Bacteriostatic antibiotics promote the evolution of CRISPR-Cas immunity

No statistical methods were used to pre-determine sample size. The experiments were not randomized, and investigators were not blinded to allocation during experiments and outcome assessment.

Bacteria, phages and growth conditions

Except when stated otherwise, evolution experiments and phage assays used P. aeruginosa UCBPP-PA14 (PA14) and lytic phage DMS3vir11. UCBPP-PA14 csy3::lacZ was used for phage stock amplification, phage titre determination and estimation of Cas protein expression. Competition experiments used a surface mutant (3A) derived from PA14 csy3::lacZ and a CRISPR-resistant mutant (BIM4, bacteriophage insensitive mutant with 2 additional acquired spacers against DMS3vir) derived from PA14, both of which have been previously described38. DMS3vir and a mutant expressing anti-CRISPR against PA14 IF system, DMS3vir AcrIF1, were used for determination of resistance phenotypes13. Evolution experiments in the presence of Chl also used PA14-cat, a Chl-resistant mutant of PA14 carrying the cat gene inserted into the genome using a variant of plasmid pBAM39 carrying cat. To this end, the cat gene was amplified from plasmid pKD340 using primers TAGATTTAAATGATCGGCACGTAAGAGGTT and CTGACCCTTGTCTTACGCCCCGCCCTGCCACT, then ligated into pBAM1 after digestion with SwaI and PshAI. For microfluidics experiments, we used PA14 flgK::Tn5B30(TcR)41. All bacterial strains were grown at 37 °C in LB broth or M9 medium (22 mM Na2HPO4; 22 mM KH2PO4; 8.6 mM NaCl; 20 mM NH4Cl; 1 mM MgSO4; and 0.1 mM CaCl2) supplemented with 40 mM glucose, glycerol or pyruvate. All liquid cultures were grown with 180 rpm shaking. For experiments using M9, overnight pre-cultures were themselves grown in M9 with the same carbon source.

Determination of antibiotic activity

For MIC (minimum inhibitory concentration) determination, overnight cultures (∼5.109 cells/mL) were diluted 104>-fold in LB medium. 20 μL of the diluted cultures were inoculated into 96-well microplate wells containing 180 μL of LB supplemented with antibiotics using 2-fold serial dilutions of the antibiotic. After 18 h growth at 37 °C, MIC was determined as the lowest antibiotic concentration with no visible growth. To determine the MBC (minimal bactericidal concentration), the content of wells with no visible growth was plated on LB-agar and further incubated overnight. MBC was defined as the lowest antibiotic concentration resulting in 99.9% decrease in initial inoculum cell density (< 5 CFU in 100 μL). MBC/MIC ratio was used to estimate if antibiotic activity was bacteriostatic or bactericidal: a high MBC/MIC ratio indicates that the concentration sufficient to prevent growth is much lower than the concentration required to kill the majority of cells42. In our assay, antibiotics with average MBC/MIC ratio >1 were the ones that are commonly recognized as being bacteriostatic (Tm, Erm, Chl and Tc).

Evolution experiments

Evolution experiments were performed in glass vials containing 6 mL growth medium and appropriate antibiotics at the concentrations shown in Extended Data Fig. 1. 60 µL from overnight cultures were co-inoculated with 104 plaque-forming units (p.f.u.) of phage DMS3vir, with the exception of the experiment in Extended Data Fig. 10, where two different phage inocula of 104 and 109 p.f.u. were used. 1:100 volume was then transferred every 24 h into fresh medium for 3 days with the exception of the experiment in Extended Data Fig. 10, which was carried out for 5 days. Each treatment contained 6 biological replicates. Cell densities and phage titers were monitored daily with serial dilution in M9 salts (after chloroform treatment for phages), and enumeration of colonies on LB-agar and enumeration of plaques on a lawn of PA14 csy3::lacZ cells. The identification of phage resistance type (sensitive, CRISPR-Cas or SM) was performed by cross-streaking 24 randomly selected colonies on DMS3vir and DMS3vir-AcrF1 phages: SM clones are resistant to both phages and have a characteristic smooth colony morphology, whereas clones with CRISPR-Cas immunity are resistant to DMS3vir but sensitive to DMS3vir-AcrF113.

Determination of bacterial growth rate by optical density

Overnight cultures were diluted 100-fold into fresh growth media. Growth of 200 μL of culture was measured in a 96-well plate by measuring optical density at λ=600nm (OD600) for 14 to 24 h at 37 °C in a BioTek Synergy 2 Plate reader, with 5 s shaking before each measurement. All growth curves were performed in at least 8 replicates. Exponential growth rate was determined in R using the package growthrates43.

Determination of bacterial doubling time by microfluidics

The mother machine device was fabricated and handled as previously reported15,16. Briefly, overnight cultures in LB were spun down via centrifugation for 5 minutes at 4000 rpm at room temperature (Eppendorf 5810 R). The supernatant was filtered twice (Medical Millex-GS Filter, 0.22 µm, Millipore Corp.) and used to re-suspend the bacteria to an OD600 of 75. 2 µl of the bacterial suspension was injected into the microfluidic mother machine device and incubated at 37 °C until there were 1-2 bacteria in the lateral side channels. Fluorinated ethylene propylene tubing (1/32” × 0.008”) was connected to the inlet and outlet holes and connected to a computerized pressure-based flow control system (MFCS-4C, Fluigent) controlled by MAESFLO software (Fluigent) and outlet reservoir respectively. Spent media was flushed through the device to wash excess bacteria out of the main channel at 300 µL/h for 8 minutes to completely exchange the fluid in the device and tubing. The chip was mounted on an inverted microscope (IX73 Olympus, Tokyo, Japan) and images were acquired in bright-field via a 60×, 1.2 N.A. objective (UPLSAPO60XW, Olympus) and a sCMOS camera (Zyla 4.2, Andor, Belfast, UK) with a 0.03s exposure. The microfluidic device was moved by two automated stages (M-545.USC and P-545.3C7, Physik Instrumente, Karlsruhe, Germany, for coarse and fine movements, respectively) to image multiple fields of view in a sequential manner. The imaging setup was controlled by LabView. After acquiring the first set of images, we flowed each of the investigated antibiotics dissolved in LB at the appropriate concentration at 300 µL/h for 8 minutes before lowering the flow rate to 100 µL/h for 3 hours. The entire assay was carried out at 37 °C in an environmental chamber surrounding the microscope. Bacterial doubling times were extracted from the acquired image sets as previously reported17. Briefly, we tracked each individual bacterium and its progeny throughout each experiment and doubling times were measured as the lapses of time between successive bacterial divisions that were assessed by eye through the images loaded in ImageJ and considered to have happened when two daughter cells became clearly distinguishable from their respective parental cell.

One-step phage growth assays

Overnight cultures of PA14 were first diluted into 6 mL growth medium ± antibiotic treatment in glass vials (N=4). For experiments with 30 min of pre-exposure to antibiotics, cells were diluted 25-fold into fresh media with antibiotics and grown for 30 min before phage addition. For experiments with 12 h of pre-exposure to antibiotics, cells were diluted 100-fold into fresh media with antibiotics and grown for 12 h before phage addition. Bactericidal treatments were excluded from further analysis because they caused a 4-fold to 570-fold decrease in cell density after 12 h, making it impossible to determine the latent period of phage under those conditions. After growing in the presence of antibiotics, approximately 5.107 p.f.u. of DMS3vir were added in each vial, and vials were vortexed and incubated at 37 °C for 15 min, allowing phage adsorption. Cultures were then diluted 1000-fold into 6 mL growth medium ± antibiotic treatment to limit further adsorption and re-infection, vortexed again and transferred to 24-well plates for parallel processing. Samples were taken immediately (t=0) and then approximately every 20 minutes. The first samples were diluted in M9 salts and plated on LB-agar to quantify cell densities; all samples were chloroform-treated and plated on PA14 csy3::lacZ lawns. Phage densities measured after chloroform treatment correspond to the sum of free phages and mature phage particles inside infected cells.

Determination of antibiotic effects on infection success

Four overnight cultures of PA14 were diluted in parallel 100-fold into LB with or without antibiotics. After 2 h growth at 37 °C, DMS3vir phages were added to a final concentration of 1000 p.f.u./mL (equivalent to 5 p.f.u. in 200 μL) and the vials were vortexed. After 15 min at 37 °C, vials were vortexed again and 24* 200 μL of each individual culture were aliquoted into 24 wells of a 96-well plate. Plates were incubated at 37 °C for 22 h, then 20 μL of each well were spotted on a lawn of PA14 csy3::lacZ cells in two replicates. With an average phage inoculum of 5 phages, the distribution of phages across wells is expected to follow a Poisson distribution with 0.7% wells containing 0 phages and 1.3% wells containing more than 10 phages. The control treatment with no antibiotics was consistent with this, as 1 in 96 wells produced no lysis. Lysis indicated that the founding phages reproduced. The number of wells in which phages failed to reproduce was counted for each treatment, and significance was determined by chi-square tests between antibiotic and no-antibiotic treatment.

Measurement of mutation towards SM

To evaluate the frequency of SM cells in the absence of phage selection, cells were grown in LB ± antibiotic treatment for 24 h. After 24 h, cultures were serially diluted in M9 salts, then dilutions were plated both on LB-agar to calculate total cell density, and on LB-agar containing a high concentration of DMS3vir, which was generated by covering the agar surface with a phage stock of 108 p.f.u./μL. The density of SM mutants was calculated from counting the number of colonies growing on top of DMS3vir. Three independent experiments were run with 6 experimental replicates each.

Spacer acquisition assay

20 μL of PA14 overnight culture were first diluted 1:50 into 1 mL LB with or without antibiotics in 24-well plates, in 8 replicates per treatment. After 30 min of growth at 37 °C, 2.109 DMS3vir phages were added per well, and cultures were incubated at 37 °C for 3h. The density of phage-sensitive cells was measured by plating 100 μL on LB-agar after 104-fold dilution in M9 salts. The density of phage-resistant cells was measured by directly plating 100 μL of cultures on LB-agar without dilution: the phage density on these plates was sufficient to prevent growth of sensitive colonies. The majority of colonies had a smooth morphology characteristic of SM clones. We confirmed that smooth colonies were resistant to both DMS3vir and DMS3vir+AcrIF1, whereas non-smooth colonies were resistant to DMS3vir but sensitive to DMS3vir+AcrIF1, and were therefore CRISPR immune. In each culture, the proportion of CRISPR-Cas immune clones within the total population of resistant clones (CRISPR-Cas and SM) was calculated.

Competition assays

Competition experiments were performed in 6 mL LB supplemented in the presence or absence of antibiotics. They were initiated by inoculating 60 μL of a 1:1 mix of LB overnight cultures of CRISPR-Cas immune (BIM2) and surface mutant (3A) clones. For treatments including phages, 8.109 p.f.u. DMS3vir were added per vial. Samples were serially diluted at 0 and 24 h and plated on LB agar supplemented with 50 μg/mL X-Gal, to determine the ratio of the surface mutant that carries the LacZ gene and therefore forms blue colonies, and the BIM2, which forms white colonies. The selection rate of the CRISPR-Cas clone was calculated as mBIM2-m3A, with m the Malthusian parameter defined as log(density(t1)/density(t0)) 44. We used selection rate rather than relative fitness because some treatments led to an absolute decline in the abundance of the CRISPR-Cas clone.

Cas expression assay

We used PA14 csy3::lacZ as a reporter strain for Cas gene expression, using the β-Galactosidase fluorogenic substrate 4-Methylumbelliferyl β-D-galactoside45 (MUG). An overnight culture of PA14 csy3::lacZ was diluted 100-fold into 6 mL of LB with or without antibiotics. After 5 h of growth, OD600 was recorded and 100 μL aliquots were immediately frozen at -80 °C. Prior to the assay, the frozen 96-well plate was defrosted and 10 μL were transferred to a new plate and frozen again at -80 °C for 1 h. After transfer to 37 °C, 100 μL reagent solution (0.25 mg/mL MUG and 2 mg/mL lysozyme in phosphate-buffered saline) was added to each well. Fluorescence was measured for 30 min in a Thermo Scientific Varioskan flash plate reader at 37 °C, with excitation and emission wavelengths respectively 365 nm and 450 nm. The 15 min timepoint was used for analysis. Relative fluorescence was calculated as (fluorescence at 15 min – fluorescence at 0 min) / OD600.


All statistical analyses were done with R version 3.4.1 46, and package cowplot47. Evolution experiments with antibiotics were not all performed simultaneously (Figure 1, Extended Data Fig. 1, Extended Data Fig. 3): for these, we used individual Student t-tests comparing each treatment to the associated no-antibiotic treatment. For experiments in M9 that used two different phage inocula (Extended Data Fig. 10), the effect of the two treatments was modelled as: proportion of CRISPR-Cas immune clones ∼ carbon source + phage inoculum.

Article TitleBacteriostatic antibiotics promote the evolution of CRISPR-Cas immunity


Phage therapy can be used in combination with antibiotics to combat infections with bacterial pathogens13. However, bacteria can rapidly evolve phage resistance via receptor mutation, or using their CRISPR-Cas adaptive immune systems4, which insert short phage-derived sequences into CRISPR loci in the bacterial genome5 to guide sequence-specific cleavage of cognate sequences6. Unlike CRISPR-Cas immunity, mutation of the phage receptor leads to attenuated virulence when the opportunistic pathogen Pseudomonas aeruginosa is infected with its phage DMS3vir7, which underscores the need to predict how phage resistance evolves under clinically relevant conditions. Here, using eight antibiotics with various modes of action, we show that bacteriostatic antibiotics (which inhibit cell growth without killing) specifically promote evolution of CRISPR-Cas immunity in P. aeruginosa by slowing down phage development and providing more time for cells to acquire phage-derived sequences and mount an immune response. Our data show that some antimicrobial treatments can contribute to the evolution of phage-resistant pathogens with high virulence.

Login or Signup to leave a comment
Find your community. Ask questions. Science is better when we troubleshoot together.
Find your community. Ask questions. Science is better when we troubleshoot together.

Have a question?

Contact or check out our support page.