Antibiotics Are Man's Greatest Invention

Comparative fitness analysis of D-cycloserine resistant mutants reveals both fitness-neutral and high-fitness cost genotypes


Unless otherwise stated, all the chemicals and reagents used in this study were purchased from Sigma-Aldrich.

Bacterial strains and conditions

Mycobacterium tuberculosis laboratory strain H37Rv was routinely cultured in Middlebrook 7H9 broth (Difco) supplemented with either 10% (v/v) of ADC enrichment (Difco), 0.05% (v/v) tyloxapol, and 0.02% (v/v) glycerol or with 10% of (v/v) ADGNT solution (0.5% bovine serum albumin, 2% dextrose 2% glycerol, 0.85% NaCl, and 0.4% tyloxapol) (c7H9). Liquid cultures were grown at 37 °C either in 50 mL centrifugation tubes with rotation of 40 rpm or in roller bottles with a rotation of 2 rpm. For solid growth experiments, M. tuberculosis H37Rv was grown in Middlebrook 7H10 (Difco) or Middlebrook 7H11 (Difco) agar medium supplemented with 10% (v/v) of OADC enrichment (Difco) and 0.5% (v/v) glycerol (c7H10 or c7H11).

Calculation of DCS mutation rates

Mutation rates were calculated using the fluctuation assay and the Poison distribution methodology with few modifications48. Briefly, M. tuberculosis, starter cultures were inoculated from freezer stocks until the culture reached an OD600 of 1 and ~300,000 cells were used to inoculate 120 mL of c7H9, giving a total cell count of 10,000 cells per 4 mL culture. This volume was immediately divided to start 30 cultures of 4 mL each in 15 mL centrifugation tubes. Cultures were grown at 37 °C with rotation of 40 rpm until saturation (around 4 weeks). Following this, 26 cultures were spun at 3220 × g for 10 min at 4 °C. Cultures were then resuspended in 250–500 μL of 7H9 media and spotted onto multiwell plates containing c7H10 agar medium supplemented with the drug under investigation (DCS at 64 and 100 µg mL−1, RIF at 2 µg mL−1, and INH at 1 µg mL−1). Once spread plates were allowed to dry and were subsequently incubated at 37 °C for 1 month. Cell counts were determined by serial dilution of four remaining random cultures. Following that, the mutation rates can be calculated using the following formulas:

For the proportion (P0) of the cultures with no mutants

$$P_0 = frac{{{mathrm{no.}};{mathrm{of}};{mathrm{cultures}};{mathrm{with}};{mathrm{no}};{mathrm{mutants/growth}}}}{{{mathrm{total}};{mathrm{no.}};{mathrm{of}};{mathrm{cultures}},left( {mathrm{i.e.}}, 26 right)}}$$


For the number of mutation per culture (m)

$$m = – ln left( {P_0} right)$$


Finally for the mutation rate (μ)

$$mu = frac{m}{{{mathrm{total}};{mathrm{CFUs}}}}$$


Isolation of spontaneous D-cycloserine-resistant mutants

A fresh culture of M. tuberculosis H37Rv was set up in 5 ml c7H9 and incubated shaking at 37 °C until mid-exponential phase. The optical density was recorded and the culture diluted to a final OD600 of 0.000005 (~500 cells mL−1) in a total volume of 200 mL c7H9. Aliquots were taken and spread on 7H11 plates for CFU analysis of initial culture concentration. In total, 4 mL aliquots were then transferred into 60 individual 15 mL sterile polystyrene culture tubes, and incubated with rotation at 37 °C until saturation (~1 month). Aliquots were taken from select tubes and plated for CFU analysis of final culture concentrations. Tube contents were then inoculated, in full, onto individual 47 mm 0.22 -µm nitrocellulose membrane filters by vacuum, and the resulting bacteria-laden filters were transferred to individual c7H10 plates containing DCS at 100 µg mL−1 (5 × MIC). Plates were incubated at 37 °C, with filters transferred to fresh plates every 8–9 days for a period of 5 to 7 weeks. Individual colonies were then picked and inoculated into fresh c7H9, and freezer stocks generated. In addition, potential mutants were also spreaded on c7H10 plates containing 100 µg mL−1 of DCS to verify heritable acquisition of DCS resistance.

Minimum inhibition concentration assays

Cultures were inoculated to a final OD600 of 0.001 into 100 µL of 7H9 in 96-well microtiter plates, with twofold serial dilutions of test compound going down each row (final column left without any drug). Top concentrations were chosen on a per-drug basis. Plates were incubated at 37 °C for 6 days, and then a solution containing 0.02% of resazurin dye was added on all the wells and incubated for another day. Minimum inhibition concentrations (MIC90) was calculated using visual determination of the lowest concentration of drug at which there was no change in the resazurin color.

Determination of MICs were also performed using the automated BD BACTEC™ MGIT™ 960 Instrument. Positive MGIT tubes containing the DCSR and parent wild-type strains were used to inoculate new tubes containing DCS at final concentrations of 0, 2, 4, 8, 16, 32, 64, and 128, µg mL−1, in duplicates. In the MGIT, the lowest concentration of DCS that prevented the generation of florescence from the DCS-containing tube within 2 days of the no drug control tube was defined as the MIC.

Whole-genome sequencing

In all, 2 mL of stationary phase cultures of desired strains were harvested by centrifugation, resuspended in 0.2 mL TE buffer, and heat-killed at 80 °C for 1 h. Genomic DNA was extracted using the Zymo gDNA extraction and genomic DNA clean and concentrator kits, following the manufacturer’s protocols. Quality and quantity of genomic DNA were confirmed by agarose gel electrophoresis, and DNA sequencing performed on an Illumina HiSeq 2000 instrument at the Cricks Genomic Science Technology Platform.

Mapping genome data and SNP calling

Sequencing reads for the WT and the 11 DCSR mutant strains were mapped against the M. tuberculosis H37Rv reference genome (AL123456) as single-end data using the Burrows-Wheeler algorithm described in BWA49 (Supplementary Data 1). Mapping outputs were used to generate SNP calls using SAMtools49. Filtering of SNP calls was performed by keeping those SNPs with minimum mapping quality of 10 and maximum read depth of 400. These high confidence calls were used to generate a nonredundant list of variable positions called in at least one strain and used to recover the base call in all other strains. Finally, SNPs positions that involved heterozygous calls or felt in a repetitive or mobile element were removed. Functional annotation of each of the polymorphic positions identified was performed using ANNOVAR50. The SNPs identified are listed in Supplementary Data 2. All genome data is available in ArrayExpress under accession number E-MTAB-5935.

Sanger sequencing

DNA fragments including the identified SNPs were amplified via PCR with specific primers outlined in Supplementary Data 4, PCR amplicons were then purified using a Qiagen PCR clean up kit and then sent for Sanger sequencing at GATC Ltd (Germany). Sequencing files were visualized and compared to wild-type using the Unipro UGENE software.

Competition assays

M. tuberculosis H37Rv parent strain and all DCS-resistant mutants were grown at 37 °C in c7H9 media until late logarithmic phase separately in roller bottles. The cultures were then diluted and a new set of cultures initiated by co-culturing together the parent strain and each of the DCS-resistant mutant at an equal ratio (2 × 105 CFUs) and incubated over the course of the competition assay for 56 days. An aliquot of the mixed culture was taken at the following timepoints, 0, 7, 14, 28, and 56 days, serially diluted and plated in triplicate in c7H10 plates with and without 30 µg mL−1 of DCS for CFU determination. CFUs were counted following 1 month of incubation at 37 °C. Three independent experiments were performed with duplicate cultures in each of them. The relative fitness of the DCS susceptible over the DCS-resistant strains was calculated using the formula:

$$W = {mathrm{ln}}(frac{{R_e}}{{R_b}}) div {mathrm{ln}}(frac{{S_e}}{{S_b}})$$


Where Re and Se are the number of resistant and susceptible bacilli at the endpoint (7, 14, 28, and 56 days), respectively, and Rb and Sb are the number of resistant and susceptible bacilli at the baseline, time 0.

Infection of human monocyte-derived macrophages

White blood cells were isolated from leukocyte cones (NC24) supplied by the NHS Blood and Transplant service, UK and complied with the UK Human Tissue Act regulations. GM-CSF-derived macrophages were plated in 24-well tissue culture plates at a final concentration of 2 × 105 cells per well 1 day prior to infection. M. tuberculosis cultures were grown to the mid-log phase (OD600 ~0.8). Prior to infection, M. tuberculosis cultures were harvested by centrifugation at 3000 rpm for 5 min and washed once with PBS and once with cell culture media (RPMI with 10% FCS). An equal volume of sterile glass beads (2.5–3.5 mm) that matched the pellet size was added and vigorously shaken for 1 min to make single-cell suspension. The bacteria were then resuspended in cell culture media and spun at 1200 rpm for 5 min. Following centrifugation, the supernatant containing single-bacterial cells was transferred into a new tube, OD600 was measured and the culture was diluted accordingly to 0.1 OD to make the infection solution (~8 × 108 cells mL−1). Macrophages were then washed and suspended with 0.25 mL per well of the infection solution (MOI of 1) for 2 h. Following infection, macrophages were then washed three times with PBS and suspended with cell culture media for the course of infection.

CFU determination

For counting bacterial viability, macrophages were washed once with PBS to remove extracellular bacteria and lysed with 0.5 mL of water–Tween-80 (0.05%) per well for 1 h at room temperature. The lysed solution from the triplicate wells was then used for serial tenfold dilutions in PBS–Tween-80 (0.05%). In all, 20 µL from each dilution was then plated in triplicate onto c7H11 agar plates. Agar plates were incubated for 3–4 weeks at 37 °C. CFU counts were calculated and plotted as the mean CFU per mL.

Mouse infections

C57BL/6J (WT) mice were bred and maintained under specific pathogen-free conditions at The Francis Crick Institute, Mill Hill Lab or purchased from Charles River. All studies were ethically reviewed and approved by the respective ethical review committees at the Francis Crick Institute and NIBSC. Procedures involving mice were performed in strict accordance with the United Kingdom Animals (Scientific Procedures) Act 1986 and the respective Institutes policies on the Care, Welfare and Treatment of Animals. Procedures were done under UK Home Office animal licenses 70/8045 and P7611793C. Animals were monitored by trained animal technicians at least once a day, and this frequency could increase if any adverse reactions were observed. Infections were performed in the containment level 3 animal facilities either at the Mill Hill Lab or NIBSC. For mouse infections, M. tuberculosis H37Rv and DCS-resistant strains were cultured in c7H9 broth, to an OD600 of 0.6. From this, an infection sample was prepared to enable delivery of about 100 CFUs/mouse lung using a nebulizer system (Walkers, UK) linked to a Middlebrook airborne infection device (Glas-col, Terre Haute, USA). Infection was monitored by assessing homogenized lungs from infected mice at defined time intervals. Bacterial counts were determined by plating serial dilutions of homogenates on duplicate c7H11. CFUs were counted 2–3 weeks after incubation at 37 °C. The data at each timepoint are the means of five mice per group ± standard error (SEM).

RNA extraction and quantitative qPCR

Early logarithmic phase cultures of M. tuberculosis and DCS-resistant mutants were grown in triplicates and then either treated with 10 × MIC of DCS or carrier control for 4 h. Following treatment, cultures were spun down at 3000 × g for 5 min at 4 °C, supernatant discarded and suspended in 1 mL of Trizol reagent and transferred into a 2 -mL tube containing lysing matrix B beads (MP Biomedicals). Samples were lysed using a ribolyser at 6.5 speed for 45 s twice. Following ribolysis, samples were spun down, and the RNA was purified using the Direct-zolTM−96 RNA kit (Zymo Research) following the manufacturer’s instructions with the on column DNase treatment. Following RNA purification, the RNA was further DNase treated by a vigorous treatment using the Ambion DNA Free kit according to the manufacturer’s instructions. RNA quantity and quality was estimated using NanoDrop and 1.5 µg of RNA was used to make cDNA according the High-Capacity RNA-to-DNA kit (Applied Biosystems). qPCRs were done using TaqMan probe assays to quantify the levels of a selection of genes (Supplementary Data 4) on a QuantStudio 5 Real-Time PCR System. PCR efficiencies and limit of detection were calculated for each TaqMAn assay using genomic DNA from M. tuberculosis. Differential gene expression was calculated using the cloud-based AB software.

Western blotting

Early logarithmic phase cultures of M. tuberculosis and DCS-resistant mutants were grown in and either treated with 10 × MIC of DCS or carrier control for 4 h. Following treatment, the cultures were spun down at 3000 × g for 5 min at 4 °C, supernatant was discarded, and the pellet was washed twice with PBS solution and suspended into 1 mL of protein lysis buffer (20 mM TAE pH 7.8, 8 M urea, 10% glycerol and 5 mM TCEP (Tris[2-carboxyethyl]phosphine hydrochloride)) and transferred into a 2- mL tube containing lysing matrix B beads (MP Biomedicals). Bacterial suspensions were lysed using a ribolyser at 6.5 speed, 45 s, three times with intermittent cooling on ice, followed by centrifugation at top speed for 10 min in a refrigerated microcentrifuge. The supernatants containing soluble protein fractions were then passed through a 0.22 µm pore-size spin filter, and stored at −80 °C until further use. Thirty micrograms of protein of each sample were run per lane of a SDS-PAGE gel alongside Odyssey® One-Color Protein Molecular Weight Markers (Li-Cor Bioscience) and blotted to polyvinylidine fluoride membranes using iBlot 2 Dry Blotting System (ThermoFisher Scientific). Membranes were blocked using the Odyssey Blocking buffer (Li-Cor Bioscience) for 2 h at room temperature and subsequently probed with primary rabbit-raised anti-MtAlr antibody (1:5000 dilution) overnight at 4 °C. Following washing, membranes were incubated for 1 h with IRDye 800CW Goat anti-Rabbit IgG (Li-Cor Bioscience) secondary antibody, and the immunoblots were developed using the Odyssey CLx Imaging System (Li-Cor Bioscience).

Recombinant Alr purification

MtAlr D322N was generated using QuickChange mutagenesis (Agilent; cloning primers; Supplementary Data 4), and pET28:MtAlr as a template. Recombinant proteins (including Escherichia coli L-alanine dehydrogenase; EcLADH) were expressed in E. coli BL21(DE3) and purified by nickel-affinity chromatography51. Both enzymes were co-expressed with the mycobacterial and E. coli GroEL and GroES chaperone proteins, respectively. Purified histidine-tags were removed using thrombin (Restriction grade, GE Healthcare) and dialyzed three times to remove the cleaved tags in 20 mM TEA or 20 mM NaPO4, 150 mM NaCl, pH 7.8, concentrated, flash frozen, and stored at −80 °C until activity assays.

MtAlr activity measurements

Alanine racemase activity was measured in the L-Ala to D-Ala direction51. Briefly, reactions containing 50 mM buffer (see below), 5 mM α-ketogluratate, 0.25 mM NADH, 18–28 U ml-1 LDH (used as a mixture of pyruvate kinase/lactate dehydrogenase for convenience; Sigma), 4 μM BsDAAT, and varying concentrations of L-Ala were combined with 10–100 nM recombinant enzyme and monitored at 340 nm. Buffers used for each pH were HEPES (7.62, 7.12), Tris (7.92), TAPS (8.18), MOPS (6.86), CHES (8.9), PIPES (6.57), and MES (6.3). All reactions were performed at 37 °C. Km and kcat values were calculated by fitting data to the Michaelis–Menten equation.

DCS inactivation reactions comprised 5 µM recombinant Alr, 20 mM sodium phosphate buffer (pH 7.5 or 6.85), 150 mM NaCl, and varying concentrations of DCS. Reactions were commenced upon addition of DCS. At distinct timepoints during the incubation (including a time 0, before DCS was added), aliquots were removed and added to activity reactions, comprising 50 mM CHES pH 8.9, 5 mM NAD+, 20 mM D-Ala, and 30 µL mL−1 EcLADH. Activity reaction progression was monitored at 340 nm.

Activity at time t was re-calculated as a percentage of the initial activity of the enzyme (activity at t0). Plots of ln(% activity) vs. time were linear in all cases. The slope of each plot allowed calculation of the kobs for each inhibitor concentration tested, per Eqs. (1) and (2):

$$V(t)/V(0) = a times e^{left( { – k_{obs} times t} right)}$$


$$lnleft( {V(t)/V(0)} right) = k_{obs} times t + a$$


Where a is a Y-axis normalization factor, V(t) is activity at time t, and V(0) is activity at time 0. Nonlinear regression analysis of replots of kobs vs. inhibitor concentration allowed calculation of kinact and Ki, per Eq. (3):

$$k_{obs} = left( {k_{inact} times left[ I right]} right)/left( {K_i + left[ I right]} right)$$


Lipid extraction

Early logarithmic phase cultures of M. tuberculosis and DCS-resistant mutants were grown for lipid profiling. Apolar lipids were extracted following established procedures as previously described52. Briefly, 20 mL of M. tuberculosis H37Rv and DCS-resistant strains were cultured in c7H9 broth, to an OD600 of 2, harvested via centrifugation at 3000 × g for 5 min and transferred to 1 mL of PBS solution in a 2 mL screw cap tubes and heat inactivated at 90 °C for 2 h. Lipids were extracted from the remaining biomass in a mixture of methanol: chloroform (2:1) overnight, followed by a second extraction of methanol: chloroform (1:1) for 5 h at room temperature. Pooled lipid mixture was then dried at 50 °C under a nitrogen stream. Extracted lipids were dissolved in chloroform and subsequently analyzed using thin layer chromatography. The extracted lipids were then analyzed using two-dimensional TLC with system A solvents consisting of petroleum ether–ethyl acetate (98:2), three times in the first direction and petroleum ether–acetone (98:2), once in the second direction. TLC plates were visualized using a 5% phosphomolybdic acid solution in ethanol followed by gentle charring of the plates with a heat-gun.

Statistical analysis

GraphPad Prism 7 was used for all statistical analysis. Three independent experiments per assay were performed and statistical analysis was carried out by unpaired t test of the biological replicates of each experiment.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.