Mycobacterium smegmatis mc²-155 and M. tuberculosis H37Ra were obtained from the National Institutes for Food and Drug Control of China. Mycobacteria were grown in Middlebrook 7H9 broth (BD Difco) supplemented with 0.5% glycerol, 0.05% Tween 80, and 10% oleic acid-albumin-dextrose-catalase (OADC) (BD Difco) at 37°C. E. coli DH5α was grown in the LB medium at 37°C. When required, antibiotics were added as follows: zeocin 100 μg/ml, kanamycin 25 μg/ml, or hygromycin 150 μg/ml. Mycobacteria were plated on Middlebrook 7H10 agar (BD Difco) supplemented with 10% OADC for colony-forming unit (CFU) determination and were grown at 37°C. Recombinant purified proteins of CnpB (MSMEG_2630), DisA (MSMEG_6080), Ag85 (MSMEG_6398), and pyridoxamine 5′-phosphate oxidase (PdxH, MSMEG_5675) and their polyclonal antibody sera were prepared in our laboratory. The strains and plasmids used in this study are listed in Supplementary Table 1. Female BALB/c mice aged 6–8 weeks were purchased from the Animal Center of Air Force Medical University.

Mycobacterium smegmatis ΔcnpB mutant was constructed by homologous recombination using a previously reported approach (Van Kessel and Hatfull, 2007). Briefly, upstream and downstream homologous fragments of cnpB were amplified by polymerase chain reaction (PCR) with M. smegmatis mc²-155 genomic DNA as the template. The two flanking DNA fragments were sequentially subcloned into pMSG360zeo at EcoRV/HindIII and XbaI/KpnI restriction sites. The linear recombineering substrates flanked by homologous arms (loxp-zeo-loxp) were generated by EcoRV and KpnI digestion. The linear DNA substrates were electroporated into M. smegmatis cells that carried plasmid of pJV53. Then, cells were plated on 7H10 agar containing 100 μg/ml zeocin. The recombinant clone was verified by PCR and western blot analysis. The plasmid of pJV53 expressing gp60/gp61 was lost spontaneously. Primers used in this study are listed in Supplementary Table 2.

To complement the mutant, the cnpB open reading frame (ORF) was cloned into the single-copy expression vector of PW51 between KpnI/BamHI sites, and this plasmid of PW51-cnpB was transformed into ΔcnpB to complement the mutation (ΔcnpB::C). Meanwhile, cnpB ORF was cloned into the multiple-copy shuttle expression vector of PW54 at HindIII site. The recombinant plasmid of PW54-cnpB was transformed into ΔcnpB to generate cnpB-overexpressed strain (ΔcnpB::O). Transformants were verified by PCR and western blot analysis.

Mycobacterium smegmatis strains were grown in 15 ml of 7H9 broth supplemented with 0.5% glycerol, 0.05% Tween 80, and 10% OADC at 37°C until OD₆₀₀ = 1.0. Bacteria were harvested and washed three times with phosphate-buffered saline (PBS). Pellets were resuspended in precooled PBS supplemented with proteinase inhibitor cocktail (Roche) and then lysed by MINI-BEAD BEATER (BioSpec). The lysates were centrifuged at 12,000 rpm for 20 min at 4°C, and the supernatants were collected. Besides, the bacterial culture supernatants were concentrated by vacuum freeze-drying for western blot analysis. The concentrations of proteins were measured by the bicinchoninic acid assay (BCA). Phenotypes of M. smegmatis strains were confirmed by western blot analysis with polyclonal antibodies as the primary antibody, as shown in Supplementary Table 1. PdxH was used as a loading control, and recombinant proteins were used as positive controls.

Mycobacterium smegmatis strains were grown to OD₆₀₀ = 1.0 in 7H9 broth with 10% OADC. Intracellular nucleotides were extracted from bacterial pellets according to the reported method (Burhenne and Kaever, 2013). Briefly, bacterial pellets were resuspended in extraction buffer (acetonitrile: methanol: water = 2:2:1), incubated on ice, boiled at 95°C for 10 min, and incubated on ice. The lysates were centrifuged for 10 min at 20,800 × g at 4°C. Supernatants were transferred to a new tube. The extraction procedure was repeated two times. The pooled extractions were lyophilized and dissolved in ddH₂O. c-di-AMP levels were determined by high-performance liquid chromatography (HPLC) as previously reported (Ning et al., 2019). Chemically synthesized c-di-AMP (InvivoGen) was used as a standard substance to generate a standard curve. The concentrations of c-di-AMP in M. smegmatis strains (1 g wet weight) were displayed as nanomoles per liter (nM).

Mycobacterium smegmatis strains (2 × 10⁵CFU) were grown in 15 ml 7H9 broth with 10% OADC or in Sauton’s medium in 50 ml tube under shaking (80 rpm and 180 rpm) or stationary conditions at 37°C. The growth of each strain was determined by measuring OD₆₀₀ at indicated time points. The bacterial length was determined according to our previous study (Ning et al., 2019). Briefly, bacteria in the late-log phase were collected, washed with PBS, and resuspended in PBS. Bacterial suspensions were dropped on copper grids (200 mesh) for 10 min incubation, and excess bacterial suspension was removed by the filter paper. Bacteria were stained with phosphotungstic acid, and the bacteria were observed by transmission electron microscopy (TEM) (TECNAIG2 Spirit Biotwin). The bacterial length was measured by ImageJ software. The individual cells were manually traced in ImageJ and converted bacterial arbitrary units to actual size according to the scale of 5 μm. Bacterial cells in clumps were omitted during measurement.

Mycobacterium smegmatis strains were grown in a 7H9 medium with 10% OADC to OD₆₀₀ = 0.8–1.0 at 37°C. Bacteria pellets were resuspended and washed two times by Sauton’s medium in the absence of Tween 80. The cultures were diluted by Sauton’s medium at OD₆₀₀ = 0.2 and inoculated into 24-well (2 ml/well, for imaging) or 96-well (0.2 ml/well, for quantification) plates. The plates were incubated statically at 37°C for 3–5 days. Biomass was quantified by a crystal violet assay. Briefly, biofilms in the well were washed with ddH₂O three times to remove the medium. The biofilms were dried at 37°C and incubated with 200 μl of 1% (w/v) crystal violet for 30 min and washed two times with ddH₂O. About 200 μl of 95% ethanol was added to dissolve the biofilm. The OD₅₇₀ was recorded by a microplate reader. Each experiment was performed in a set of three replicates.

Mycobacterium smegmatis strains were inoculated in a 7H9 medium and cultured for 3 days (stationary phase) at 80 rpm at 37°C. Subsequently, the cultures were transferred to a 37°C incubator for static culture avoiding light. After incubation at 14 and 60 weeks, bacteria were harvested and washed with PBS. Red autofluorescence of the bacterial cells was monitored by the flow cytometer (BeckmanCoulter EXPO32) in the FL3 channel (excitation 488 nm with a 610LP emission filter). Fresh bacteria (3 days) were used as the control. Flow cytometry data were analyzed using the FlowJo V10.

The 14- and 60-week supernatants of M. smegmatis strains were collected and filtered with a 0.22 μm filter membrane. Absorption spectra of supernatants were recorded by a multifunctional microplate reader (Tecan Spark) from 100 to 1,000 nm. Then, fluorescence measurements were taken with TECAN Spark at the excitation wavelength (λ_excitation) of 400 nm.

Total RNAs were extracted from M. smegmatis by TRIzol reagent (Invitrogen) according to the manufacturer’s protocol. RNA quality was assessed on an Agilent 2100 Bioanalyzer (Agilent Technologies) and RNase-free agarose gel electrophoresis. mRNA was enriched by removing rRNA by Ribo-Zero Magnetic Kit (Epicenter) for library construction. RNA sequencing (RNA-seq) was performed using Illumina Novaseq6000 by Gene Denovo Biotechnology Co., Ltd., (Guangzhou, China). The expression levels of differential genes were further verified by real-time quantitative PCR (qRT-PCR). In total, 500 ng total RNA was reverse-transcribed using the HiScript III cDNA Synthesis Kit (Vazyme). The transcriptional levels of target genes were determined by qRT-PCR using SYRB Master Mix (Vazyme). Primers used in the qRT-PCR assay are listed in Supplementary Table 2. The relative transcriptional levels of genes were presented as 2^{–Δ Δ Ct} with M. smegmatis sigA as a reference gene.

Female BALB/c mice were anesthetized with an intraperitoneal injection of 50 mg/kg pentobarbital sodium. Mice were immunized with 10⁷CFU M. smegmatis in 100 μl PBS for three times at 2-week intervals through a subcutaneous (s.c) route on the back. An equal volume of PBS was injected into control mice. Then, 4 weeks after immunization, mice were challenged intravenously (i.v) with 5 × 10⁴ CFU M. tuberculosis H37Ra in 200 μl PBS.

Whole blood was collected at 4 weeks post-immunization and 8 weeks post-infection. The whole blood was kept at 37°C for 30 min to collect the supernatant as serum. Plates were coated with M. smegmatis total protein extract of mycobacteria and CnpB. The serum was diluted at 1:200 with PBS. Antigen-specific IgG subclasses were determined by indirect enzyme-linked immunosorbent assay (ELISA). For cytokine measurement of IFN-γ, IL-2, and IL-10, splenocyte supernatants were collected after 72 h stimulation with M. smegmatis protein extract (25 μg/ml) as described in our previous study (Ning et al., 2019). Double antibody sandwich ELISA was performed according to the manufacturer’s instructions (Thermo Fisher Scientific).

Spleen single-cell suspensions were prepared as we previously described (Ning et al., 2019). For cells proliferation assay, 1 × 10⁶ cells were seeded in 96-well microplates and stimulated with CnpB (5 μg/ml), M. smegmatis protein extract (25 μg/ml), or M. tuberculosis H37Ra protein extract (25 μg/ml) as indicated. After stimulation at 37°C for 68 h with 5% CO₂, MTS reagents (Promega) were added in each well and plates were incubated for another 4 h at 37°C. The absorbance at 490 nm was recorded using a microplate reader. The stimulated index was calculated as described in our previous study (Lu et al., 2018).

For histopathology, the upper lobes of the left lungs were fixed in 10% buffered formalin for 24 h. Then, sections of 5 μm in thickness of tissues were cut into glass slides. Hematoxylin–eosin (H&E) staining for pathohistological analysis was performed by the Department of Histopathology (Air Force Medical University, China). Sections were observed under a light microscope (Olympus). At 8 weeks after M. tuberculosis infection, the lung and the spleen were removed and homogenized in sterile PBS. Homogenates were diluted and spread on 7H10 agar plates with 10% OADC enrichment. Plates were then incubated at 37°C for 3 weeks. The data are represented in Log₁₀CFU (lgCFU).

Animal experiments were performed according to the approval and guidance of the Institutional Animal Ethics Center of the Air Force Medical University, and the approval number is 20190213.

Statistical analysis was performed by GraphPad Prism 8.0 software. Statistical significance was determined by Student’s t-test or for multiple comparisons by ANOVA. P < 0.05 were considered statistically significant. Data are shown as means ± SEM or as individual data points, i.e., “*” denotes p < 0.05, “^**” denotes p < 0.01, “^***” denotes p < 0.001, and “^****” denotes p < 0.0001.

Mycobacterium smegmatis CnpB shared 73% identity (247/338 aa) with M. tuberculosis CnpB (Supplementary Figure 1A). In addition, M. smegmatis CnpB carries a highly conserved motif of Asp-His-His (DHH) (Supplementary Figure 1B), which is indispensable for the degradation of c-di-AMP (Yang et al., 2014; Tang et al., 2015). In this study, cnpB mutant (ΔcnpB) of M. smegmatis mc²155 was generated by homologous recombination through Che9c gp60/gp61 recombinase-assisted genome engineering (Figures 1A,B). Meanwhile, complemented (ΔcnpB::C) and overexpressed (ΔcnpB::O) strains of ΔcnpB were generated by different plasmids with the cnpB gene from M. smegmatis. These strains were verified by PCR and western blot analysis (Figures 1B,C). CnpB expression was restored in ΔcnpB::C and ΔcnpB::O (Figure 1C). High-performance liquid chromatography (HPLC) analysis showed that intracellular c-di-AMP concentration was elevated 1.5-folds in ΔcnpB than that of wild-type strain (p < 0.01) (Figure 1D). The level of c-di-AMP in ΔcnpB::C was lower than that of ΔcnpB but was comparable to that of wild-type (Figure 1D). However, the c-di-AMP level of ΔcnpB::O was about a quarter of that in ΔcnpB (Figure 1D), which could be explained that more c-di-AMP was hydrolyzed in M. smegmatis cnpB-overexpressing strain. These results showed that the c-di-AMP level of M. smegmatis could be controlled by the expression level of CnpB.

It has been found that increased c-di-AMP level by DisA overexpression in M. smegmatis led to abnormal phenotypes, including cell expansion, cell bulge phenotype, bacterial aggregation, and loss of motility (Zhang and He, 2013; Tang et al., 2015). It was reported that the morphology of CnpB-overexpressed strain was normal as wild-type M. smegmatis (Tang et al., 2015). In our study, the ΔcnpB colonies showed a smaller, smooth, and moist phenotype compared with that of the wild-type (Figure 2A). The colonies of ΔcnpB::O showed a similar appearance but has fewer wrinkles (Figure 2A). Colony diameter measurements showed that ΔcnpB formed smaller colonies with 28% reduction in diameter (p < 0.001) (Figures 2A,B). ΔcnpB::C just partially restored the colony size compared with ΔcnpB (p < 0.05) and smaller than wild-type (p < 0.001) (Figures 2A,B). However, ΔcnpB::O with the lowest c-di-AMP level exhibited a comparable colony size with wild-type, but other appearances seemed to be different from wild-type (Figures 2A,B).

Elevated c-di-AMP level in M. tuberculosis by CnpB deletion or in BCG by DisA overexpression led to a reduction of the single bacterial length (Yang et al., 2014; Ning et al., 2019). We speculated that the smaller bacterial colony is due to the smaller bacterial length. As expected, transmission electron microscope (TEM) observation showed that ΔcnpB (4.16 ± 1.73 μm) exhibited a bacterial length reduction of approximately 28% relative to the wild-type (5.83 ± 1.15 μm) M. smegmatis (Figures 2C,D). These results confirmed previous reports that the c-di-AMP level affected colony morphologies as well as bacterial length in mycobacteria and suggested that other phenotypes were controlled by c-di-AMP.

In M. tuberculosis, deletion of c-di-AMP phosphodiesterase did not affect M. tuberculosis growth by OD₆₀₀ (Yang et al., 2014), and a similar result was observed in L. monocytogenes (Witte et al., 2013). The growth of wild-type, ΔcnpB, ΔcnpB::C, and ΔcnpB::O was monitored in Middlebrook 7H9 medium and Sauton’s medium. Our data showed that the growth rate of ΔcnpB was slower in the exponential phase in 7H9 determined at both OD₆₀₀ and CFU detections but reached similar levels at the stationary phase (Figures 3A–D). Meanwhile, ΔcnpB::C and ΔcnpB::O had similar growth rates as that of wild-type in 7H9 broth without the addition of OADC (Figures 3A,B). Similar results were observed when bacteria were grown in 7H9 + OADC (Figures 3C,D). Sauton’s medium is a minimal medium that does not contain exogenous proteins and is commonly used for the preparation of mycobacterial culture filtrates (Venkataswamy et al., 2012). In Sauton’s medium with or without OADC enrichment, ΔcnpB had a significantly reduced growth rate (Figures 3E–H). Both ΔcnpB::C and ΔcnpB::O showed a similar growth rates/CFUs to that of wild-type in 7H9 and Sauton’s medium with OADC (Figures 3E–H). The reduction of ΔcnpB growth was also found under slow shaking conditions (80 rpm) (Supplementary Figures 2A–H). Collectively, elevated c-di-AMP level impaired M. smegmatis growth, especially in media with less complex nutrients.

It has been reported that smooth or rough/wrinkled colony morphology was associated with the biofilm formation in several bacteria (Friedman and Kolter, 2004; Uhlich et al., 2006; Bharti et al., 2020). Further, the ability of ΔcnpB strain to form biofilm was determined in Sauton’s medium at 48, 72, and 120 h respectively, and biomass was quantified by the crystal violet assay. It was observed that ΔcnpB could not form the biofilm with rough surfaces like M. smegmatis and showed a smooth pellicle with low biomass quantifications by crystal violet assay in 48 and 72 h culture (Figures 4A,B). Though ΔcnpB::C showed a decrease in biofilm formation in 72 h culture, but more biomass quantification was detected in 120 h culture than ΔcnpB and was comparable to the wild-type strain (Figure 4C). To our surprise, ΔcnpB::O with decreased c-di-AMP level showed similar biofilm formation as ΔcnpB::C (Figures 4A–C). Nevertheless, our experimental results demonstrated that c-di-AMP could regulate the biofilm formation of mycobacterium, so the mechanisms related to biofilm regulation should be investigated further.

Unexpectedly, we observed brown coloration in the culture of ΔcnpB, ΔcnpB::C, and ΔcnpB::O over 14 weeks, and supernatants of ΔcnpB::C and ΔcnpB::O showed lighter brown color than that of ΔcnpB (Figures 5A,B). The brown coloration of ΔcnpB::C and ΔcnpB::O supernatants gradually decreases with incubation time and were comparable to wild-type of M. smegmatis in 60-week culture (Figure 5A). After centrifuging, the brown coloration was retained in supernatant rather than bacterial pellets of M. smegmatis strains, suggesting that the unknown brown material substance was water soluble (Figure 5B). In addition, the brown pigment could not be removed by dialysis but be salted out from ΔcnpB supernatant by saturated (NH₄)₂SO₄, which suggested that the brown pigment could bind to protein (Figure 5C and Supplementary Figure 3A). According to the flow cytometry experiments, ΔcnpB cells from long-term (14 and 60 weeks) rather than fresh (3 days) culture revealed red fluorescence (Figure 5D). Then, an absorbance peak at 400 nm was found in ΔcnpB supernatant of 14-week culture recorded by full-wavelength (100–1,000 nm) scanning (Figure 5E) and a lower absorbance peak in that of ΔcnpB::C and ΔcnpB::O (Supplementary Figure 3B). It was observed that M. smegmatis produced a dark brown fluorescent pigment in long-term culture, which was identified as porphyrin by mass spectrometry analysis and ¹H-NMR spectra (Nikitushkin et al., 2016). Porphyrin has the typical fluorescence at 400 nm and fluorescence emission maxima at 625 and 675 nm (excitation wavelength, 400 nm) (Gouterman, 1959, 1961; Dolphin, 1978; Nikitushkin et al., 2016). Further, we found the characteristic fluorescence of ΔcnpB supernatant at 625 and 675 nm with an excitation wavelength of 400 nm (Figure 5F and Supplementary Figure 3C). Taken together, our results suggested that the elevated c-di-AMP level may increase porphyrin production in M. smegmatis.

To further explore gene expression profiles regulated by c-di-AMP, M. smegmatis and ΔcnpB pellets of late logarithmic phase cultures were harvested for RNA-seq analysis. Differentially expressed genes (DEGs) were analyzed at the cutoff of both false discovery rate (FDR) < 0.05 and | log2FC| > 1. Overall, it was found that a total of 2,828 genes were differentially expressed between ΔcnpB and wild-type, including 1,870 upregulated and 958 downregulated compared with wild-type strain. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment showed that several major metabolism-related genes were repressed in ΔcnpB, including nitrogen, alanine, aspartate, and glutamate, microbial metabolism in diverse environments, galactose metabolism, purine, carbon, and so on (Figure 6A). On the other side, elevated c-di-AMP in M. smegmatis induced expressions of genes involved in the biosynthesis of aminoacyl-tRNA, amino acids and so on (Figure 6B). We also noticed that gene expressions of porphyrin and chlorophyll metabolism were also induced in ΔcnpB (Figure 6B).

Further, qRT-PCR was used for the analysis of gene expressions associated with the phenotypes found above. We had observed that ΔcnpB with high c-di-AMP reduced the biofilm formation in the 7H9 complete medium (Figure 4). The impaired or absent function of the two-component system results in defective biofilm formation in M. smegmatis (Nguyen et al., 2010; Li et al., 2020). Deletion of cnpB with increased c-di-AMP downregulated the genes of two-component systems involved in biofilm formation, such as MSMEG_1875 (mtrB), MSMEG_1874 (mtrA), MSMEG_0246 (prrB), and MSMEG_0244 (prrA) (Figure 6C). ΔcnpB showed porphyrin accumulation in the long-term medium (Figure 5 and Supplementary Figure 3). Elevated c-di-AMP in M. smegmatis significantly upregulated the expression of porphyrin-metabolizing enzymes in 3-day culture, including MSMEG_2618 (uroporphyrin-III C-methyltransferase), MSMEG_0954 (uroporphyrinogen-III synthase), and MSMEG_2780 (uroporphyrinogen decarboxylase) (Figure 6C), and downregulated the expression of MSMEG_4525 (putative oxygen-independent coproporphyrinogen III oxidase) (Figure 6C), which may result in more porphyrin production in ΔcnpB.

It has been reported that c-di-AMP specifically binds to K⁺ transport proteins such as KdpD and KtrA, TrkA, and TetR family proteins to inhibit potassium transporter function (Bai et al., 2014; Kim et al., 2015; Ali et al., 2017). We found that several potassium ion transport gene transcriptions were downregulated in ΔcnpB, including MSMEG_5392 (kdpA), MSMEG_5393 (kdpB), MSMEG_5394 (kdpC), MSMEG_5395 (kdpD), and MSMEG_2769 (trkB) (Figure 6C). These data suggested that c-di-AMP inhibited the uptake of potassium ions by bacteria not only because of its direct binding to potassium ion transporters but also may be related to the inhibition of transporters transcription.

It has been demonstrated that c-di-AMP modulates bacterial virulence and affects host immune responses (Blander and Barbet, 2018). There were several different protein bands appearing in supernatants between wild-type and ΔcnpB strain but they were not obvious in bacterial lysates (Supplementary Figures 4A,B). qRT-PCR showed that elevated c-di-AMP downregulated the expression of early secreted antigens of MSMEG_0066 (esat-6, esxA) and MSMEG_6398 (ag85, fbpA) in ΔcnpB (Figure 6C), and ΔcnpB seemed to secrete less Ag85 than wild-type M. smegmatis (Supplementary Figure 4C). While host cell death-associated MSMEG_1638 and PE/PPE family proteins of MSMEG_5350 (PPE 63) were upregulated, it was speculated that CnpB deletion might change the immunogenicity of M. smegmatis in vivo.

c-di-AMP could induce innate immune responses and enhance antigen-induced Th1/Th2/Th17 responses as an adjuvant (Ebensen et al., 2011; Devaux et al., 2018; Ning et al., 2021). It has been revealed that M. tuberculosisΔcnpB could trigger host innate immune responses, which could promote the clearance of bacteria (Yang et al., 2014; Dey et al., 2017). We demonstrated that recombinant BCG with elevated c-di-AMP by overexpressing DisA induced enhanced immune responses and provided similar protection as BCG in the same M. tuberculosis intravenous infected mice model (Ning et al., 2019). In this study, humoral and cellular immune responses induced by ΔcnpB were also investigated (Figure 7A). ΔcnpB subcutaneous immunization induced comparable IgG levels with the naïve group (Figure 7B). Splenocyte proliferation assay showed that ΔcnpB immunization induced significant proliferation than the naïve group responding to mycobacterial proteins (p < 0.01) but no more than that of M. smegmatis immunization (Figure 7C). ΔcnpB immunization induced increased Th1 (IFN-γ, IL-2) and Th2 (IL-10) cytokines’ secretion in supernatants of splenocytes restimulated with M. smegmatis proteins than control mice (p < 0.05), but no difference was found between two immunization groups (Figures 7D–F). These results showed that the elevated c-di-AMP level could not enhance the immunogenicity of M. smegmatis as immunized subcutaneously in mice.

Post 4 weeks of the last immunization, mice were challenged with M. tuberculosis intravenously (Figure 7A). Then, 8 weeks post-infection, ΔcnpB immunized mice showed an increase in IgG level, which were significantly higher than that of challenged mice without immunization (UN), and higher than the Ms group though no significant difference (Figure 8A). After M. tuberculosis infection, both wild-type and ΔcnpB strain-immunized mice produced significantly increased Th1 cytokines (IFN-γ, IL-2) rather than Th2 cytokine (IL-10) compared with unimmunized mice (Figures 8B–D). It was noted that ΔcnpB immunization could induce more IFN-γ and IL-2 secretion in splenocyte supernatant than that of wild-type M. smegmatis (Figures 8B,C). IL-10 secretion showed no differences between groups (Figure 8D). Overall, these results indicated that ΔcnpB immunization promoted significant Th1-type immune responses but not Th2-type immune responses after M. tuberculosis infection in mice.

ΔcnpB immunization stimulated Th1 instead of Th2 immune responses, which is considered to induce immunopathological damage after M. tuberculosis infection in mice (Da Silva et al., 2015). At 8 weeks post-infection, all infected mice showed similar pathological manifestations of chronic inflammation by H&E staining, including partially damaged alveolar structures, occasional thickening of alveolar mesenchyme, inflammatory cell infiltration, and erythrocyte and histological fluid exudation, but no granuloma or nodules were found in the slide of lungs (Figure 8E).

After mice were infected with M. tuberculosis intravenously, both M. smegmatis strains’ immunization could reduce the bacteria burdens in both lungs (0.87–1.02 lgCFU) and spleens (0.87–1.3 lgCFU) compared to the unimmunized group (p < 0.05) (Figure 8F). There were no statistical differences in bacteria loads of lungs and spleens between M. smegmatis and ΔcnpB-immunized mice (Figure 8F). Overall, ΔcnpB immunization provided similar protections as wild-type M. smegmatis against M. tuberculosis intravenous infection in mice.