Experiments were performed with the MTB reference strain H37Rv, M. bovis bacillus Calmette-Geurin (BCG) (Pasteur), and its isogenic mutants ΔespB and ΔhtpG. ΔespB and ΔhtpG were constructed as described previously (Jain et al., 2014). MTB strains were grown in Middlebrook 7H9 medium supplemented with 0.5% glycerol, 10% oleic acid-albumin-dextrose-catalase (Becton Dickinson, Franklin Lakes, NJ, United States) and 0.05% Tween-80 or on Middlebrook 7H10 agar (BD Difco Laboratories, Sparks, MD, United States) supplemented with 10% oleic acid-albumin-dextrose-catalase.

The crRNP complex was isolated using a strategy similar to that described by Wei et al. (2019). MTB cells were resuspended and disrupted in Fastprep-24 (MP Biomedicals, Santa Ana, CA, United States). The cell extract was subjected to ultracentrifugation, after which the supernatant was filtered twice with a 0.22 μm pore Millex filter unit (Millipore Sigma, Burlington, MA, United States) and loaded onto a 5 mL Q-Sepharose Fast Flow (GE Healthcare, Chicago, IL, United States) prepacked column. The proteins were eluted using a 0–1 M NaCl gradient. RNA was isolated from 200 μL of each fraction with TRIzol (Invitrogen, Carlsbad, CA, United States) according to the manufacturer’s instructions, and the fractions were analyzed via northern blotting and separated on 15% Tris-borate-ethylenediaminetetraacetic acid-urea gels. Blots were analyzed for the presence of spacer 1.04 to guide the purification process (Wei et al., 2019). The fractions, which had strong and clear signal for probes, were pooled, diluted in 50 mM Tris (pH 7.0) and 400 mM NaCl, and loaded onto a Superdex 200 10/300 GL column (GE Healthcare). Fractions from this column were selected based on signals obtained by northern blotting using a spacer 1.04 probe, pooled, and used in subsequent experiments. Native gel northern analysis using a spacer 1.04 probe was performed as described previously (Hale et al., 2009; Wei et al., 2019) to assess the mobility of crRNAs in the purified protein fraction. The protein bands containing the crRNP complex were then excised from the native gel, followed by in-gel trypsin digestion and identification by tandem MS.

The pMtbCsm1-6 plasmid, in which MTB csm2 was replaced with a homologous M. Canettii csm2 with an N-terminal His6 tag, was engineered to express the Csm complex. MTB csm1, csm3, csm4, csm5, and csm6 were codon optimized for expression in E. coli. The codon optimization sequence was previously described by Gruschow et al. (2019). Each csm gene had an independent T7 promoter, high-expression ribosome-binding site, and a T7 terminator, a strategy similar to that described by Ichikawa et al. (2017). The full csm fragment, containing csm1-csm6 together with their respective T7 promoter and T7 terminator, was synthesized by Sangon Biotech (Shanghai, China) and cloned into a pET-28a vector by Seamless Cloning (GeneArt^® Seamless Cloning and Assembly; Thermo Fisher Scientific, Waltham, MA, United States). A pMtbCas6-crRNA plasmid was constructed to express Cas6 and crRNA. The pCDFDuetTM vector, compatible with the pET-28a vector, was used to co-express Cas6. A CRISPR sequence (3R2S) containing three repeats and two spacer 1.01 units from the CRISPR 1 locus (Supplementary Table S1) was synthesized by Sangon Biotech and cloned into a pCDFDuetTM vector together with cas6, and espB and htpG were amplified from the MTB H37Rv genome and cloned into pET22b and pACYADuet-1 vectors separately.

Condon-optimized MtbMBP-csm1 (Rv2823c), Mtbcsm2 (Rv2822c) Mtbcsm3 (Rv2821c), Mtbcsm4 (Rv2820c), Mtbcsm5 (Rv2819c), and Mtbcsm6 (Rv2818c) were cloned into pET-28a vector, separately, to generate individual Csm-expressing plasmids. Individual Csm plasmid (pCsm1, pCsm3, pCsm4, pCsm5, pCsm6) and HtpG was each transformed into E. coli BL21 (DE3) (Novagen) grown at 37°C with 50 μg/ml of kanamycin selection. Protein expression was induced by adding 0.3 mM IPTG (isopropyl β-D-1-galactopyranoside) at OD600 = 0.6 at 16°C overnight. Cells were collected and resuspended in lysis buffer [20 mM Tris-HCl, pH 7.5, 300 mM NaCl, 20 mM imidazole, 5% (v/v) glycerol], and then lysed by sonication. Supernatant by centrifugation was loaded onto HisPur Ni-NTA resin (Thermo Scientific) and the bound protein was washed by washing buffer [20 mM Tris (pH7.5), 300 mM NaCL, 100 mM imidazole, 5% (v/v) glycerol] and eluted by elution buffer containg 20 mM Tris (pH7.5), 300 mM NaCL, 500 mM imidazole, 5% (v/v) glycerol. Individual Csm proteins (MtbCsm1, MtbCsm3, MtbCsm4, MtbCsm5, MtbCsm6) and HtpG was separately further purified by size-exclusion chromatography (Superdex 75 10/300 GL, GE Healthcare) in a buffer containing 20 mM Tris-HCl (pH 7.5), 300 mM NaCl, 5% (v/v) glycerol. Fractions containing the Csm complex and HtpG were identified by Coomassie Brilliant Blue G250 staining SDS-PAGE, flash-frozen with nitrogen and collected at −80°C. MtbCsm2 was purified from inclusion bodies of E. coli. The purification process was carried out following established procedure (Wang et al., 2009; Zhang et al., 2023), including protein inclusion body washing, inclusion body dissolving, urea gradient dialysis and protein refolding. The protein supernatant was pooled for further studies.

The pMtbCsm1-6, pMtbCas6-crRNA, pEspB, and pHtpG plasmids were cotransformed into E. coli BL21 (DE3) cells (Novagen). When the optical density at 600 nm reached 0.6, protein expression was induced with 0.1 mM isopropyl-1-thio-β-D-galactopyranoside for 5 h at 28°C. Cells producing the His6-tagged Csm complex were resuspended in 25 mM Tris-HCl (pH 8.0) containing 300 mM NaCl and lysed using a high-pressure cell disrupter. After centrifugation, the supernatant was applied to a Ni Sepharose column (GE Healthcare) pre-equilibrated with 25 mM Tris-HCl (pH 8.0) containing 0.3 M NaCl and 20 mM imidazole. The bound proteins were eluted with buffer containing different concentrations of imidazole (20, 50, 100, and 500 mM). Target fractions were then collected and analyzed using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Coomassie Brilliant Blue G250 staining.

A biolayer interferometry experiment was carried out at 25°C on Octet RED96e (GE Healthcare) to test the interaction of protein-RNA interaction. To measure and compare the strength of the interactions of MTB EspB and HtpG with different RNAs, 3′-biotinylated RNAs were immobilized on a streptavidin-coated sensor chip, and protein (EspB/HtpG) was injected at creasing concentrations over the surface (500–16,000 nM). Surface plasmon resonance (SPR) experiments were carried out on Biacore 8K (GE Healthcare) at 25°C to test protein-protein interactions. The ligand was covalently immobilized on the dextran surface of a CM5 optical chip. The analyte samples were injected at different concentrations (25–3,200 nM) over the surface at a flow rate of 30 μL/min for 2 min. All experiments were repeated at least thrice.

Plasmids containing the C-termini of csm, htpG, and espB fused with enhanced green fluorescent protein (eGFP) were transformed into BCG and BCGΔhtpG, respectively for constitutive expression system. Samples for flow cytometry analysis were washed with 0.22 μm-filtered 1× phosphate-buffered saline (PBS). A 200 μL aliquot of each sample was transferred to a fluorescence-activated cell sorting analysis tube, and bacterial cells were injected into the flow cytometer within a few minutes after resuspension. A FACScalibur flow cytometer (Becton Dickinson) equipped with a 15-mW, air-cooled argon ion laser was used as the excitation light source (488 nm), and fluorescence (500–549 nm) was detected using a fluorescence detector. The analysis of each sample was stopped after 100,000 cells had been counted.

Competent E. coli BL21 strains containing a csm module (Mtbcsm1-6, Mtbcas6 and crRNA 1.01 spacer) were transformed with 100 ng of either the target or nontarget plasmid. Heat shock transformation was performed, and 800 μL of Luria-Bertani medium was immediately added to the competent cells (50 μL), and incubated with shaking at 200 rpm at 37°C for 45 min. Three sequential 1:10 dilutions was prepared for each transformation, and 20 μL of each dilution was spread onto Luria-Bertani agar plates. Plates were incubated overnight at 37°C, and images of the colonies were captured on a Gel Doc XR (Bio-Rad, Hercules, California, United States) using white light. Assays were performed in triplicate and analyzed using two-way analysis of variance.

The cells were harvested and lysed in PBS. Lysates were incubated with the corresponding antibody or nonspecific immunoglobulin G antibody (negative control) at 4°C. Protein G agarose beads were then added to the mixtures and incubated for 2 h on a rotatory platform at 4°C. Bound proteins were eluted by adding SDS-PAGE treatment buffer.

Samples were separated on 12% SDS-PAGE and then transferred to nitrocellulose membranes. The membranes were blocked with 1x PBS with Tween 20 (PBST) with 5% (w/v) nonfat milk for 1 h at room temperature and then incubated overnight at 4°C with a 1:1000 dilution of the appropriate antiserum. After washing the membranes with PBST three times, bound primary antibodies were detected using an Amersham Imager 600 detection system (GE Healthcare).

DNA oligonucleotide (Fourth spacer from the MTB CRISPR1 locus) probes were end-labeled in a 20 μL kinase reaction containing 10 pmol γ-32P-adenosine triphosphate and 20 U T4 polynucleotide kinase (New England Biolabs, Ipswich, MA, United States) at 37°C for 1 h. Ten micrograms of total RNA were separated on 10% or 15% denaturing acrylamide gels and electroblotted onto Amersham Hybond-XL membranes (GE Healthcare, Waukesha, WI, United States) (Wei et al., 2019). After Pre UV-Cross-linking, the membranes were incubated with labeled probes for 3–6 h in Amersham Rapid-hyb Buffer (GE Healthcare), washed, and exposed to Kodak film over night or for 3 days. All experiments were repeated at least three times (Wei et al., 2019).

Lanes of interest were manually cut out of the preparative gel. After de-staining, dithiothreitol reduction and alkylation of iodoacetamide, trypsin digestion were performed as described elsewhere (Wilm et al., 1996; Fandino et al., 2005). Peptide solutions were extracted three times with a solution containing 60% acetonitrile (ACN) and 0.1% formic acid. Peptides were analyzed by nanoLC-LTQ-MS/MS (ThermoFinnigan, San Jose, CA) equipped with a nano-electrospray ion source, separated with an online reversed-phase LC column (100 μm × 150 mm, 3 μm, Pepmap C18) and gradually eluted with a 0%–100% gradient of buffer B (ACN: water: formic acid, 80:19.9:0.1) in buffer A (ACN: water: formic acid, 2:97.9:0.1) for 70 min at flow rate of 500 nL/min. MS parameters included 2.1 kV electrospray voltage, 300–1,800 m/z range, 90 s exclusion duration and 210°C source temperature.

To identify accurate peptides, LC-MS/MS data was evaluated with Bioworks software (version 3.3.1, Thermo Fisher Sceintific) software and sequences identical to the peptides used to generate the LC-MS/MS data were searched in the National Center for Biotechnology Information database. The search parameters were as follows: species: MTB (H37Rv); digest: trypsin; sites: two missed cleavage sites per peptide allowed; mass error: precursor ion <3 Da and fragment ion <1 Da. The search was set with cysteine carbami-domethylation as the fixed modification and methionine oxidation as the variable modification. The filtering criteria of the MS/MS search result were as follows: 1) The correlation coefficients (Xcorr) of peptides with +1, +2, and +3 charges after complete digestion were greater than 1.90, 2.5, and 3.75, respectively; 2) Sp value of ≥500, Rsp value of ≤5; 3) Each identified protein containing at least two specific peptide segments to support. Results were manually verified and validated.

A summary of results are displayed in Supplementary Table S1. CRISPR/Cas proteins may act together with non-Csm proteins in certain biological functions. For example, T. thermophilus (type III-A) crRNP interacts with RNA and DNA via RNAP to cleave RNA in an RNAP-bound transcription bubble (Liu et al., 2019). Csm proteins Csm1-5 form a complex with mature crRNAs that are involved in the cleavage of nucleic acids (Staals et al., 2014; Tamulaitis et al., 2017). To investigate whether other proteins interact with the MTB crRNP complex (Csm1-5 and crRNA) in vivo, we isolated the crRNP complex using the strategy employed by Wei et al. (2019). Northern blotting with an anti-1.04 probe (fourth spacer from the CRISPR1 locus) was used to locate the purified mature crRNP complex on the native gels (Figures 1A, B). The band corresponding to the crRNP complex was gel digested, and the tryptic peptides of the complexes was identified by the nano LC-TLQ-MS/MS. The significant peptide including duplicate sequences from separate MS/MS spectra was detected and shown the sequences of the protein coverage regions. Tryptic digestion provided a high number of cleavages and produced long peptides resulting in a high percentage (32%∼75%) of sequence coverage. The MS/MS search included seven proteins (Csm 2-6, Chaperone htpG, and ESX-1 secretion-associated protein EspB) were involved in the MTB H37Rv strain; however, Csm1 was lost (Figure 1C; Supplementary Table S1). Although Csm4 protein was identified from only two peptides, the data was reliable since the peptide spectra was manually validated; therefore, protein identification was effective and sufficient.

Csm6 did not interact with the crRNP complex; however, its RNase activity was stimulated by short cyclic oligoadenylate signaling molecules generated by the crRNP complex, Csm1, upon target RNA recognition. Csm1 was not identified, possibly owing to partial degradation during isolation.

To investigate whether EspB and HtpG could bind crRNA and participate in the crRNP complex, we first explored whether HtpG and EspB interacted with crRNA using BLI. In brief, the results indicated that EspB could bind to MTB crRNA 1.01 (Figure 2B). Relevant kinetic paramenters calculated by BLI was listed in Supplementary Table S3.

Mature crRNAs in MTB are of the same length (∼71 nt), and their structure (intact spacer flanked by 3′ and 5′ repeat sequence tags, with the 3′ tag forming a hairpin structure) closely resembles that of type I system mature crRNAs. The sequence or structure of the 3′ hairpin B in the repeat RNA of MTB crRNAs is important for MTBCas6 recognition or cleavage of the substrate RNA. The accuracy of mature crRNA cleavage in MTB depends on the presence of the 3′ hairpin B structure and its base sequence (Zhang et al., 2021). We investigated whether the binding ability of EspB to crRNA 1.01 was influenced by the 3′ hairpin B structure and its base sequence by mutating GAGA to UCUC within the 3′ hairpin B [Seen the secondary structure in Figure 2A (Zuker, 2003)]. The results showed that EspB could bind repeat RNA sequence and binding ability was not affected by mutated sequences (Figure 2B). However, Figure 2C showed that HtpG could not directly interact with repeat RNA or repeat mutant RNA. Consistent with this result, Yosef et al. (2011) only found that HtpG was essential for the function level of CRISPR system.

We first investigated whether HtpG and/or EspB interacted with the crRNP complex in vivo. We simultaneously expressed various proteins in E. coli BL21 (DE3) competent cells using the following plasmids: 1) codon-optimized MTB csm genes: csm1-His-tagged mca2-csm3-csm4-csm5-csm6 genes [MTBCsm2 was replaced by its homologue from M. canettii (Supply and Brosch, 2017) (McaCsm2; 71% amino acid identity), as the expression and purification of MTBCsm2 was problematic]; 2) cas6 repeat-spacer sequence; 3) espB gene; and 4) htpG gene (Figure 3A). Cell lysates were purified with HisPur Ni-NTA resin targeting the 6-His tag on Mca2. Proteins were eluted at different concentrations of imidazole (50, 100, and 500 mM), and six bands (corresponding to different elution concentrations) were observed on SDS-PAGE. Nano LC-TLQ-MS/MS also identified two non-Cas system proteins, namely, EspB and HtpG, indicating that EspB interacted with the MTB crRNP complex in vitro (Supplementary Table S1).

Subsequent SPR results confirmed that EspB could interact with Csm1, Csm2, Csm3, and Csm6; however, we were unable to detect any interaction between EspB and Csm4 in vitro (Figure 3B). As Csm5 is unstable when released from the complex in vitro, we performed a co-immunoprecipitation assay in M. bovis BCG, whose CRISPR/Cas system is similar to that of MTB to validate its interaction with EspB. Western blotting demonstrated that EspB interacted with Csm5 in vivo (Figure 3C).

Even though HtpG could not interact with the crRNP complex, SPR experiments showed that HtpG interacted with EspB and Csm3 in vitro (Figure 3D; Supplementary Figure S1). The kinetic paramenter analysis was listed in Supplementary Table S4.

As HtpG is essential for maintaining functional levels of Cas3 in E. coli (type I CRISPR/Cas system) (Yosef et al., 2011), we hypothesized that its expression might also affect the accumulation of Csm proteins in Mycobacteria. We transformed plasmids carrying csm1-egfp), csm2-egfp, csm3-egfp, csm4-egfp, csm5-egfp, and csm6-egfp into wild-type M. bovis BCG and BCGΔhtpG, and eGFP intensities were measured using flow cytometry to determine the accumulation of Csm1-6 proteins. The results showed that Csm2 and Csm3 levels increased while Csm1, Csm4 and Csm6 decreased in both BCGΔhtpG relative to the wild-type control. These results indicate that htpG can influence the accumulation of Csm proteins (Figure 4).

However, when we transformed the espB-egfp plasmid into wild-type BCG and BCGΔhtpG, we obtained colonies of BCGΔhtpG overexpressing espB-egfp but not wild-type BCG overexpressing espB-egfp. These results indicate that overexpression of EspB can lead to the death of wild-type M. bovis BCG, while loss of htpG (BCGΔhtpG) confers a higher survival capability against EspB accumulation.

The MTB CRISPR/Cas system (type III-A), the crRNP complex, provides resistance against invasive phages and plasmids. Since EspB is involved in the MTB EspB-crRNP complex and HtpG overexpression affects the accumulation of Csm proteins, we investigated whether EspB and HtpG overexpression also influences the efficiency of MTB CRISPR/Cas anti-plasmid immunity. We simultaneously transformed foreign DNA from four plasmids [two expression plasmids containing csm1-His-tagged mca2-csm3-csm4-csm5-csm6 and cas6 + crRNA 1.01 (the first spacer in the MTB CRISPR1 locus), an artificial foreign plasmid designed to target crRNA 1.01, and a plasmid containing non-target crRNA 1.01 as a control] into E. coli BL21 (DE3) competent cells. The transformed plasmids containing non-target crRNA 1.01 were unaffected by the MTB Csm-crRNA complex; however, plasmids targeting crRNA 1.01 could not be transformed into E. coli BL21 (DE3) cells and no colonies were present on the culture plates. These results indicate that the reconstituted crRNA-Csm complex is active in defense and can provide immunity against certain foreign nucleic acids.

Next, we overexpressed EspB and HtpG in our recombinant E. coli BL21 (DE3) MTB CRISPR/Cas complex-containing cell line. When the artificial foreign plasmid targeting crRNA 1.01 was transformed into these cells (Figure 5A), EspB overexpression did not affect the efficiency of plasmid transformation. However, HtpG overexpression significantly inhibited the ability of the MTB crRNP complex to protect against foreign DNA (Figures 5B, C). To check the HtpG overexpressed in E. coli BL21(DE3) MTB CRISPR/Cas system, HtpG was identified and confirmed by LC-MS/MS (Supplementary Table S2). These findings done in this work may provide a useful guideline for clinical application and has important practical values.