Functional impact of a conservative missense mutation in the Mycobacterium bovis BCG Moreau heat shock response regulator hrcA gene

Abstract

Background

The genus Mycobacterium includes important human pathogens such as M. tuberculosis, the etiological agent of human tuberculosis. A critical factor in its pathogenic success is its ability to adapt efficiently to various stress conditions, including heat shock. This response mitigates the accumulation of misfolded proteins by upregulating heat shock proteins, primarily chaperones, through complex regulatory networks, including transcriptional repressors such as HspR and HrcA. The vaccine strain M. bovis BCG Moreau genotypically differs from M. tuberculosis in several aspects, including a conservative missense mutation in hrcA, substituting valine with leucine at residue 67 near its DNA-binding domain. This genetic variation could impact stress adaptation and vaccine-associated traits through altered regulation of heat shock proteins.

Results

Structural modeling of BCG Moreau’s HrcA revealed local rearrangements accommodating the larger leucine side chain, with maintained overall protein architecture. Electrophoretic mobility shift assays (EMSA) demonstrated reduced binding of Moreau’s HrcA to CIRCE elements upstream of groES and groEL2 compared to the BCG Pasteur strain. RT-qPCR revealed unexpected patterns, with greater transcriptional induction of the groES operon and groEL2 in the Pasteur strain under heat shock conditions. Western blotting analysis corroborated these findings, showing increased GroEL2 levels in both intracellular and secreted fractions of BCG Pasteur. These observations suggest the involvement of multiple regulatory layers beyond direct HrcA-mediated repression.

Conclusions

Our findings suggest that M. bovis BCG Moreau may employ distinct mechanisms to modulate its heat shock response, with the V67L mutation in HrcA potentially contributing to these differences. The lower baseline expression of immunogenic heat shock proteins in BCG Moreau, combined with its distinct regulatory response to stress, may explain its reported characteristics of reduced adverse reactions while maintaining protective immunity. These results provide new insights into strain-specific variations in heat shock protein regulation and their potential impact on vaccine performance.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12866-026-05246-7.

Background

Species within the group of mycobacteria exhibit diverse life strategies adapted to their specific niches, ranging from free-living organisms that can cause opportunistic infections under certain conditions to pathogens, such as Mycobacterium tuberculosis (Mtb), the etiological agent of human tuberculosis [1]. The success of Mtb as a pathogen is attributed to various characteristics, including its lipid-rich envelope composition, its ability to enter a dormant state for prolonged periods, and its capacity to respond to different types of stress [2, 3].

Upon entering macrophages, the bacillus faces various host defense mechanisms, such as exposure to reactive oxygen and nitrogen species, hypoxia, and nutritional stress [4]. Under these conditions, an upregulation of heat shock proteins (HSPs) is observed [5–7]. These proteins are highly conserved across all organisms, indicating their fundamental role in cellular processes. The term originates from the observation that their expression increases under heat stress conditions. However, it is now well established that various other stressors can also induce the expression of certain HSPs, such as hypoxia, heavy metals, nutritional deprivation, oxidative agents, ethanol, and infections [8]. Heat shock and other forms of stress lead to protein aggregation and misfolding. These improperly folded proteins/peptides act as triggers for the activation of specific stress response pathways. For example, it is known that the HspR repressor binds tightly to Hsp70 [9], forming a heterodimer that constitutes the functional repressor, thereby preventing the transcription of its regulon. In the presence of misfolded proteins, Hsp70 is sequestered into the cytoplasm to perform its chaperone function, disrupting the functional heterodimer and releasing the HspR regulon from inhibition [10]. Importantly, HSPs are expressed at significant levels under normal physiological conditions in both eukaryotes and prokaryotes.

HSPs are commonly classified into families based on their molecular weight [3]. At the molecular level, their functions include modulation of receptor-mediated protein responses, signal transduction, intracellular transport and peptide translocation, chaperone activity, prevention of denatured protein aggregation in vitro (e.g., Hsp60), folding of nascent proteins (e.g., Hsp60 and Hsp70), facilitation of protein degradation as a co-factor in proteolytic systems (e.g., Hsp60), and autoregulation of heat shock responses (e.g., Hsp70) [11].

The increased expression of HSPs can be associated with both positive and negative gene regulation [12]. Positive regulation often involves alternative sigma factors [13, 14], while negative regulation relies on transcriptional repressors such as HspR and HrcA [10, 15–17]. For instance, in Bacillus subtilis, regulation of the 70 kDa HSP, or DnaK, is mediated by enhanced transcription due to the removal of a transcriptional repressor [18]. Additionally, positive regulation of DnaK through an alternative sigma factor is observed in Escherichia coli [19]. In B. subtilis, a single repressor, HrcA, co-regulates the dnaK (Hsp70) gene alongside the groE (Hsp60) family, which includes groEL1, groEL2 (Hsp65), and groES (Hsp10) [20]. In contrast, the loss of HspR, a regulator of Hsp70, did not show any detectable effect on Hsp60 expression in Mtb [21]. This observation suggests the involvement of a second repressor regulating this set of HSPs in Mtb, in this case, HrcA.

The study of the heat shock response serves as a reliable and convenient laboratory model for investigating transcriptional regulation and is also a crucial element in the pathogenesis of Mtb. HSPs are induced following the phagocytosis of the bacteria by host cells [22], likely contributing to the microorganism’s survival within the phagosomal environment. Moreover, it has been demonstrated in Mtb that partial disruption of the heat shock response significantly impacts virulence, impairing the bacteria’s ability to establish chronic infection [23]. However, HSPs are also utilized by the host cell as source of peptides for presentation to the class I MHC complex and as signals for the immune system [16]. Due to this immune role of HSPs, their increased expression results in reduced survival of Mtb during prolonged infection [9]. Therefore, the regulation of the expression of these proteins must be finely controlled, allowing the pathogen to mount an appropriate stress response while avoiding the elicitation of an immune response in the host that could eliminate the bacillus.

Vaccination with the Bacillus Calmette-Guérin (BCG) strain remains the only prophylactic method against human tuberculosis, effective against extra-pulmonary cases and in children. It was developed in the early 20th century by the in vitro attenuation of a clinical strain of M. bovis [24]. After its distribution to health institutes worldwide, genotypic and phenotypic diversification occurred due to different cultivation methods, lack of effective storage techniques, and absence of seed lots [25]. Today, it is known that there is a family of BCG strains, with the one used in Brazil until 2017 being BCG Moreau [26]. Following the sequencing of its genome and comparison with other strains, such as BCG Pasteur (used as a reference), several differences were observed, including large regions of deletion/duplication as well as small polymorphisms. One such mutation, a non-synonymous one, results in the substitution of valine with leucine in the HrcA protein, which regulates the transcriptional repression of certain heat shock-induced genes [17]. Thus, the functional impact of this mutation in the Moreau HrcA regulator was analyzed. Structural analyses were performed, along with electrophoretic mobility shift assays (EMSA). Furthermore, to initiate the study of the heat shock response in BCG Moreau and to investigate potential impacts of the HrcA mutation on this broader phenotype, we analyzed the transcription of several genes from the HrcA regulon following heat shock using RT-qPCR, and detection of GroEL2 expression by western blotting. These results are important for understanding the regulation of HSPs expression, which in the context of a vaccine is critical for the survival of the bacillus as well as for the production of antigens, both essential for inducing an appropriate host immune response.

Methods

Computational modelling and structural analysis

The three-dimensional structure of HrcA from Mycobacterium tuberculosis H37Rv (Mtb H37Rv) was predicted using AlphaFold 3 [27] through its web interface (https://golgi.sandbox.google.com/). The modeling process employed default parameters, including five independent modeling cycles. The generated models were evaluated based on their predicted Local Distance Difference Test (pLDDT) scores, which range from 0 to 100. The model exhibiting the highest pLDDT score was selected for subsequent analyses. It should be noted that only the HrcA protein structure was modeled in this study; protein-DNA docking was not performed due to the absence of experimentally determined HrcA-DNA complex structures suitable for comparative modeling.

To investigate the structural implications of the V67L mutation observed in the Moreau strain, the variant was constructed in silico using UCSF ChimeraX [28]. The mutation was introduced using the ‘swapaa’ command, with rotamer selection guided by the Dunbrack backbone-dependent rotamer library [29]. The V67L mutation localizes to the HrcA dimerization domain near the DNA-binding interface (residues 50–80).

Both wild-type and mutant structures underwent energy minimization using the AMBER 22 package implementing the ff19SB force field. The minimization protocol consisted of two sequential phases: (1) an initial 200-step minimization (100 steps steepest descent followed by 100 steps conjugate gradient) with harmonic restraints (5 kcal mol⁻¹ Å⁻²) applied to all non-hydrogen atoms, and (2) an unrestrained 200-step minimization (100 steps each of steepest descent and conjugate gradient).

The structural quality of both models, before and after energy minimization, was evaluated using the Structure Assessment platform of SWISS-MODEL [30]. The assessment included Ramachandran plot analysis to verify backbone geometry through φ/ψ angle distributions, identifying favorable, allowed, and disallowed regions. Stereochemical quality was assessed using MolProbity, which examined various structural parameters including bond lengths and angles, peptide bond planarity, side-chain rotamer conformations, and atomic clashes.

The evolutionary conservation of residue position 67 was analyzed using the ConSurf server, which estimates the evolutionary conservation of amino acid positions based on phylogenetic relationships between homologous sequences.

Structural visualization and analysis were performed using UCSF ChimeraX [28]. Two-dimensional protein-ligand interaction maps were generated using Discovery Studio Visualizer 2021.

Cloning and recombinant protein production

The coding sequences of hrcA from M. bovis BCG Moreau and Pasteur strains were cloned into the pET28a expression vector. Genomic DNA from these strains served as the PCR template, with Platinum SuperFi DNA polymerase (Invitrogen) and specific primers (forward: CATGCCATGGGAAGCGCCGACGAGCG and reverse: CGAATTCCGAGCACCCAGGACGTCGCC), adding an NcoI restriction site at the amplicon 5’ end and an EcoRI site at the 3’ end. After proper digestion and vector and insert ligation, resulting constructs were transformed into E. coli TOP10 cells, and recombinant plasmids were verified by colony PCR and Sanger sequencing, prior to E. coli BL21(DE3) transformation. Recombinant protein expression was induced with 1 mM isopropyl-β-D-thiogalactopyranoside (IPTG) for 3 h at 37 °C, 200 rpm in Luria-Bertani medium containing 25 µg/mL kanamycin.

After lysis using a BeadBeater apparatus, cell lysate was clarified by centrifugation and used for recombinant protein purification. This was achieved via Immobilized Metal Affinity Chromatography (IMAC) using a HisTrap HP 1 mL column (GE Healthcare) loaded with Ni²⁺ ions. Elution was carried out with gradient steps of 10, 20, 30, 40, and 100% of the elution buffer containing 500 mM imidazole. Purity was assessed by 12% SDS-PAGE stained with Coomassie Brilliant Blue R-250, and fractions containing the target protein were pooled, dialyzed against PBS, and utilized for immunization of BALB/c mice. Serum from immunized mice was analyzed by western blotting against the recombinant protein to confirm antibody production. Reactive sera were combined to generate a polyclonal anti-HrcA antibody (α-HrcA). The experimental procedures were approved by the Animal Care and Use Committee at IOC, Fiocruz (CEUA/IOC License LW-46/11).

Electrophoretic mobility shift assay

To prepare probes for EMSA, 200 and 167 bp fragments containing upstream regions of groEL2 and groES, respectively, were amplified using the following primers pairs: groEL2 (for: GCCGAACATACTCACCCGGAT; rev: ACGGACGGCTCAAGTTGTCCT); groES (for: ACCCGGACATTGCACCTGG; rev: GGTGATTCGGCATTCGTCC). Both reverse primers were biotinylated at the 5′ end. PCR products were purified using the Wizard SV Gel and PCR Clean-Up System kit (Promega), and 2 ng of purified DNA was used in binding reactions with increasing concentrations of recombinant, purified HrcA protein, both from Pasteur and Moreau strains. Each binding reaction, in a final volume of 20 µL, contained 2 µL of 10X binding buffer, 1 µL of 50% glycerol, and 1 µL of poly dIdC (1 µg/µL) from the LightShift™ EMSA Optimization & Control Kit (Thermo). The protein was added last, and reactions were incubated for 30 min at room temperature. Following incubation, a loading buffer with glycerol was added, and the samples were loaded onto 5% polyacrylamide gels in 0.5X TBE buffer. Electrophoresis was performed at 100 V for 30 min.

Post-electrophoresis, DNA–protein complexes were transferred to positively charged Hybond-N+ nylon membranes (GE) for 1 h at 180 mA in 0.5X TBE buffer using a Mini-Transblot apparatus (BioRad). DNA was cross-linked to the membrane by UV exposure for 5 min on each side. The biotin-labeled PCR product was visualized using the LightShift Chemiluminescent EMSA Kit (Thermo), following the manufacturer’s protocol. Membranes were then exposed to X-ray film for signal detection and subsequently developed.

Real-time PCR

RT-qPCR assays were conducted to compare transcription levels of seven genes involved in heat shock response in BCG Moreau and Pasteur. Both strains were cultivated at 37 °C and subjected to a heat shock condition when incubated at 42 °C. Control cultures were maintained at 37 °C Samples were collected at different time points after increasing temperature.

Total RNA was isolated from the bacterial samples using TRIzol (Thermo Fisher Scientific) and the RiboPure™ Bacteria kit (Thermo Fisher Scientific), following the manufacturer’s instructions. Briefly, bacterial cells were disrupted mechanically with zirconium beads using a bead-beater (Biospec Products Inc.). Chloroform extraction was performed, and the RNA-containing aqueous phase was collected and further purified with the RiboPure™ Bacteria kit. To eliminate any DNA contamination, Turbo-DNase treatment (Ambion-Life, Lithuania) was applied according to the manufacturer’s protocol, and the RNA was then eluted in RNase-free water.

For cDNA synthesis, 600 ng of total RNA was reverse transcribed using random primers and the SuperScript III First-Strand Synthesis SuperMix (Thermo Fisher Scientific). The synthesized cDNA was then used as a template in RT-qPCR analysis.

Primers specific to the target genes (dnaK, ppiA, hrcA, groES, groEL1, groEL2, dnaJ2 and hspR) were designed (Table 1), with sigA serving as the internal normalization control. PCR reactions were performed using a 7500 Real-Time PCR System (Applied Biosystems) in 96-well plates under the following conditions: initial denaturation at 95 °C for 10 min (1 cycle); followed by 40 cycles of 95 °C for 30 s, 60 °C for 30 s, and 72 °C for 30 s. After each assay, a dissociation curve was generated to verify primer specificity. Gene expression levels were calculated using relative quantification with the comparative Ct method (ΔΔCt) as described previously [31]. Relative mRNA levels were normalized to sigA mRNA [32, 33], and all qPCR reactions were performed in triplicate. Fold changes were calculated, setting the time zero expression level as 1.0.

Table 1.

Primer list for RT-qPCR assay

Primer name		Sequence (5´-3´)
locus	Direction
dnaK	For	ATGGCTCGTGCCGGTCGGGAT
	Rev	CGCACGGTGCGATCGACGTT
ppiA	For	ATGGCAGACTGTGATTCCGTGA
	Rev	TCGTAGAACGGGCCGGACGG
hrcA	For	ATGGGAAGCGCCGACGAGCG
	Rev	AGGCGTCCGGAACTGGTGTG
groES	For	GTGGCGAAGGTGAACATCAAGC
	Rev	GTCACCCTCCGCAACGTCCA
groEL1	For	ATGAGCAAGCTGATCGAATACGA
	Rev	GTCTTCAAACGGATCTTCCAGC
groEL2	For	ATGGCCAAGACAATTGCGTACG
	Rev	CTTCTCGTACGGATCCTGGAG
dnaJ2	For	GTGGCACGCGATTATTACGGGC
	Rev	CCAGGTCGACGATGCGACGTT
hspR	For	ATGGCGAAGAACCCAAAGGACG
	Rev	TCCTGCGAGAGGTGCTGCAC

Western blotting

Samples from the heat shock assay used for the RT-qPCR were processed for western blotting analysis. The culture filtrate was prepared by separating the culture supernatants from bacterial cells, followed by centrifugation at 2,500 × g for 10 min at 4 °C. The supernatant was then filtered through a 0.22 μm low protein-binding membrane. Then, total protein was precipitated using 17% (v/v) trichloroacetic acid (TCA) and subsequently washed with cold acetone. The total soluble intracellular protein fraction was obtained following cell lysis and subsequent clarification by centrifugation. Protein contents were measured using a NanoDrop spectrophotometer and 20 µg of protein was then subjected to separation by 15% SDS-PAGE and subsequently transferred to a nitrocellulose membrane (Hybond C, GE) using a Bio-Rad Mini Trans-Blot system at a constant voltage of 100 V for 1 h. The quality of protein transfer was verified through reversible Mem Code staining (Pierce). The membrane was then blocked overnight (16 h) at 4 °C in TBS-T buffer (Tris-buffered saline with 0.05% (v/v) Tween 20) supplemented with 10% (w/v) skim milk. After blocking, the membrane was rinsed with TBS-T and incubated for 2 h with an α-GroEL2 polyclonal antibody diluted 1:1,000, followed by a 1-hour incubation with an HRP-conjugated goat anti-mouse IgG secondary antibody at a dilution of 1:10,000. Each antibody incubation step was followed by three washes in TBS-T, with additional final washes in TBS. All antibody incubations were conducted in TBS buffer containing 5% (w/v) skim milk. The blot was visualized using the SuperSignal detection kit (Pierce), following the protocol provided by the manufacturer.

Images from MemCode stained membranes and blots were captured using a GS-800™ auto-calibrating imaging densitometer (Bio-Rad). Quantification of GroEL2 reactive bands was performed by calculating their relative ‘volume’, which represents the cumulative pixel intensity within each band. To adjust for minor variations, each band volume was normalized to the total lane density from the corresponding MemCode image. To assess heat shock induction, fold change was determined by comparing the normalized density values of post-induction samples to those of the respective control samples without temperature increment.

Results

Structural analysis of the V67L variation between Mtb H37Rv and BCG Moreau HrcA regulators

Comparative genomic analyses of M. bovis BCG Moreau, the Pasteur strain (a reference within the BCG family), and Mtb H37Rv revealed several genetic differences. Among these, a notable transversion substitution (G to C) occurs at the first nucleotide of hrcA codon 67, resulting in the replacement of a valine with a leucine, thus a conservative missense mutation (Fig. 1).

Fig. 1.

The hrcA gene of M. bovis BCG Moreau harbors a conservative missense mutation when compared to its homolog in M. tuberculosis H37Rv. Alignments of the hrcA nucleotide sequences (upper panel) and the predicted amino acid sequences (lower panel) were performed using the online tool Clustal Omega. These alignments reveal that the nucleotide substitution at position 199 (the first nucleotide of codon 67) results in the replacement of a valine residue, as found in M. tuberculosis H37Rv, with a leucine residue in M. bovis BCG Moreau. (*) Conserved sites and (:) sites with conservative substitutions

The AlphaFold-generated structural model of Mtb H37Rv HrcA demonstrated robust quality metrics, with a predicted Template Modeling score (pTM) of 0.77, indicating high confidence in the global fold prediction (Fig. 2A). Per-residue confidence analysis through predicted Local Distance Difference Test (pLDDT) scores revealed that the majority of the protein structure exhibits scores above 70, with the functionally significant V67 residue showing a particularly high pLDDT score of 95. As expected for heat-responsive transcriptional regulators, surface-exposed loops displayed moderate (pLDDT 50–70) to lower (< 50) confidence scores, consistent with their inherent conformational flexibility.

Fig. 2.

Structural analysis and evolutionary conservation of the V67L variation in Mycobacterium HrcA. A Three-dimensional structural model of Mtb H37Rv HrcA generated by AlphaFold 3, colored according to pLDDT confidence scores (> 90: blue; 70–90: green; 50–70: yellow; <50: red). The position of residue Val67 is highlighted in ball and sticks representation and indicated by a red arrow. B ConSurf analysis showing evolutionary conservation patterns across 150 bacterial species, with position 67 displaying moderate conservation (grade 6) and strong preference for β-branched hydrophobic residues. The position of residue Val67 is highlighted in ball and sticks representation and indicated by a red arrow. C Two-dimensional interaction maps depicting the molecular networks surrounding position 67 in both Mtb H37Rv (left) and Moreau (right) HrcA variants, highlighting key stabilizing interactions and variant-specific contacts. D Superimposed structural comparison of the local environment around position 67, showing the conformational adaptation to accommodate the V67L substitution. The Mtb H37Rv (blue) and Moreau (red) variants demonstrate subtle differences in aromatic interactions with Phe19 and local packing arrangements while maintaining core stabilizing contacts. Yellow dashed lines indicate the hydrogen bond between the backbone of residue 67 and Thr58

Comprehensive validation of both Mtb H37Rv and BCG Moreau HrcA models demonstrated significant improvements following energy minimization (Tables 2 and 3). The overall model quality, assessed by MolProbity score, showed marked enhancement from 1.45 to 0.50 for Mtb H37Rv and from 1.35 to 0.50 for the Moreau variant. Notably, the minimization process completely resolved atomic clashes, with clash scores reducing from 6.76 to 5.22 to 0.00 in both structures. Ramachandran plot analysis revealed excellent backbone geometry, with the percentage of residues in favored regions improving from 97.65% to 98.12% post-minimization, accompanied by complete elimination of Ramachandran outliers. Side-chain geometry remained consistently good, with rotamer outliers staying stable at approximately 0.37–0.38% across all models. Both structures exhibited ideal covalent geometry, with no C-beta deviations or bad bonds detected, while maintaining acceptably low levels of suboptimal bond angles.

Table 2.

Experimental design framework: bacterial strains, genetic constructs, and analytical approaches

Bacterial strain	contructs	Purpose/endpoints measured
E. coli BL21(DE3)	pET28a_hrcA(Moreau)	Recombinant protein production for EMSA assay
	pET28a_hrcA(Pasteur)
M. bovis BCG Moreau	---	RNA extraction for qRT-PCR / Protein extraction for western blotting
M. bovis BCG Pasteur

Table 3.

Structural validation metrics for HrcA models

Parameter	Mtb H37Rv	Mtb H37Rv_min	Moreau	Moreau_min
MolProbity Score	145	0.50	1.35	0.50
Clash Score	6.76	0.00	5.22	0.00
Ramachandran Favoured (%)	97.65	98.12	97.65	98.12
Ramachandran Outliers (%)	0.59	0.00	0.59	0.0
Rotamer Outliers (%)	0.37	0.38	0.37	0.38
C-Beta Deviations	0	0	0	0
Bad Bonds	0/2588	0/2518	0/2587	0/2517
Bad Angles	8/3509	13/3425	7/3508	7/3424

Evolutionary analysis of position 67 using ConSurf revealed moderate conservation across bacterial species, with a normalized conservation score of -0.198, corresponding to grade 6 on the ConSurf 9-grade conservation scale (Fig. 2B). Analysis of 150 aligned sequences demonstrated a strong preference for β-branched hydrophobic amino acids, with isoleucine appearing as the predominant residue (51%), closely followed by valine (48%). Notably, leucine was observed in less than 1% of sequences, exclusively in Corynebacterium matruchotii ATCC 14,266 (Uniprot E0DGG1). This stark preference for β-branched amino acids suggests specific structural constraints at this buried position, while the presence of leucine in C. matruchotii represents a notable deviation that maintains hydrophobic character despite different side chain geometry.

Detailed analysis of molecular interaction networks surrounding position 67 revealed both conserved and variant-specific patterns between Mtb H37Rv and Moreau HrcA structures (Fig. 2C). Both variants maintain several key stabilizing interactions, including a hydrogen bond with Thr58, alkyl interactions with Pro25 and Pro60, and van der Waals contacts with Ile57, Arg66, Pro68, and Thr69. However, subtle differences were observed in their interaction networks. The Mtb H37Rv HrcA exhibits Pi-Alkyl interactions with Phe19 and an additional van der Waals contact with Ile26, while the Moreau HrcA displays Pi-sigma interactions with Phe19. These variations in aromatic interactions likely reflect structural adaptations accommodating the V67L substitution, where the marginally larger and more flexible leucine side chain in Moreau HrcA influences local packing arrangements (Fig. 2D).

Functional impact on the DNA binding ability

In order to study the impact of this mutation on protein-DNA interactions, the hrcA genes from M. bovis BCG Moreau and BCG Pasteur (which has an identical sequence to the M. tuberculosis H37Rv gene) were cloned into the pET28a vector for heterologous expression in E. coli. The results showed that the recombinant proteins were predominantly expressed in the soluble form.

For the gel shift assays (EMSA), the upstream regions of the groEL2 and groES genes were used as target DNA. These assays revealed subtle differences between the proteins. When 100 pmol were used, HrcA from Moreau produced a partial shift, whereas the same concentration of the Pasteur protein resulted in a complete shift (Fig. 3).

Fig. 3.

HrcA from M. bovis BCG Moreau exhibits a subtle difference in its ability to bind CIRCE elements located in the upstream regions of groEL2 and groES compared to the HrcA from the Pasteur strain. Electrophoretic Mobility Shift Assays (EMSAs) were performed using biotin-labeled, PCR amplified upstream sequences of groEL2 and groES with 0, 10, 100, and 200 pmol of recombinant HrcA (rHrcA) from M. bovis BCG Moreau or Pasteur. Bovine Serum Albumin (BSA) was used as a negative control, ensuring no binding to the DNA fragments. Additionally, a control assay was performed using rHrcA from Pasteur at the lowest (0 pmol) and highest (200 pmol) quantities, incubated with the upstream sequence of rv3406, which is known to contain a binding site for the regulator Rv3405c

Expression of heat shock-regulated genes after thermal stress induction

qRT-PCR analysis was performed on BCG Moreau and Pasteur cultures exposed to 42 °C for 15, 30, and 60 min, compared to control cultures maintained at 37 °C (Fig. 4). The results revealed significant differences in the transcriptional induction levels of genes encoding GroEL family proteins (groES, groEL1, and groEL2) following heat shock. A more pronounced transcriptional response of these genes was observed in BCG Pasteur compared to BCG Moreau. Similar induction patterns were noted for groES and groEL1 in BCG Pasteur, consistent with both genes being under the control of the same promoter. Additionally, the difference in induction between the strains was less pronounced for this operon compared to the transcription of groEL2. This may be attributed to the fact that the operon is regulated by both HrcA and HspR (the latter showing no predicted structural differences between the strains analyzed), whereas groEL2 is regulated solely by HrcA [17]. Although a CIRCE sequence is present upstream of hrcA, no differences were observed in the regulation of this gene or dnaJ2 (located downstream of hrcA) between Pasteur and Moreau.

Fig. 4.

Analysis of the heat shock response genes transcriptional induction in M. bovis BCG Moreau and Pasteur strains. RT-qPCR assays were performed on cultures grown at 37 °C and exposed to a temperature increase (42 °C) for various durations (0, 15, 30, and 60 min), targeting the genes dnaK, groEL1, groES, groEL2, hrcA, dnaJ2, hspR, and ppiA. Fold-change values were calculated relative to the 0-minute time point, which was set as 1.0 for analysis. Statistical analysis was conducted using a one-way ANOVA test. Asterisks without a connecting bar indicate comparisons between each time point and time 0, while bars represent other pairwise comparisons. Significance levels were denoted as follows: (*) p < 0.05; (**) p < 0.01; (***) p < 0.001. Data are presented as the mean ± SD of three independent experiments

Western blotting analysis using anti-GroEL2 antibody on proteins from the culture supernatant and intracellular fraction of M. bovis BCG Moreau and BCG Pasteur, subjected to heat shock (42 °C) for the three time points described in the previous assay, revealed an expression pattern consistent with the trends observed in qRT-PCR experiments (Fig. 5).

Fig. 5.

The transcriptional induction differences observed in heat shock response genes are less pronounced at the protein level when comparing M. bovis BCG Moreau and Pasteur strains. Total intracellular proteins and culture filtrate proteins from bacteria exposed to elevated temperatures (42 °C) were extracted, separated by SDS-PAGE, and transferred onto nitrocellulose membranes. Protein complexity was assessed using MemCode reversible stain (top images), and GroEL2 levels were analyzed by western blotting using a specific polyclonal antibody (bottom images). The GroEL2 signals were quantified and normalized to the total lane pixel intensity from the MemCode staining. The results are shown in the lower graphs. Fold-change values were calculated relative to the 0-minute time point, set as 1.0 for analysis. Statistical analysis was performed using a one-way ANOVA test. Asterisks without a connecting bar indicate comparisons between each time point and time 0, while bars represent other pairwise comparisons. Significance levels were denoted as follows: (*) p < 0.05; (**) p < 0.01; (***) p < 0.001. Data are presented as the mean ± SD of three independent experiments

Discussion

The present study investigates the potential impact of a conservative missense mutation in the heat shock response regulator HrcA in M. bovis BCG Moreau, in comparison with the isoform present in BCG Pasteur (which is sequence-identical to that of Mtb). To this end, we initially assessed the possible effects of this mutation through in silico structural analyses. Both models display a similar overall structure, with minor local adjustments that accommodate the difference in side-chain volume between valine (in Pasteur and Mtb) and leucine (in Moreau). The maintenance of key stabilizing interactions alongside subtle variations in the interaction networks suggests that while both variants preserve similar overall structural characteristics, they may exhibit fine differences in local protein dynamics or stability. The V67L substitution represents a conservative change in amino acid properties, with leucine being slightly larger and more flexible than valine. Given position 67’s location within a buried hydrophobic core, as indicated by ConSurf analysis, this substitution could subtly influence local protein packing. The observed differences in aromatic interactions with Phe19 and the presence of variant-specific contacts suggest that the protein structure has undergone minor reorganization to accommodate the larger leucine side chain, potentially affecting local structural stability.

From a functional perspective, the conservative nature of the V67L substitution, maintaining hydrophobic character without altering charge or polarity, suggests that primary DNA-binding and transcriptional repression functions are likely preserved. However, the altered interaction network and potential changes in local protein dynamics could indirectly modulate HrcA function through subtle effects on protein-protein interactions or allosteric regulation. The position’s moderate evolutionary conservation (ConSurf grade 6) and strong preference for β-branched amino acids across bacterial species suggest some flexibility in amino acid identity while maintaining specific structural constraints. These structural adaptations, though modest, could potentially contribute to the distinct properties of Mtb H37Rv and Moreau strains through fine-tuned modifications in protein behavior and regulatory function.

The HrcA protein acts as a transcriptional repressor that binds to upstream sequences of certain genes, thereby repressing gene transcription [17]. Microarray experiments with Mtb mutants carrying a deleted hrcA gene demonstrated increased transcription of the HSP60 family genes, groES, groEL1, and groEL2, highlighting the regulation of these genes via the CIRCE-HrcA mechanism [16, 34]. The CIRCE (Controlling Inverted Repeat of Chaperone Expression) element is a palindromic DNA sequence (TTAGCACTC-N9-GAGTGCTAA) bound by HrcA to repress heat shock genes. Its conservation across bacteria underscores its role in stress response regulation [17]. To initiate the study of the heat shock response in these BCG strains and to infer whether the alteration in the HrcA protein of BCG Moreau would lead to changes in genes regulated by this transcriptional factor, we selected eight genes known to be regulated after heat shock (dnaK, ppiA, hrcA, groES, groEL1, groEL2, dnaJ2, and hspR) to evaluate their expression levels under thermal stress conditions by RT-qPCR. Some of these genes are well-documented components of the HrcA regulon and possess CIRCE elements upstream the coding sequence, such as hrcA itself, the groES operon (groES and groEL1), and the paralog groEL2 found in the genome [17]. Additionally, dnaJ2 is located downstream of hrcA. Furthermore, ppiA and dnaK were included due to prior data indicating higher expression levels in Pasteur compared to Moreau under non-stress conditions [35]. The hspR gene was added because it plays a key role in regulating the expression of various HSPs genes, including those in the Hsp70 operon (dnaK, grpE, and dnaJ) [16], as well as co-regulating, together with HrcA, the groES operon (whose promoter region contains CIRCE [HrcA] and HAIR [HspR] sequences) [17]. Our observed fold-induction of groEL2 (~ 3.9× in Pasteur after 15 min vs. 2.0× in Moreau after 60 min) aligns with prior reports in Mtb [16], while dnaK induction (2.0 after 60 min× vs. 1.8× after 60 min) might mirror strain-specific SigH activity [36]. In transcriptomic experiments comparing wild-type Mtb and a ΔsigH mutant, it was observed that immediately after the induction of oxidative stress (t = 0), triggered by exposure of the bacilli to diamide, the wild-type strain exhibited higher transcript levels of groEL1 and dnaK. Thirty minutes after induction, higher transcript levels of groEL2 were also observed in the wild-type strain. These results indicate that more than one molecular mechanism may control the expression of heat shock response proteins and that each gene may be subject to time-dependent regulation.

After quantifying the levels of these transcripts by RT–qPCR, we performed Western blot analyses to detect GroEL2, in order to assess whether RNA and protein levels for this target were consistent. For these assays, both intracellular and secreted protein fractions were analyzed. It is well-established that GroEL2, in addition to its chaperone function within the intracellular context, can be secreted and detected in a cleaved form through processing by the serine protease Hip1 [37]. This allows GroEL2 to modulate macrophage proinflammatory responses as well as influence dendritic cell functions, which are the primary antigen-presenting cells of the immune system [38]. Therefore, alterations in GroEL2 expression would be observed in both the intracellular protein extract and the secreted protein fraction. Within the bacterial cell, GroEL2 protein has a multimeric conformation, but after secretion and Hip1 action (proteolysis cleavage at its N terminus between amino acid residues Arg12 and Gly13) it acquires a monomeric conformation. The western blotting results support these findings. In the intracellular fraction, both bands corresponding to the monomeric form (denatured under SDS-PAGE conditions) and higher molecular weight bands were observed. These larger bands potentially indicate GroEL2’s chaperone activity, suggesting that it interacts with possibly misfolded proteins under heat shock conditions (not observed in the control condition). Conversely, in the secreted protein fraction, GroEL2 does not appear to perform such a function, as it was predominantly detected in its monomeric form. In this context, GroEL2’s role does not involve interactions with other proteins under the experimental conditions.

At first glance, the results from EMSA, RT-qPCR, and western blotting experiments might appear conflicting. Despite the lower capacity of Moreau’s HrcA to interact with DNA sequences containing the CIRCE element, heat shock-induced gene expression was more pronounced in Pasteur, which theoretically possesses a repressor with higher affinity for the promoter regions of its regulon genes. It is important to note that the comparisons made here involve different genetic backgrounds, as distinct BCG strains exhibit genotypic differences that may explain the observed discrepancies. For example, proteomic analyses comparing the culture filtrate profiles of BCG Moreau and BCG Pasteur under non-stress growth conditions revealed at least four proteins expressed at twice the level in BCG Pasteur: peptidyl-prolyl cis-trans isomerase (PPIaseA, rv0009), trigger factor (TF, rv2462c), Hsp65 (GroEL2, rv0440), and Hsp70 (DnaK, rv0350). These proteins are known for their roles as chaperones in the heat shock response, among other functions [35]. The ppiA gene is located within a duplication region specific to BCG Pasteur (DU1). This gene duplication alone could explain the increased expression; however, it is noteworthy that RT-qPCR did not reveal differences in transcript levels of this gene between BCG Moreau and Pasteur in response to heat shock. Another gene within this duplication region, sigH, encodes an alternative sigma factor involved in responses to heat and oxidative stress, positively regulating the expression of several genes, including dnaK and possibly groEL2 [39]. Under stress conditions, SigH is released from its anti-SigH factor, promoting the transcription of its target genes [40], such as groEL2 [39]. This suggests that groEL2 may undergo a second layer of regulation, influenced by the HrcA repressor [41]. This alternative sigma factor, sigH, has been under investigation as a knockout target and a potential candidate for novel vaccine development. These results underscore its critical function as a stress-response pathway in mycobacteria, which modulates oxidative stress. Disruption of this pathway results in a mutant strain incapable of counteracting host-derived oxidants in vivo, demonstrating stringent regulation by the primate innate immune system [42]. Furthermore, the presence of the DU1 duplication in the Pasteur strain may introduce confounding variables, hindering direct comparative analysis of the HrcA mutation’s phenotypic effects across the studied strains. The potential link between differences in sigH gene dosage and our findings remains speculative, and further experiments will be necessary to test this hypothesis.

Furthermore, these western blotting data demonstrate lower GroEL2 expression induction in the Moreau strain, resulting in reduced fluctuations of this protein’s abundance during stress conditions in both intracellular and extracellular environments. As previously mentioned, GroEL2 secretion - modulated by Hip1 protease - influences dendritic cell responses [37]. Consequently, the lower extracellular GroEL2 levels observed in Moreau may attenuate proinflammatory reactions, aligning with its clinical profile of reduced adverse effects [26].

Thus, BCG Pasteur, due to its duplication, could exhibit higher expression of these HSPs independently of differences in HrcA. Future experiments using the same genetic background - deleting hrcA in one BCG strain (either Pasteur or Moreau), complementing with both alleles, and comparing HSP expression - will be conducted to more accurately assess the functional impact of the HrcA alteration on the global HSP expression phenotype. Furthermore, it is known that multiple regulatory layers control the expression of these proteins in response to stress, such as heat shock. This includes: negative regulation via repressor interaction with operator regions in gene promoters; interactions between repressors and other proteins (e.g., HspR and Hsp70 [16], or HrcA and its still-unknown partner [17]); and positive regulation through increased expression of alternative sigma factors or their release from the respective anti-sigma factors. Regulatory mechanisms involving small RNAs (sRNAs) are also implicated, such as the F6 sRNA regulated by the alternative sigma factor SigF, which controls hrcA expression [43]. Therefore, our analyses remain preliminary and may reflect the influence of various regulatory layers acting on both repressor expression/activity and the co-regulated genes. Briefly, while EMSA revealed reduced DNA binding by Moreau HrcA, the higher heat shock gene expression in Pasteur may reflect its DU1 duplication (containing extra copies of sigH and ppiA), which amplifies stress responses independently of HrcA. This genetic background difference complicates direct attribution to HrcA alone, necessitating future studies in isogenic strains.

Overall, subtle variations in HSP expression and/or localization can profoundly impact interactions between the bacteria (in this case, different BCG vaccine strains) and the host immune system, ultimately influencing vaccine efficacy. Coordinated regulation of HSPs expression is critically important for stress responses in M. tuberculosis, as failure in this process could result in greater immune recognition of the bacilli and reduced survival [26].

BCG Moreau has been reported in some cases as producing fewer adverse reactions while maintaining a high degree of immunogenicity among the other vaccinal strains. This may be attributed to a deletion that impairs the synthesis of PDIM/PGL. Another specific deletion in BCG Moreau involves the rv3887c gene, which encodes a membrane protein within the type VII ESX-2 secretion system [44]. Although the role of rv3887c remains unknown, its homology with ESX-1 secretion components indicates that it may influence immunogenicity [44, 45]. Additionally, the lower reactivity of the BCG Moreau vaccine might be linked to the expression levels of HSPs (e.g., DnaK, GroES, GroEL1, and GroEL2), which are immunogenic and recognized as T-cell epitopes [46]. These HSPs are expressed at higher levels in BCG Pasteur compared to BCG Moreau. Thus, understanding the regulation of HSP expression is critical for elucidating the characteristics of vaccine strains, given the importance of these polypeptides in bacillus survival and the elicitation of an appropriate immune response to the vaccine.

Conclusions

This study provides comprehensive insights into the functional implications of a conservative missense mutation (V67L) in the hrcA gene of M. bovis BCG Moreau vaccine strain. Our integrated approach, combining structural biology and molecular analyses, reveals several key findings regarding the heat shock response regulation in mycobacterial vaccine strains. To our knowledge, this work constitutes the first report detailing both structural and functional features of this regulator in the BCG family.

The structural analysis demonstrated that the V67L substitution in BCG Moreau’s HrcA leads to subtle yet significant conformational adjustments in the protein’s architecture. While maintaining overall structural integrity, these changes appear to influence the protein’s DNA-binding capabilities, as evidenced by reduced affinity for CIRCE elements in EMSA analyses. However, the relationship between altered DNA binding and transcriptional regulation proved to be more complex than initially anticipated.

Notably, our expression analyses revealed an unexpected pattern wherein BCG Pasteur exhibited stronger transcriptional induction of heat shock proteins despite possessing an HrcA variant with apparently higher DNA-binding affinity. This observation highlights the complexity of heat shock regulation in mycobacteria, suggesting the involvement of multiple regulatory layers beyond direct HrcA-mediated repression. The presence of strain-specific genetic elements, such as the DU1 duplication in BCG Pasteur containing additional copies of stress-response genes, may contribute to these differences.

The differential expression patterns of heat shock proteins between Moreau and Pasteur strains may have important implications for vaccine efficacy. Lower baseline expression of immunogenic heat shock proteins in BCG Moreau, combined with its distinct regulatory response to stress, could contribute to its reported characteristics of reduced adverse reactions while maintaining protective immunity. These findings add to our understanding of strain-specific variations in BCG vaccines and their potential impact on vaccine performance.

The analyses presented in this study aim to infer a possible functional impact associated with HrcA mutation in M. bovis BCG Moreau. Although these findings are of interest and contribute to the understanding of the heat shock response, several limitations should be considered. Previous studies have demonstrated the importance of SigH, encoded within the genomic duplication present in BCG Pasteur, in regulating the response to this and other stresses. However, a direct connection between our observations and the difference in sigH dosage between the strains used in this study, BCG Moreau and BCG Pasteur, remains speculative. As noted above, additional experiments using isogenic strains will be required to better isolate these variables, although multiple factors likely contribute to the regulation of this biological process. Thus, this interpretation should be regarded only as a possible explanation rather than definitive evidence, and further experiments will be necessary to determine whether SigH indeed controls the expression of these heat shock proteins in this system. Additionally, investigation of potential strain-specific regulatory partners and the interaction of HrcA with other stress response pathways could provide further insights into the complex regulation of heat shock proteins in mycobacterial vaccine strains.

These results contribute to our broader understanding of how subtle genetic variations can influence bacterial stress responses and potentially impact vaccine characteristics. This knowledge may prove valuable for the rational design and optimization of future tuberculosis vaccines.

Abbreviations

Mtb

Mycobacterium tuberculosis

BCG

Bacillus Calmette-Guérin

HSPs

heat shock proteins

EMSA

electrophoretic mobility shift assay

PCR

Polymerase Chain Reaction

Authors’ contributions

LHFG and LML conceived and designed the study; LHFG, DA, PRC and LML conducted the experiments; LHFG, DA, PRC, MGAS and LML analyzed the data; MGAS and DA wrote the manuscript, and all authors read and approved the final version. All authors have read and agreed to the published version of the manuscript.

Funding

This work was funded by Fiocruz and the Brazilian National Council for Scientific and Technological Development (project CNPq-PAPES VI 421923/2017-2).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.