Mycobacterium tuberculosis H37Rv (SysteMTb), Mycobacterium smegmatis mc2155 (Snapper et al., 1990) and Escherichia coli NEB-5α and NEB-10β (New England Biolabs UK Ltd) were used in this study. All the strains were grown at 37°C, either in a shaking incubator (E. coli and M. smegmatis) or in a roller apparatus (M. tuberculosis). All work involving live M. tuberculosis was performed in a dedicated Biosafety Level 3 (BSL3) laboratory. M. tuberculosis and M. smegmatis were grown on Middlebrook 7H11 agar medium (BD Diagnostics) supplemented with 0.5% glycerol and 10% oleic acid albumin-dextrose-catalase (OADC) enrichment (BD Diagnostics). Liquid cultures of mycobacteria were grown in Middlebrook 7H9 broth (BD Diagnostics) supplemented with 0.2% glycerol, 10% albumin-dextrose-catalase (ADC) enrichment (BD Diagnostics) and 0.05% Tween 80 (Sigma). For the nutrient starvation and nitric oxide expression screens, M. tuberculosis was cultured in 100 ml of liquid medium in 1 L roller bottles rolling at 2 rpm, unless stated otherwise. E. coli was cultured on Luria-Bertani (LB) agar and in LB liquid medium. Kanamycin was added where appropriate [25 μg ml^–1 for mycobacteria, 50 μg ml^–1 for E. coli (Sigma)]. The exponential and stationary phases of growth were defined based on the growth curves obtained. For E. coli, exponential growth was considered for time points from 2 to 5 h and stationary growth from 6 to 9 h. For M. smegmatis, the exponential phase of growth was defined for time points from 6 to 12 h and stationary for time points 24 to 48 h. Finally, for M. tuberculosis, exponential growth phase was considered for time points 2–8 days and non-exponential growth from 11 days, which corresponded to 1 week after OD₆₀₀ had reached 1.0.

Mycobacterial genomic DNA was prepared using the InstaGene Matrix kit (BioRad) or the PureLink Genomic DNA Kit (Invitrogen), following manufacturers’ instructions. All the oligonucleotides used in this study are listed in Table 1 and were synthesized by Sigma-Aldrich or IDT. DNA sequences were determined by Source BioScience¹ using the Sanger method.

The plasmids used in this study are described in Table 2. The integrating expression vector pMV306trpT was created by cloning five copies of the transcriptional terminator trp from pEJ414 (Papavinasasundaram et al., 2001) into pMV306 (Stover et al., 1991) using KpnI + XbaI. Cloning of the terminator was confirmed by sequencing. The firefly luciferase (ffluc) gene was PCR amplified using primers FFluc_EcoRI_F and FFluc_SalI_R (Table 1) containing restriction sites as indicated and using pMV306hsp + FFluc (Andreu et al., 2010) as a template. Primers were designed to only amplify the coding region and hence exclude the optimized Shine–Dalgarno sequence. The sequence of the PCR product was confirmed by DNA sequencing. The ffluc gene was cloned into pMV306tprT after digestion with EcoRI-SalI to create pMV306trpT-FFluc(_delSD), herein referred as pTC1. The desA1 (Rv0824c) and desA2 (Rv1094) gene pair from M. tuberculosis were selected as Shine–Dalgarno and leaderless candidates, respectively. Reporter constructs were generated by fusing the candidate regions from desA1 and desA2 genes to the ffluc gene in pTC1 using NcoI-EcoRI as follows: (a) the Shine–Dalgarno reporter vector pTC1 + PdesA1:desA1’ was created by fusing the 50 bp promoter region, followed by the 5′UTR and six N-terminal amino acids from desA1; (b) the leaderless reporter pTC1 + PdesA2:desA2’ was created by fusing the 50 nt promoter region followed by the six N-terminal amino acids from desA2; (c) the Shine–Dalgarno reporter pTC1 + PdesA2:desA1’ was created by swapping the desA1 promoter in pTC1 + PdesA1:desA1’ by the desA2 promoter; and (d) the leaderless pTC1 + PdesA1:desA2’ reporter was created by swapping the desA2 promoter in pTC1 + PdesA2:desA2’ by the desA1 promoter. All candidate regions were introduced by oligo cloning and the sequences of the DNA oligos are available in Table 1. Briefly, oligos were reannealed by mixing equal volumes of both complementary oligos and placing them in a thermocycler with the following settings: first heat at 95°C for 2 min; second cool down to 25°C and incubate for 45 min and finally cool down to 4°C for further storage. All the sequences from the reporter constructs were confirmed by sequencing. Reporter plasmids were transformed into E. coli, and electroporated into M. smegmatis and M. tuberculosis as previously described (Goude et al., 2015). Correct construction of the reporter strains was verified by sequencing of specific PCR amplified regions (Table 1) from either plasmid DNA isolated from E. coli or chromosomal DNA isolated from M. smegmatis and M. tuberculosis. The generated strains were named according to the reporter fusion they carried.

Samples from M. tuberculosis pTC1 + PdesA1:desA1’, pTC1 + PdesA2:desA2’, pTC1 + PdesA2:desA1’ and pTC1 + PdesA1:desA2’ reporter strains were harvested from mid-exponential and stationary phase cultures and immediately processed for RNA extraction. For each sample, 30 mL of culture were spun down and RNA isolated using the FastRNA Pro blue kit (MP Biomedicals) following manufacturer’s instructions. All RNA samples were treated with Turbo DNase (Ambion) to remove any DNA contamination. The concentration and quality of RNA samples were assessed by Nanodrop (ND-1000, Labtech) and by running an Agilent RNA chip (2100 Bioanalyser). One microgram of purified RNA was then reverse transcribed into cDNA using the 5′RACE kit from Invitrogen, according to manufacturer’s instructions, and the obtained cDNA was further amplified using the specific oligo FFluc_R1 (Table 1) together with the oligos provided in the 5′ RACE kit. Amplified products were sequenced and the TSSs for each construct identified.

To assess the translation kinetics of the leaderless and SD reporters, luciferase activity assays were performed. The substrate for FFluc, D-luciferin (GoldBio^®), was prepared in distilled water at 94 mM (30 mg ml^–1) and filter sterilized. All stocks were stored at −20°C and diluted in equivalent broth media or PBS (without calcium or magnesium) immediately before use. Working solutions were kept on ice in the dark during preparation. For the experiments, two or three independent cultures of each strain were grown as described in the bacterial strains and growth conditions section within Materials and methods, and each culture was measured in duplicate or triplicate. Data presented in Figure 1B was acquired using a TD 20/20 tube luminometer (Turner Designs) using 500 μl of sample and 500 μl of D-luciferin solution (final concentration of 150 μg ml^–1) and default settings. All other data was acquired at 37°C with a FLUOstar Omega microplate reader (BMG Labtech). For measuring luminescence during the THP-1 infection experiments, infected macrophages were washed twice with Hank’s Balanced Salt Solution (HBSS) (ThermoFisher Scientific UK Ltd) to eliminate any extracellular bacteria and subsequently lysed using 0.1% Triton X-100 in HBSS and luminescence from released intracellular bacteria was measured. For performing the measurements, 96-well polystyrene microplates were prepared with 100 μl sample/well, and 10 μl of D-luciferin solution (final concentration of 150 μg ml^–1) was injected row by row and luminescence was immediately measured after adding the substrate for 3 s. Equivalent blank samples only containing broth medium were processed in parallel to each experiment and the luminescence measurements treated as the background luminescence. These control values were equivalent to luminescence measurements obtained for adequately processed cultures of M. tuberculosis, M. smegmatis and E. coli carrying the corresponding empty vector pTC1. Data was analyzed using the Mars software package (BMG Labtech) to calculate relative light units (RLUs).

For the nutrient starvation experiments, M. tuberculosis H37Rv was grown in 100 ml cultures until mid-exponential phase using 1 L roller bottles (Nalgene). When the cultures reached an OD₆₀₀ of 0.4–0.6, cells were harvested, washed twice with PBS and resuspended in 100 ml of fresh PBS supplemented with 0.025% tyloxapol (Sigma). Cultures were then incubated in 1 L roller bottles for 28 days. Parallel cultures were kept in 7H9 media as controls. For each reporter tested, two independent transformants were grown. The samples were collected from nutrient-starved and control cultures over 28 days for luminescence assays and for RNA extraction. Experiments were performed in duplicate.

For the nitric oxide experiments, M. tuberculosis H37Rv was grown in 100 ml cultures until mid-exponential phase using 1 L roller bottles. Diethylenetriamine/nitric oxide adduct (DETA/NO; Sigma) was prepared by dissolving in 0.1 M NaOH to reach a working concentration of 100 mM. When the cultures reached an OD₆₀₀ of 0.4–0.6, DETA/NO was added to a final concentration of 0.25 mM. Cultures were subsequently incubated in 1 L roller bottles for 7 days in parallel to control cultures kept in 7H9 with an equivalent concentration of NaOH to that of the DETA/NO cultures. For each reporter tested, two independent transformants were included. Samples were collected after 6, 24, 48, 72, 96, and 120 h for luminescence assays and for RNA extraction. Experiments were performed in duplicate.

THP-1 line monocytes (ATCC^®) were cultured routinely in RPMI-1640 medium containing HEPES and L-Glutamine (ThermoFisher Scientific UK Ltd), supplemented into RPMI complete medium with 10% heat-inactivated fetal bovine serum (FBS; Sigma) and 0.05 mM 2-mercaptoethanol (Sigma) at 37°C in 5% CO₂. The monocytes were differentiated into THP-1 macrophages by using the inducing reagent phorbol 12-myristate 13-acetate (PMA; Sigma) at a final concentration of 0.050 μg ml^–1. Cells were seeded in 24-well plates at a density of 10⁴ cells/well in 500 μL of medium and differentiated for 4 days prior to infection with M. tuberculosis. Mid-log phase cultures of M. tuberculosis reporter strains were washed twice with PBS, diluted in RPMI-1640 and added to THP-1 macrophage layers at a concentration of 10⁵ CFU/well (MOI of 10). After 3 h of infection at 37°C in 5% CO₂, macrophages were washed twice with Hank’s Balanced Salt Solution (HBSS) (ThermoFisher Scientific UK Ltd) to eliminate any extracellular bacteria. Lastly, 0.5 ml of complete RPMI was added to each well. Cells were either immediately processed or incubated further to reach 24, 48, 72, 96, and 168 h (7 days) post-infection. Bacterial intracellular survival and growth were assessed by lysis of the monolayers after addition of 0.1% Triton X-100 in HBSS and enumeration of bacteria by plating of serial dilutions in PBS-Tween onto 7H11 medium. Colonies were counted after 3–4 weeks and their count confirmed after 6 weeks of incubation at 37°C and the average CFU/ml determined. Inoculum represents the CFU/ml of the bacterial population used for the infections. Infection experiments were conducted in triplicate.

For the quantification of transcript levels through quantitative real-time PCR, parallel samples to the ones used for the luminescence screens were harvested for RNA isolation. For each sample, 10 mL of culture were spun down and RNA isolated using the FastRNA Pro blue kit (MP Biomedicals), following manufacturer’s instructions. All RNA samples were treated with Turbo DNase (Ambion) to remove any DNA contamination. The concentration and quality of the RNA samples were assessed by Nanodrop (ND-1000, Labtech) and by running an Agilent RNA chip (2100 Bioanalyser). Either 500 ng or 1 μg of the purified RNA were used for reverse transcription into cDNA using the High-Capacity cDNA Reverse Transcription kit (Applied Biosystems). Quantitative real-time PCR was carried out on a 7500 Fast Real-Time PCR system (Applied Biosystems) and using the KAPA SYBR FAST kit (Applied Biosystems), following the manufacturer’s protocol. Specific primers targeting the 16s rRNA were used as an (Table 1) internal control. Relative quantification of the expression levels was calculated using the 2^–ΔΔCT method (Livak and Schmittgen, 2001).

Statistical analysis was performed using GraphPad Prism 9.0.1 (GraphPad Software, San Diego, CA, United States²). Data was tested for normal distribution using the D’Agostino & Pearson omnibus test. Differences in translation efficiencies between reporters were assessed using unpaired multiple t-tests with False Discovery Rate correction when comparing two groups, or by one-way ANOVA with Tukey’s correction for multiple testing when comparing more than two groups. Non-parametric analysis was performed using the Mann–Whitney test or the Kruskal–Wallis test when comparing two or more groups respectively.

To study the role of leaderless translation in M. tuberculosis response to stress and infection, we first devised a system to easily quantify translation in M. tuberculosis by engineering a promoterless firefly luciferase (FFluc) encoding vector that could be used to generate suitable integrative translational reporters. As backbone we used the mycobacterial integrating vector pMV306 (Stover et al., 1991) with five copies of the transcriptional terminator trp (Papavinasasundaram et al., 2001). We then cloned the ffluc coding region codon-optimized for M. tuberculosis from pMV306hsp + FFluc (Andreu et al., 2010) and removed the optimized Shine–Dalgarno sequence to create pMV306trpT-FFluc(_delSD), which we refer herein to as pTC1 (Figure 1A). Median bioluminescence production in the resulting strain was significantly increased to that in the H37Rv wild-type strain with no ffluc gene (unpaired t-test, p = 0.01), indicating that there was some background transcription and translation of the ffluc gene in the promoterless vector (Figure 1B).

To compare translational efficiency between leaderless and Shine–Dalgarno transcripts, we selected the desA1 (Rv0824c) and desA2 (Rv1094) gene pair which encode homologous acyl-ACP desaturases sharing 30% primary sequence identity. Both genes are strongly expressed during exponential growth and have a typical TANNNT -10 promoter motif; however, while desA1 contains a 5′UTR with the Shine–Dalgarno sequence, desA2 is expressed as a leaderless transcript (Cortes et al., 2013). We obtained the Shine–Dalgarno reporter construct pTC1 + PdesA1:desA1’ by fusing the desA1 50 bp promoter region, the 5′UTR and the region encoding the first six amino acids of desA1 to ffluc in pTC1 (Figure 1C); similarly, we generated the leaderless reporter construct pTC1 + PdesA2:desA2’ by fusing the desA2 50 bp promoter region and the region encoding the first six amino acids of desA2 to ffluc in pTC1 (Figure 1C). To make sure that any differences observed between the Shine–Dalgarno and leaderless reporters were due to translational regulation and not to transcriptional regulation, we swapped the 50-bp desA1 and desA2 promoters in pTC1 + PdesA1:desA1’ and pTC1 + PdesA2:desA2’ to create the Shine–Dalgarno reporter pTC1 + PdesA2:desA1’ and the leaderless reporter pTC1 + PdesA1:desA2’ respectively (Figure 1C). All reporter sequences were verified by sequencing and plasmids were subsequently electroporated in M. tuberculosis. Successful integration was verified by polymerase chain reaction (PCR) followed by sequencing. Additionally, we verified that transcription of the reporters was driven by the desired promoter by determining the transcriptional start sites using 5′ rapid amplification of cDNA ends (RACE). Finally, we verified translation of the FFluc reporter by measuring luminescence of the reporter strains over a 5-day period of exponential growth as an indirect measurement of protein expression. All reporter strains pairs showed comparable levels of luminescence (Kruskal–Wallis test) (Figure 1D) suggesting comparable levels of luciferase production. As a result, we selected the Shine–Dalgarno pTC1 + PdesA1:desA1’ and leaderless pTC1 + PdesA1:desA2’ reporter pair, for which pTC1 + PdesA1:desA1’ contains a naturally occurring canonical Shine–Dalgarno organization for further experiments.

Our results so far suggested that the leaderless reporter was translated with similar efficiency to that of the Shine–Dalgarno reporter at least during exponential growth (Figure 1D). To identify possible differences in the translation efficiencies of the Shine–Dalgarno and leaderless reporters that could be associated with the growth status of the bacteria we used pTC1 + PdesA1:desA1’ and pTC1 + PdesA1:desA2’ to closely monitor luminescence production during the transition from exponential to non-exponential growth. As a control, we also measured luminescence production of the reporters in M. smegmatis, a close relative of M. tuberculosis that also encodes a high percentage of leaderless proteins in its genome (Shell et al., 2015). Previous experiments using fluorescent reporters have shown that leaderless translation in M. smegmatis has similar efficiency to that of a Shine–Dalgarno reporter during exponential growth conditions (Nguyen et al., 2020). Additionally, we measured luminescence production of the reporters in E. coli, a bacterium with scarce leaderless transcripts that are translated at low levels by the 70S monosome (Moll et al., 2002).

Introduction of the reporter vectors did not affect bacterial growth in any of the three bacterial models tested (Figures 2A–C). As predicted, during exponential growth luminescence production from the leaderless reporter in E. coli was 10-fold lower than that from the Shine–Dalgarno reporter (one-way ANOVA, p < 0.0001; Figures 2D,G and Supplementary Figure 1A). By contrast, in M. smegmatis, luminescence from the leaderless reporter was significantly higher than that from the Shine–Dalgarno reporter (one-way ANOVA, p < 0.001) (Figures 2E,H and Supplementary Figure 1A), whereas in M. tuberculosis luminescence production of the leaderless reporter was comparable to that of the Shine–Dalgarno reporter (Figures 2F,I and Supplementary Figure 1A). Upon exit from exponential growth, median luminescence levels increased for both reporters in E. coli, but the increase was not statistically significant in the case of the leaderless reporter (Figure 2G). In M. smegmatis, the mean luminescence levels of the Shine–Dalgarno reporter significantly decreased whilst the mean luminescence levels of the leaderless reporter were maintained (one-way ANOVA, p < 0.0001) (Figure 2H). In the case of M. tuberculosis, both Shine–Dalgarno and leaderless translation were significantly increased upon exit from exponential growth (one-way ANOVA, p < 0.0001) (Figure 2I).

In summary, our results show a robust translation of the leaderless reporter during exponential growth in M. tuberculosis, to a level comparable to that of the Shine–Dalgarno reporter. Upon exit from exponential growth, the mean levels of both reporters significantly increased compared to those during exponential growth, but mean levels of leaderless translation were significantly higher than mean levels of Shine–Dalgarno translation.

Despite mean levels of leaderless translation being significantly higher than mean levels of Shine–Dalgarno translation upon exit from exponential growth, our results did not indicate that the ratio of translation of leaderless and Shine–Dalgarno transcripts was different during the transition from exponential to non-exponential growth in M. tuberculosis (Supplementary Figure 1B). Next, we studied what happens during stress conditions. In particular, we determined the luminescence production of leaderless and Shine–Dalgarno reporters in M. tuberculosis during nutrient starvation and following a transient stress with a sub-lethal concentration of NO. These two conditions are representative of conditions encountered by M. tuberculosis during non-replicating growth (Betts et al., 2002) and active infection in humans (Nathan, 2006; Robinson et al., 2014), respectively. In each condition tested, we addressed translation kinetics of the leaderless and Shine–Dalgarno reporters by measuring luminescence produced from the pTC1 + PdesA1:desA1’ and pTC1 + PdesA1:desA2’ reporter vectors and quantified changes at the transcriptional level through quantitative real-time PCR to rule out changes due to differences in promoter activities.

For the nutrient starvation experiments, cells growing exponentially in nutrient-rich media were washed with PBS and used as the inoculum (time zero). PBS (nutrient starvation) and nutrient-rich media (7H9 media, control) cultures were incubated for 28 days. The control cultures grew exponentially for 5 days before reaching stationary phase (Figure 3A). The control cultures reached stationary phase earlier than the cultures in Figure 2C, likely because of the higher starting OD (see section “Materials and Methods”). As expected, nutrient limitation affected the growth and luminescence production of both the leaderless and Shine–Dalgarno M. tuberculosis reporter strains compared to those in the control cultures (Figures 3A,B). In both the control and the nutrient starvation cultures, the leaderless reporter was translated at a higher level than the Shine–Dalgarno reporter (multiple t-tests, p < 0.03) (Figure 3B and Supplementary Figure 2). When comparing changes in the luminescence production of the reporters during nutrient starvation with those growing in control conditions, we found a significantly higher increase in luminescence production of the leaderless reporter compared to that of the Shine–Dalgarno reporter after 24 h of nutrient starvation (multiple t-tests, p < 0.001) (Figure 3C). This difference was not driven by transcriptional changes, as transcription levels of the Shine-Dalgarno and leaderless reporters did not change significantly (one-way ANOVA, p > 0.163) (Figure 3D). This less pronounced effect of nutrient starvation on leaderless-mediated luminescence production was transient; a comparably high reduction of luminescence production of both types of reporters was observed at 72 h post-nutrient starvation and later time points (Figure 3C).

Next, we studied changes in translation during exposure to a sub-lethal concentration of NO by monitoring changes in luminescence in exponentially growing cultures of the M. tuberculosis leaderless and Shine–Dalgarno reporter strains exposed to 0.25 mM NO or sodium hydroxide (NaOH, used to dissolve the NO adduct, see section “Materials and Methods”) for 7 days. We chose 0.25 mM NO as it has been previously demonstrated that exposure to intermediate levels of NO (0.5 and 1.0 mM) results in a temporary cessation of growth (Voskuil et al., 2011). As expected, exposure to NO caused a rapid but transient growth arrest of both the Shine–Dalgarno and leaderless reporter strains that could be observed from 48 h post-NO exposure (Figure 4A). Throughout the experiment, the leaderless reporter was translated at a higher level than the Shine–Dalgarno reporter in both the NO-treated and the NaOH control cultures (multiple t-tests, p < 0.03) (Figure 4B). Indeed, although the transient inhibition in growth resulted in a decrease in luminescence production of both reporters compared with the corresponding controls, both reporters were affected to a similar degree and hence translation of the leaderless reporter was higher than that of the Shine–Dalgarno reporter (Figures 4B,C). As DETA/NO is reported to release NO with a half-life of 5.5 h under similar M. tuberculosis growth conditions (Voskuil et al., 2003), it is likely the effects here observed are due to the initial NO exposure alone. Forty-eight hours after NO treatment started, both reporter strains resumed growth and luminescence production, suggesting that they had recovered from the stress (Figures 4A,B). At this point, the percentage of luminescence production of the leaderless reporter in comparison with that in the corresponding control was significantly higher than that of the Shine-Dalgarno reporter (27% vs. 14%, respectively, multiple t-tests, p < 0.002), suggesting a quicker recovery of leaderless translation following exposure to NO treatment (Figure 4C). These differences were not driven by transcriptional changes, as we found no significant differences in the transcription levels of Shine–Dalgarno and leaderless transcripts (Tukey’s multiple comparisons test, p > 0.162) (Figure 4D).

Overall, the results show that luminescence production of the leaderless reporter is significantly less affected by nutrient starvation and recovers more quickly from nitrosative stress than luminescence production from the Shine-Dalgarno reporter.

Finally, we studied luminescence production of the leaderless and Shine–Dalgarno reporters during M. tuberculosis intracellular growth in macrophages. To this end we measured luminescence of the reporter strains during infection of PMA-activated THP-1 cells. We used an MOI of 10 to ensure sufficient intracellular bacteria would be present to allow detection of reporter luminescence.

Following infection of THP-1 cells, no increase in luminescence and growth was observed for the reporter strains during the first 24 h (Figures 5A,B). At 48 h post infection, both reporter strains resumed growth and translation, suggesting that they had adapted to the intracellular environment. At 24 and 48 h post infection, there was a significant decrease in the growth of the leaderless reporter compared to that of the Shine–Dalgarno reporter (Figure 5A; multiple t-tests, p < 0.0001), but this reduction on viable bacteria for the leaderless reporter did not correlate with a significant reduction in luminescence production (Supplementary Figure 3A). When corrected by growth, luminescence from the leaderless reporter was higher than that from the Shine–Dalgarno reporter and this difference was significant between 24 and 72 h post infection (multiple t-tests, p < 0.002) (Figure 5B). To quantify how translation of the leaderless and Shine–Dalgarno reporters responded during the different stages of macrophage infection, we normalized the luminescence levels of the reporters at different time points against the luminescence levels at time 0. This revealed that during the first 48 h of infection, before bacterial growth resumed, translation of the leaderless reporter was less affected than that of the Shine–Dalgarno reporter, and this difference was statistically significant at 48 h post infection (multiple t-test, p < 0.05) (Figure 5C). As growth resumed following adaptation to the intracellular environment, the luminescence production of both types of transcripts increased to comparable levels (Figure 5C and Supplementary Figure 3B). Our data show that during the early stages of macrophage infection, the luminescence production of the leaderless reporter is significantly less affected by exposure to the intracellular host environment than that of the Shine–Dalgarno.