Mycobacterium bovis BCG (Pasteur 1173P2) overproducing the KdpF peptide (pKdpF) and its control strain (containing the empty vector, pMV261) were used in this study (Gannoun-Zaki et al., 2013). The overexpression of kdpF gene from the pMV261 vector (Stover et al., 1991) is triggered by the hsp60 promoter and is therefore constitutive. Both strains were grown at 37°C in liquid culture in Sauton's medium containing 0.025% tyloxapol (Sigma), in the presence of kanamycin (25 μg/ml) as required or onto Middlebrook 7H10 agar (Difco) plates supplemented with Oleic Acid-Dextrose-Catalase (OADC). Plates were incubated at 37°C for 2–3 weeks prior to visual counting of the colony forming units (CFUs).

Peptide synthesis was performed by a solid-phase method using the Fmoc methodology on an automated microwave peptide synthesizer (Liberty-1, CEM, Orsay, France) as described previously (Belon et al., 2016). Due to failure in the synthesis of the entire KdpF peptide (MTTVDNIVGLVIAVALMAFLFAALLFPEKF), a shorter peptide deleted for 5 amino-acids at the N-terminus (MIVGLVIAVALMAFLFAALLFPEKF) was synthesized. The final product was resuspended at 3.5 mM in DMSO/water (25%/75%, vol/vol) and analyzed by MALDI-TOF mass spectrometry.

The human monocytic cell line THP-1 was obtained from the ATCC collection (TIB-202™). Cells were grown at 37°C in a 5% CO₂ atmosphere in RPMI 1640 with Glutamax I (Life Technology) supplemented with 10% fetal calf serum, 0.1 mM non-essential amino acids and 0.15% sodium pyruvate (Life Technology). Cultures were maintained at a density not > 5 × 10⁵ cells/ml.

For infection experiments, THP-1 cell suspensions were adjusted to a concentration of 2 × 10⁵ cells/ml in complete RPMI medium, seeded in 24-well plates, and cells were allowed to adhere and differentiate in the presence of 60 nM Phorbol 12-Myristate 13-Acetate (PMA, Sigma) for 72 h at 37°C. The medium of differentiated THP-1 cells was removed and cells were infected with exponentially growing M. bovis BCG cultures (OD₆₀₀ = 0.8) at a multiplicity of infection (MOI) of 5:1 bacteria per macrophage as previously described (Corrales et al., 2012). After 3 h of incubation, the infected cells were washed 3 times with PBS buffer to remove any extracellular bacteria, and then incubated in fresh medium for 6 days. For scoring, cells were lysed with 0.1% Triton X-100 in PBS at selected times post-infection and serial dilutions of the lysate were plated onto agar plates.

To inhibit reactive nitrogen intermediates (RNIs) production, 200 μM N-ω-nitro-L-arginine methyl ester hydrochloride (L-NAME, Sigma) was added to the culture medium of activated macrophages 1 h before infection and maintained throughout the infection process. Data were normalized to multiplication of the M. bovis BCG pMV261 strain in non-treated L-NAME macrophages to account for differences in bacterial multiplication between the two bacterial strains in macrophages L-NAME-treated or not.

To test the effect of synthetic KdpF on intramacrophage bacterial replication, M. bovis BCG pMV261 was incubated with synthetic KdpF (70 μM) during 10 min at room temperature. These pre-treated bacteria were used to infect THP-1 macrophages and plated after 3 h and 6 days post infection onto agar plates.

Mycobacterium bovis BCG pKdpF and M. bovis BCG pMV261 were grown to mid log phase and diluted to obtain 3 × 10⁷ bacteria/ml. Tubes containing the bacilli were incubated for 5 h at 37°C in Sauton's medium containing 0.1 mM or 1 mM of the NO donor SNAP (S-nitroso-N-acetyl-DL-penicillamine, Sigma) (Ferrero et al., 1999) supplemented with 10 mM CuSO₄ (Nelli et al., 2000). Serial dilutions were plated onto Middlebrook 7H10 agar containing 25 μg/ml kanamycin.

To test the effect of synthetic KdpF peptide on mycobacterial susceptibility to NO stress, tubes containing M. bovis BCG pMV261 (1.5 × 10⁷ bacteria) were incubated at 37°C for 5 h in Sauton's medium containing 70 μM of synthetic KdpF and 1 mM of the NO donor SNAP, supplemented with 10 mM CuSO₄. Serial dilutions were plated as described earlier. The obtained CFUs from the control cultures (also supplemented with 10 mM CuSO₄-DMSO 25%) were considered as 100% viability.

THP-1 cells were infected with M. bovis BCG pMV261 and M. bovis BCG pKdpF as described earlier. Supernatants of infected THP-1 cells were collected after 3 h and 6 days. The production of NO was indirectly measured by assaying for the presence of nitrite (NO2−) using the Griess reagent (Green et al., 1982). Aliquots of 100 μl of culture media were distributed in 96-well microtiter plates and equal volumes of Griess reagent solution were added. The reaction was allowed to proceed for 15 min at room temperature, and the optical density at 540 nm was measured. The amounts of produced NO were quantified from a standard curve using dilutions of NaNO₂ (100 nM–100 μM).

RNA was prepared as previously described (Gannoun-Zaki et al., 2013) from 5 ml of mid-logarithmic bacterial cultures grown in Sauton's medium supplemented with 10 mM CuSO₄ or grown for 5 h in Sauton's medium supplemented with 1 mM SNAP. Quantitative real-time PCR was performed using an in-house SYBR Green mix (Lutfalla and Uze, 2006) and a 480 light cycler instrument (Roche). PCR conditions were as follows: 3 min denaturation at 98°C, 45 cycles of 98°C for 5 s, 68°C for 10 s, and 72°C for 10 s. The 16S ribosomal RNA gene (rrs) was used as internal control to calculate the relative level of gene expression. The sequences of primers for qRT-PCR are listed in Table 1.

The comparison of gene expression level in M. bovis BCG strains cultured in presence or not of 1 mM SNAP was performed with REST 2009 software (Vandesompele et al., 2002) and plotted using GraphPad-Prism software.

The Bacterial Adenylate Cyclase Two-Hybrid (BACTH) system (Karimova et al., 1998) was used to evaluate the interaction between KdpF and MmpL7, NarI, NarK2, Rv2617c, or SecG. The KdpF-T25 and MmpL7-T18 constructions were previously described (Gannoun-Zaki et al., 2013). To generate the pUT18-NarI, pUT18-NarK2, and pUT18-SecG constructions (plasmids are listed in Table 2), genes were PCR amplified from M. tuberculosis genomic DNA using the primers listed in Table 1. The plasmid pUT18-Rv2617c was provided by Dr. Fabiana Bigi (Klepp et al., 2009). Recombinant plasmids derived from pUT18 and pKT25 genes were co-transformed into E. coli BTH101 bacteria and BACTH assay were conducted at 30°C as described (Karimova et al., 1998). A level of β-galactosidase activity at least 5-fold higher than the one of the control vectors indicates an interaction between the protein partners.

Escherichia coli BTH101 strain transformed with plasmid encoding MmpL7-T18, NarI-T18, or Rv2617c-T18 fusions was grown overnight at 30°C in 250 μl of LB Broth containing 100 μg/ml ampicillin, 0.5 mM IPTG and 70 μM of synthetic KdpF peptide. A control experiment was performed using E. coli BTH101 carrying the SecG-T18 plasmid, a membrane protein that does not interact with KdpF peptide. Bacterial cultures were centrifuged, resuspended in 50 μl Laemmli buffer and incubated for 5 min at 95°C. Bacterial lysates were loaded on 12.5% SDS-PAGE and transferred onto nitrocellulose membrane (Invitrogen) for immunoblotting. Mouse anti-T18 antibodies (monoclonal antibody sc-13582, Santa-Cruz) was used at 1:1000 dilution. Mouse anti-DnaK antibody (Tebubio) at 1:5000 dilution was used as loading control. Secondary antibodies labeled with IRDye 800 CW infrared dyes (LICOR) were used and the images were acquired on LICOR Odyssey Fc Imaging System to quantify the amount of proteins.

Dose-effect experiment was similarly conducted with various KdpF synthetic peptide concentrations (14, 35, 70, and 140 μM, respectively).

Statistical analyses were performed by using an unpaired Student t-test. Only P < 0.05 were considered as statistically significant.

In a previous study, we showed that replication inside human macrophages of M. bovis BCG overproducing KdpF was significantly reduced in comparison to the M. bovis BCG-pMV261 control strain (Gannoun-Zaki et al., 2013). In order to determine whether the NO could be responsible for the impaired intracellular multiplication of KdpF overproducing bacteria, we first determined the susceptibility of bacterial strains to NO and then measured its production in infected THP-1 macrophages. For this purpose, M. bovis BCG-pMV261 and M. bovis BCG-pKdpF strains were treated for 5 h with 0.1 and 1 mM SNAP, a compound releasing NO (Ferrero et al., 1999). The addition of 0.1 and 1 mM SNAP in M. bovis BCG-pMV261 bacteria led to a slight diminution of bacterial viability (73 and 61%, respectively). Notably, the effect of SNAP on viability was significantly more pronounced for M. bovis BCG-pKdpF strain (39 and 24% for 0.1 and 1 mM SNAP-treated bacteria, respectively) (Figure 1A). This shows that KdpF-overproducing bacteria are more susceptible to nitrosative stress than bacteria carrying the pMV261 control vector.

We next quantified the amount of NaNO₂ released from infected macrophages into the culture medium using the Griess reagent. A significant higher amount of NaNO₂ was obtained in macrophages infected with the KdpF overproducing strain as compared to the M. bovis BCG-pMV261 infected cells after 6 days of infection (23.12 vs. 11.38 μmol NaNO₂/well), (Figure 1B). This indicates that macrophages infected with KdpF overproducing bacteria accumulate more NO2− as compared to M. bovis BCG-pMV261 infected macrophages. This result suggests that the M. bovis BCG-pKdpF strain has hampered NO detoxification.

To evaluate whether the NO accumulation is responsible for the decreased intracellular replication of M. bovis BCG-pKdpF strain, infections were carried on human THP1-macrophages pre-treated or not with the reactive nitrogen intermediates (RNIs) inhibitor (L-NAME). In the absence of L-NAME, the strain overproducing KdpF showed a lower replication rate than M. bovis BCG-pMV261 strain after 6 days of infection as measured by CFUs counting (Figure 2A), which agrees with our previous results reported with primary human macrophages (Gannoun-Zaki et al., 2013). However, this defect was suppressed upon treatment of THP1- macrophages with L-NAME since similar CFUs counts were obtained after macrophage infection with both strains (Figure 2B). A representation of the multiplication rate (ratio J6/J0), considering non-treated pMV261 strain as 100%, clearly shows that the significant difference between the two strains is suppressed upon SNAP treatment (Figure 2C). Conversely, NO inhibition by L-NAME led to a slight increase of intracellular replication percentage (180% ± 27) of M. bovis BCG-pMV261 strain when compared to non-treated infected cells, thus revealing that NO stress limits bacterial intracellular replication. On the other hand, we have considered the ROS-mediated killing as well by using the specific ROS inhibitor NAC (N-Acetyl-L-Cystein) prior to macrophage infection. As shown in Figure S2, no recovery of bacterial replication was obtained in NAC-treated macrophages infected with M. bovis BCG overproducing KdpF. Taken together, our results show that RNIs produced upon infection are involved in the reduced replication of the KdpF overproducing strain, suggesting that the increased susceptibility of this strain to nitrosative stress is correlated to the growth defect inside macrophages.

Reactive nitrogen intermediates are detoxified through several reactions involving Nar and Nir proteins family to finally produce non-toxic ammonium. The effect of KdpF on nitrosative stress proteins was first investigated at the transcriptional level by quantifying gene expression in both M. bovis BCG-pKdpF and M. bovis BCG-pMV261 strains. Among the genes involved in the process of NO detoxification, we focused our study on two membrane proteins encoded by narI and narK2 genes. Besides these genes, rv2617c and mmpL7 genes described to be, respectively up- and down-regulated in mycobacteria upon NO stress were used as controls (Ohno et al., 2003; Cossu et al., 2013). For this purpose, quantitative RT-PCR was carried out using RNA extracted from both M. bovis BCG strains grown for 5 h in the presence or absence of 1 mM SNAP. Results obtained from the SNAP-treated M. bovis BCG-pMV261 strain showed that the expression of the rv2617c gene was clearly upregulated by nitrosative stress (Figure 3A), to an extent that was even greater than the one described in M. tuberculosis (Ohno et al., 2003). On the other hand, the nar genes were not upregulated in SNAP-treated bacteria (Figure 3A), which contrasts with what has been described in M. tuberculosis (Cossu et al., 2013). As expected, expression of the M. bovis BCG mmpL7 gene was strongly repressed in presence of SNAP, as previously described for PDIM synthesis genes in M. tuberculosis (Cossu et al., 2013). The regulation of narI, narK2, rv2617c, and mmpL7 gene expression upon SNAP treatment was then investigated in the M. bovis BCG strain overproducing KdpF (Figure 3B). A similar pattern of expression, as obtained with M. bovis BCG-pMV261 strain, was observed, thus indicating that KdpF does not play a role on transcriptional regulation of these genes under nitrosative stress.

Membrane peptides have been shown to possess regulatory functions through direct interaction with membrane proteins (Alix and Blanc-Potard, 2009). Since we demonstrated that KdpF had no effect on transcriptional regulation of genes involved in nitrosative stress, we then hypothesized that KdpF could modulate the function or stability of these proteins through direct interaction.

We thus examined the interaction between KdpF and the membrane proteins NarI, NarK2, Rv2617c, and MmpL7 (Figure 4A) using the bacterial two-hybrid (BACTH) system that has been previously validated for detecting interactions between inner-membrane proteins in living bacteria (Karimova et al., 2005). Constructs were made to fuse the T18 subunit of the adenylate cyclase to these diverse membrane proteins (plasmid constructs are described in Table 2) and interaction was tested with the KdpF peptide fused to the T25 subunit (Gannoun-Zaki et al., 2013).

Plasmids encoding the T18 and T25 fusion proteins were then introduced into an E. coli cya mutant (BTH101) and functional complementation was determined by measuring β-galactosidase activity (Figure 4B). We have previously shown that the MmpL7 transporter, involved in the PDIM transport highly interacted with KdpF (Gannoun-Zaki et al., 2013), and the MmpL7-T18 construct was thus used as positive control. A high level of β-galactosidase activity (1866 Miller units) was observed with Rv2617c-T18 plasmid, thus indicating a robust interaction between Rv2617c and KdpF. An interaction was also observed between NarI and KdpF, and NarK2 and KdpF (720 and 868 Miller units, respectively). Similar levels of β-galactosidase activity were obtained when reversing the bait and the prey for Rv2617c and NarI (2030 Miller units for assay with Rv2617c-T25 and KdpF-T18 and 1200 Miller units for assay with NarI-T25 and KdpF-T18, data not shown), thus confirming significant interaction. On the other hand, no interaction (β-galactosidase activity similar to the negative control pUT18/KdpF-T25) was detected between KdpF and SecG, a membrane protein unrelated to nitrosative stress used as control. Cumulatively, these results indicate that KdpF peptide can interact significantly and specifically with membrane proteins related to nitrosative stress, which could in turn modulate the regulation of these proteins through protein-protein interaction.

In order to investigate whether the interaction of KdpF peptide with protein partners could interfere with protein stability, we have used a synthetic KdpF peptide that was exogenously added to BTH101 bacteria expressing diverse T18 fusion proteins. While the full-length peptide failed to be synthetized in sufficient soluble amount, a shorter peptide lacking the first 5 amino-acids in its N-terminus was produced (see Materials and Methods). We first validated that this shorter peptide, possessing the same hydrophobic domain as KdpF, carried identical properties in a bacterial two-hybrid assay (Figure S1). Then, the synthetic KdpF peptide was directly added to the culture medium of BTH101 bacteria transformed with various Mtb-proteins-T18 derivative plasmids related to nitrosative stress (MmpL7, NarI, NarK2, and Rv2617c). Equal amount of bacterial lysates treated or not with synthetic peptide were analyzed by western blotting using T18-specific antibodies. Our data showed that addition of synthetic peptide (70 μM) resulted in a strong (12-fold) decreased level of NarI-T18 protein, a mild (3-fold) decreased level of MmpL7-T18 protein, whereas only a slight decrease in protein amount was observed for Rv2617c-T18 construct (Figure 5A). On the other hand, no effect was observed upon addition of the synthetic KdpF peptide on the stability of the SecG-T18 control protein. We further performed a dose-response assay with increased concentrations of synthetic KdpF peptide (from 14 to 140 μM) and the NarI-T18 fusion protein (Figure 5B). Our data show a correlation between the increase of KdpF concentration and the decrease of NarI-T18 protein level. Taken together, these results suggest that upon interaction, KdpF peptide could decrease the stability of proteins related to nitrosative stress, mainly NarI and to a lesser extent, MmpL7.

In order to assess the biological activity of exogenously added KdpF peptide on mycobacteria, M. bovis BCG-pMV261 strain was first pre-incubated or not with 70 μM of synthetic KdpF peptide, and grown in culture media supplemented or not with the NO donor SNAP. No effect of the exogenous KdpF peptide was observed on bacterial viability when bacteria were grown in the control culture media containing copper, DMSO (Figure 6A). However, when the strain was grown in presence of 1 mM SNAP, a significant diminution of bacterial viability (59%) was measured when bacteria were preloaded with synthetic peptide (Figure 6B). These results suggest that the addition of synthetic KdpF peptide to M. bovis BCG-pMV261 bacteria increases the susceptibility of mycobacteria to nitrosative stress, thus mimicking the effect of endogenous KdpF overproduction.

We next investigated the effect of exogenous KdpF peptide addition on mycobacterial intramacrophage replication. The M. bovis BCG-pMV261 strain was incubated with 70 μM of synthetic peptide prior to THP-1 macrophage infection. Bacterial multiplication, evaluated by CFUs counts after 6 days of infection, was significantly decreased (68% of multiplication) in bacteria pre-incubated with the exogenous peptide in comparison to the M. bovis BCG-pMV261 control bacteria (Figure 6C). Therefore, these results indicate that the exogenous synthetic KdpF peptide impairs intracellular multiplication, again mimicking the effect of endogenous KdpF overproduction, and thus confirming its role as a potent molecule with anti-virulence properties.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.