Bacterial strains of M. smegmatis mc² 155 were grown in 7H9 liquid medium (Difco) supplemented with 0.5% glycerol. According to Betts et al. (2002), with minor modification, the nutrition starvation culture (Albrethsen et al., 2013; Duan et al., 2016) was cultured at 37°C, 110 rpm, until OD₆₀₀ = 0.8, and the culture was pelleted washed twice with 1× PBS, resuspended in 1× PBS, transferred to standing flasks or microplate, and incubated in the bottles (Shuniu) for 0, 48 or 72 h, until the treatment with antibiotics or amino acids. The final concentration of antibiotic is 100 μg/mL. The viability was determined by serial dilutions and plating on 7H10 agar.

For the drug treatment assay after nutrition starvation, the starved M. smegmatis culture was diluted 10 times in M9 medium without glucose at 0, 48, and 72 h time points, and 1 mL was aliquoted into each well of a 24-well plate, every sample with three replicates. Rifampicin was added to duplicate wells of cultures at a final concentration of 48.6–291.6 μM. Then concentration of 0–20 mM glutamine was added to M. smegmatis cultures. The MIC (μM) of rifampicin for M. smegmatis mc² 155 was 4 μg/mL. The aqueous solutions of antibiotics and amino acids were filter sterilized by passing through a 0.22 μM syringe-fitted filter. Cultures were incubated with or without antibiotics at 37°C for 0–72 h, followed by serial diluting and plating on 7H10 agar to determine the bacterial viability.

Biofilm formation (Recht and Kolter, 2001; Bardouniotis et al., 2003) in 7H9 liquid with 10% glucose was carried out in six-well plates with or without glutamine. For rifampicin treatment, in vitro biofilms were grown in triplicate for 48 h on 96-well plates as described (Lebeaux et al., 2014) with minor modification. The planktonic bacteria were removed by washing with 1× PBS and biofilms were treated for 24 h with increasing concentrations of rifampicin (48.6–145.8 μM) with or without glutamine in M9 medium. There is no glucose in M9 medium while the control wells without antibiotics.

Mycobacterium smegmatis was cultured in 250 mL bottles under starvation conditions, each sample with three repeats. For antibiotic treated samples, three 50 mL cultures were pelleted by centrifugation and washed twice with 1× PBS, followed by resuspending in M9 media without glucose, in the presence of rifampicin (145.8 μM), incubated 12 h with or without glutamine (Duan et al., 2016). For metabolome analysis, samples were pelleted and subjected to a μHPLC (1290 Infinity LC, Agilent Technologies) (Ivanisevic et al., 2013) coupled to a quadrupole time-of-flight (AB Sciex TripleTOF 6600) analysis. The raw MS data were converted into the MzXML files by using a ProteoWizard MSConvert tool and processed using the XCMS (Benton et al., 2008) for feature detection, retention time correction, and alignment. The metabolites were identified by accuracy mass (<25 ppm) and MS/MS data, which matched with our standard database. For multivariate statistical analysis, the web-based MetaboAnalyst¹ system was used, followed by the Pareto scaling, principal component analysis (PCA) and partial least-squares-discriminant analysis (PLS-DA). Then leave one out cross-validation and response the permutation testing were used to evaluate the robustness of the model. The significantly different metabolites were determined based on the combination of a statistically significant threshold of variable influence of projection (VIP) value obtained from PLS-DA model and two-tailed Student’s t-test (P-value) on the raw data, and the metabolites with VIP values larger than 1.0 and P-values less than 0.05 were considered as significant.

The M. smegmatis was cultured in 250 mL bottles under starvation conditions, each sample with three repeats. For antibiotic treated samples, three 50 mL cultures were pelleted by centrifugation, washed twice with 1× PBS followed by resuspending in M9 medium destitute of glucose, treated with 145.8 μM rifampicin for 12 h in the presence/absence of glutamine (Duan et al., 2016). The procedure was adapted from reference (Mortazavi et al., 2008) with slight modification. Total RNA was isolated using the RNeasy mini kit (Qiagen, Germany). Paired-end libraries were synthesized by using the TruSeq RNA Sample Preparation Kit (Illumina, United States) according to TruSeq RNA Sample Preparation Guide. The cleaved RNA fragments were copied into first strand cDNA using reverse transcriptase and random primers. The cDNA fragments were synthesized afterward through an end repair process followed by the addition of a single “A” base, and then the fragments were ligated to the adapters. The products were then purified and enriched with PCR to create the final cDNA library. The cluster was generated by cBot with the library being diluted to 10 pM and then was sequenced on the Illumina HiSeq 2500 (Illumina, United States). The library construction and sequencing were performed at the Shanghai Biotechnology Corporation. The sequencing raw reads were preprocessed by filtering out the rRNA reads, sequencing adapters, short-fragment reads, and other low-quality reads. After genome mapping, the Cufflinks v2.1.1 was run with a reference annotation to generate a FPKM values for known gene models. Differentially, expressed genes were identified using Cuffdiff. The P-value significance threshold in multiple tests was set by the false discovery rate (FDR). The fold-changes were also estimated according to the FPKM in each sample. The expressed genes were selected by using the following filter criteria: FDR ≤ 0.05 and fold-change ≥ 2. The data can be found in GEO website with accession number GSE108001.

Mycobacterium smegmatis cultures undergone nutrition starvation for 48 h were pelleted by centrifuged at 8,000 rpm (7990 g), 10 min and resuspended in M9 medium deprived of glucose, followed by exposing in 145.8 μM rifampicin with or without 2 mM glutamine for 12 h. The culture was harvested and diluted to OD₆₀₀ = 0.2. For NADH measurement, cell pellets were washed twice with 1× PBS and resuspended with NADH extraction buffer (0.1 M HCl). Heat extraction was performed at 60°C for 5 min. NAD⁺ extraction buffer (0.1 M NaOH) were added to neutralize the extractions. Cells were vortexed and spun down at 12,000 rpm (13,800 g) for 5 min. The supernatant was collected for the EnzyChrom NAD⁺/NADH Assay Kit (BioAssay Systems). Then 40 μL of sample was transferred into each well of a clean, flat-bottom 96-well plate. For each reaction well, working reagents consisting of 60 μL assay buffer, 1 μL enzyme A, 1 μL enzyme B, 14 μL lactate, and 14 μL MTT were then added. The optical density (OD₅₆₅) for time “zero” at 565 nm (520–600 nm) and OD₅₆₅ after 15 min incubation at room temperature were recorded.

Mycobacterium smegmatis cultures undergone starvation for 48 h were pelleted by centrifugation at 8000 rpm (7,990 g), 10 min and resuspended in M9 medium without glucose, resuspended samples were exposed to 145.8 μM rifampicin for 24–72 h with or without 2 mM glutamine. The control cultures received no drug. For ROS analysis, aliquots were taken at indicated time points and diluted to OD₆₀₀ = 0.2. The samples were treated by the Reactive Oxygen Species Assay Kit (Beyotime). DCFH-DA was diluted 1:1000 with 1× PBS. Pelleted samples were resuspended with diluted DCFH-DA at 37°C for 1 h. The harvested cultures were pelleted by centrifugation at 12,000 rpm (13,800 g) for 1 min and then washed twice with 1× PBS. Fluorescence was analyzed on a BD FACS Calibur (BD Biosciences) as previously described (Vilchèze et al., 2013).

Mycobacterium smegmatis cultures undergone nutrition starvation for 48 h were pelleted by centrifugation at 8,000 rpm (7,990 g) for 10 min and then resuspended with M9 medium without glucose. Resuspended samples were exposed to 145.8 μM rifampicin for 24–72 h with or without 2 mM glutamine. There is no antibiotics in the control cultures. The aliquots used for TUNEL analysis were removed at indicated time points, and treated with a TUNEL reaction mix from the In situ Cell Death Detection Kit (Roche Molecular Biochemicals. The samples were washed twice with 1× PBS and fixation solution was added (final concentration 2% PFA) at 25°C for 60 min. The culture was centrifuged and the fixative was removed by suction. Resuspended cells were soaked in penetrant for 2 min on ice. The inducer IPTG was added prior to fluorescence intensity measure and the samples were analyzed by flow cytometry as described (Vilchèze et al., 2013).

Mycobacterium smegmatis cultures undergone nutrition starvation for 48 h were pelleted centrifugation at 8,000 rpm (7,990 g) for 10 min and then resuspended with M9 medium without glucose. Resuspended samples were cultured with or without 2 mM glutamine for 24–48 h. The ATP level was measured by using a BacTiter Glo kit (Promega) according to the instructions of manufacture. Fifty microliter Bactiter-Glo Reagent was added to 50 μL culture media. The contents were briefly mixed by shaking and incubated for 5 min. The background ATP was subtracted by using media alone. The luminescence was record with a fluorescence microplate reader.

To test whether glutamine can alter the M. smegmatis persister sensitivity toward the antibiotic, the survival rate of planktonic M. smegmatis exposed to rifampicin (145.8 μM) in the presence/absence of glutamine (0.5–10 mM) was compared. The result showed that glutamine combined with rifampicin inhibited bacterial growth, and this effect was dose-dependent, with a significant inhibition at a range of 2–10 mM (the survival rate after 72 h was reduced 68-, 136-, and 235.52- fold at 2, 5, and 10 mM) (Figure 1A). To determine the bactericidal potency against the starved strains, rifampicin at concentrations ranged from 48.6 to 291.6 μM was added to the starved cultures, which produced survival rate at 4.68-, 14.02-, and 17.89- fold corresponding to rifampicin concentrations at 48.6, 145.8, and 291.6 μM, respectively. The results showed that glutamine can boost the efficacy of rifampicin against M. smegmatis in a dose-dependent way (Figure 1B).

To assess whether glutamine can alter the sensitivity of M. smegmatis within the biofilm against antibiotics, the biofilm was subjected to rifampicin with or without glutamine. As expected, the survival rate of Mycobacteria within biofilm was reduce to 10 and 1% upon exposure to rifampicin at 48.6 and 145.8 μM (Figure 1C).

To investigate whether glutamine enhanced rifampicin efficacy was due to its role as nitrogen source, we treated the starved cells with various amino acids, such as valine, threonine, serine, alanine, arginine, glutamate, asparagine, and glutamine. We found that asparagine, threonine, glutamate, and valine failed to synergize the rifampicin activity (Figure 2A), though that glutamate is the deaminated product of glutamine (Figure 2C). However, arginine (14.45- fold) (Figure 2C), alanine or serine can slightly enhance rifampicin efficacy. To further explore the effect of nitrogen source on rifampicin efficacy, acetamide and urea were used to treat cells in combination with rifampicin. Both agents failed to enhance the rifampicin activity against M. smegmatis (Figure 2B). Thus, the synergistic effect on rifampicin is specific with glutamine.

To provide mechanistic insights into the synergistic effect of glutamine on rifampicin, the metabolome difference upon glutamine addition was explored. The contents of 10 amino acids were dramatically changed by the addition of glutamine (Figure 3), namely lysine (10.68- fold), pyroglutamic acid (8.12- fold), glutamine (5.07- fold), aspartate (2.94- fold), ornithine (1.69- fold), citrulline (1.72- fold), glutamate (1.33- fold), and arginine (2.01- fold). The most significantly changed amino acids were the degraded products of cell nutrients. Figure 3 showed the change of key genes involved in amino acids metabolism, such as lys1 (3.07- fold) underlying the lysine conversion into citrate via condensing with an oxaloacetate and acetyl-CoA, glsA (MSMEG_3818) (28.38- fold) responsible for the transformation of glutamine to glutamate as well as arginine, mhpB (10.03- fold) catalyzing the phenylalanine degradation into fumarate, methylcitrate dehydratase (MSMEG_6645) (4.50- fold), carboxyvinyl-carboxyphosphonate phosphorylmutase (MSMEG_2506) (4.06- fold) and MSMEG_6855 (6.75- fold) underlying the succinate formation from the asparate degradation.

To investigate whether glutamine addition can accelerate the TCA cycle, the intracellular ATP level was measured by fluorescence microplate reader with the starved cultures in the presence/absence of 2 mM of glutamine, glutamate, or arginine. Intracellular ATP level increased 2.44- fold (Figure 5A) at 72 h after glutamine was added. However, the addition of glutamate or arginine caused little increase in the intracellular ATP content (Figure 5A). To explore the effect of glutamine on the NAD⁺/NADH ratio of bacteria treated with rifampicin/glutamine, we monitored NAD⁺ and NADH concentrations in M. smegmatis treated with rifampicin alone or in combination with glutamine. The rifampicin exposure can increase NAD⁺/NADH ratio, while rifampicin–glutamine combination can lower the NAD⁺/NADH ratio (Figure 5B, left panel). The change fold of NAD⁺/NADH ratio from 24 to 48 h was determined (Figure 5B, right panel). As expected, rifampicin and glutamine exposure can increase ROS level (Figure 5C). The increased DNA damage also corroborated the higher level of ROS. The combination of rifampicin and glutamine resulted in more M. smegmatis DNA damage han rifampicin alone, as revealed by the fluorescent probe specific to the DNA damage. The DNA damage level increased at 72 h (Figure 5D). The combination of rifampicin and cysteine showed negligible synergy against the starved culture (Supplementary Figure 1). Taken together, the results implicated that glutamine might synergize the efficacy of rifampicin against M. smegmatis by increasing the production of endogenous ROS.