritional period in SSM medium revealed that cwlE deletion or cwlE and sigK double deletion did not affect the nutritional growth of Bt-59 (Figure 4B). After being cultured in SSM medium for 12 h, the total number of cells of Bt-59, Bt-59 (ΔcwlE), Bt-59 (ΔsigK), and Bt-59 (ΔcwlE-sigK) were 2.3 × 10⁸ CFU/mL, 2.4 × 10⁸ CFU/mL, 1.5 × 10⁸ CFU/mL, and 1.3 × 10⁸ CFU/mL, respectively (Figure 4C). Unlike cwlE single mutation, cwlE and sigK double mutation affected the cells number of Bt-59, similar to sigK single mutation constructed previously (Xu et al., 2020).

The expression levels of four main insecticidal protein genes (cry4Aa, cry4Ba, cry11Aa, and cyt1Aa) in Bt-59, Bt-59 (ΔcwlE), Bt-59 (ΔsigK), and Bt-59 (ΔcwlE-sigK) were revealed by qPCR (Figure 5). Compared with the wild strain, the deletion of cwlE or sigK did not affect the expression of cry4Aa and cry4Ba at T₀, T₈, and T₁₆. cry11Aa showed similar expression levels in the four strains at T₀ and T₈, while its expression was reduced in the mutants Bt-59 (ΔsigK) and Bt-59 (ΔcwlE-sigK) at T₁₆. The deletion of cwlE did not affect the expression of cyt1Aa, while the expression of cyt1Aa was reduced in cwlE and sigK double mutation strain Bt-59 (ΔcwlE-sigK). Besides, Cry proteins production of Bt-59 and its mutants were examined by SDS-PAGE after being cultured in SSM medium for 24 h. The results showed that deletion of cwlE had no effect on Cry proteins production in Bti, while the cwlE-sigK deletion decreased the expression of Cry11Aa (Supplementary Figure S3).

Optical microscopy was used to investigate the morphology of the Bt-59 and Bt-59 (ΔcwlE) grown in SSM medium to various growth phases (Figure 6). The result showed that cells of Bt-59 began to lyse at T₁₆, while cells of Bt-59 (ΔcwlE) began to lyse at T₂₀. Up to T₂₄, the cells of Bt-59 were completely lysed, and the most of the cells of Bt-59 (ΔcwlE) were lysed. After another 4 hours of culture, the cells of Bt-59 (ΔcwlE) were completely lysed. In addition, the cells of cwlE gene restoration strain Bt-59 (ΔcwlE::cwlE) began to autolyze at T₁₆, and virtually all of the cells were lysed at T₂₄, which was similar to Bt-59.

To further verify the role of cwlE in mother cell lysis of Bti, the cell wall hydrolase gene cwlE and the transcription regulatory factor gene sigK double deletion mutant Bt-59 (ΔcwlE-sigK) was constructed based on the sigK mutant strain. Optical microscopy was used to examine the morphology of Bt-59 (ΔsigK) and Bt-59 (ΔcwlE-sigK) grown in SSM medium to various growth phases (Figure 7A). The result showed that parts of cells of Bt-59 (ΔsigK) were lysed at T₄₈. Up to T₇₂, the cells of Bt-59 (ΔsigK) were completely lysed, whereas mother cell lysis of Bt-59 (ΔcwlE-sigK) could not be observed. The cell morphology of the complemented strain Bt-59 (ΔcwlE-sigK::cwlE) was comparable to that of the sigK mutant strain. Furthermore, Bt-59 (ΔcwlE-sigK) was cultured in SSM for 7 days to further examine the mother cell lysis phenotype. The results showed that the cells of Bt-59 (ΔcwlE-sigK) still did not autolyze after 7 days of growth (Figure 7B).

The early fourth-instar larvae of Culex pipiens and Aedes albopictus were fed with fermentation suspension serially diluted with water to the appropriate concentrations to determine the insecticidal activity of the cwlE and sigK double deletion mutant Bt-59 (ΔcwlE-sigK). The LC₅₀ values of Bt-59 and Bt-59 (ΔcwlE-sigK) against C. pipiens were 0.314 μL/L and 0.364 μL/L, respectively (Table 2). After exposure to UV light, the insecticidal activity of Bt-59 and Bt-59 (ΔcwlE-sigK) against C. pipiens reduced by 86.62 and 2.47%, respectively (Table 2). The LC₅₀ values of Bt-59 and Bt-59 (ΔcwlE-sigK) against A. albopictus were 2.066 μL/L and 2.075 μL/L, respectively (Table 3). After exposure to UV light, the insecticidal activity of Bt-59 and Bt-59 (ΔcwlE-sigK) against A. albopictus reduced by 52.76 and 12.43%, respectively (Table 3). The hypothesis of equality and parallelism showed that the cwlE and sigK double mutant did not affect the insecticidal activity of Bti against mosquito larvae (Figures 8A,C). In contrast, the level of decrease of insecticidal activity against mosquito larvae of Bt-59 (ΔcwlE-sigK) was less than that observed for Bt-59 after UV treatment (Figures 8B,D).