Abstract
Rapid antibiotic susceptibility testing (AST) is urgently needed for treating infections with correct antibiotics and slowing down the emergence of antibiotic resistant bacteria. Current clinical methods reply on culture and take at least 16 h. Here, using P. aeruginosa, E. coli and S. aureus as models, we show that the AST can be finished in 10 minutes by stimulated Raman scattering (SRS) imaging of D2O metabolic activities. The metabolic incorporation of D2O, which is used for biomolecule synthesis, can be monitored in a single bacterium. Time lapse experiments show that the C-D vibrational signal can be observed in a single bacterium within 10 minutes culture in D2O medium. Since water is universally used for biosynthesis in bacteria, SRS imaging of D2O metabolism has the potential to be generalizable to different bacteria species.
Antimicrobial resistance has become a growing public threat, causing nearly 1 million deaths from drug-resistant infections each year globally1. It was estimated that by 2050, antimicrobial resistance will cause 10 million deaths and $100 trillion GDP loss if no action is taken1,2. To combat this crisis, new rapid diagnostic methods for antibiotic susceptibility testing (AST) are essential to reduce the deaths caused by drug-resistant infections3, and slow down the emergence of antimicrobial resistance. The culture-based phenotypic method remains the gold standard for AST. However, this method is too slow for guidance of immediate decision for infectious disease treatment. Genotypic methods, which detect known resistance genes, can provide faster results, but are not generally applicable to different bacteria and mechanism of resistance4.
To overcome these limitations, other methods have been developed for rapid AST, including, for example, microfluidic devices that increase the detection sensitivity by confining the sample in a small area5–10, imaging-based methods that monitor the growth or morphology change at the single cell level10–12, nucleic acids-based phenotypic AST quantifying nucleic acids copy number with antibiotic treatment4,13,14, and monitoring the spectral response to antibiotic treatment by Raman spectroscopy15. While these methods significantly reduce the time for AST, they either are not generalizable to different bacteria, or rely on cell proliferation, which has difficulties for viable but nonculturable species16, or require gene marker information before testing17.
We have previously demonstrated that the AST of bacteria can be accomplished within one cell cycle (30 minutes) by measuring the metabolic activity in single bacteria under a stimulated Raman scattering (SRS) microscope18. Specifically, we monitored the metabolic activity of deuterated glucose, glucose-d7, with chirped picosecond SRS at C-D vibrationa l frequency and used the C-D signal as a marker to perform AST. Like glucose, water is also ubiquitously used for biomolecule synthesis in bacteria19, and its metabolism can be selective ly probed via monitoring the conversion of heavy water (D2O) into deuterated biomolecules at C-D vibrational frequency. Unlike glucose-d7, D2O itself does not have C-D bonds, therefore providing a better contrast for SRS metabolic imaging. The metabolic activity of D2O has been used to study the metabolic - active bacteria with spontaneous Raman spectroscopy16,19. However, the speed of spontaneous Raman is limited by the weak Raman scattering process20. Compared to spontaneous Raman, SRS has orders-of-magnitude signal enhancement, thereby enabling high speed chemical imaging of single cells 21–26. SRS imaging of D2O metabolism was recently demonstrated to be a noninvasive method to visualize metabolic dynamics in mammalian cells and live animals27.
Here, we demonstrate that femtosecond SRS imaging of D2O metabolism can determine the susceptibility of bacteria within 10 minutes. We show that the metabolic activity of D2O can be monitored in single P. aeruginosa, E. coli and S. aureus. For the broad C-D vibrational spectrum, the SRS signal of bacteria can be improved by more than 5 times by femtosecond SRS compared to the chirped SRS. The D2O metabolism in bacteria responds differently as fast as 10 minutes to different antibiotics, depending on the susceptibility of bacteria. Our work shows the promise of using SRS microscopy for metabolic activity studies and rapid AST at single bacterium level.
Experimental Section
SRS microscope
A femtosecond (fs) pulsed laser (InSight DeepSee, Spectra-Physics) with an 80-MHz repetition rate and dual outputs was employed for the SRS microscope. One 120 fs laser with tunable 680–1100 nm wavelength was served as the pump beam. The other 220 fs laser centered at 1040 nm, served as the Stokes beam, was modulated by an acousto-optical modulator (AOM, 1205-C, Isomet) at ~2.4 MHz. The two beams were collinearly combined through a dichroic mirror. When spectral focusing is needed for hyperspectral SRS, both beams were chirped with two 15 cm long SF57 glass rods. After chirping, the pulse durations of the pump and Stokes laser were 1.9 ps and 1.3 ps, respectively. For implementation of SRS imaging with femtosecond pulses, the glass rods were removed from the path. The pump and Stokes beams were directed into a lab-built laser scanning microscope. A 60x water objective (NA=1.2, UPlanApo/IR, Olympus) was used to focus the lasers to the sample, and an oil condenser (NA=1.4, U-AAC, Olympus) was used to collect the signal from the sample. Two filters (HQ825/150m, Chroma) were used to filter out the Stokes beam, the pump beam was detected by a photodiode (S3994-01, Hamamatsu) and the stimulated Raman signal was extracted by a lock-in amplifier (HF2LI, Zurich Instrument).
Sample preparation
To make D2O containing LB medium, D2O was first mixed with purified water, then LB broth powder (Sigma Aldrich) was added to the solution with a final concentration of 2% in weight. The solution was sterilized by filtering. To prepare bacteria samples, E. coli, S. aureus or P. aeruginosa were cultivated in normal LB medium for 2-3 h to reach to log phase, then bacteria were diluted in a 1:100 ratio to the D2O containing LB medium. After incubation for a controlled period, 500 μl sample was centrifuged, washed twice with purified water, and deposited to an agarose gel pad.
To prepare the agarose gel pad, ~1% in weight agarose powder was added to 5 ml purified water in a plastic tube, then the tube was heated in microwave for about 20 seconds to melt the agarose powder. About 10 μl heated agarose gel solution was added to a coverglass by a pipette, another coverglass was immediately put on the top of the agarose gel solution to make it flat. After ~2 min, one coverglass was removed from the agarose gel by sliding the two coverglasses. Bacteria in solution were deposited to the gel pad, then another coverglass was put on top of the gel pad for SRS imaging.
SRS imaging
To image bacteria at the C-D region, the pump wavelength was tuned to 849 nm, and the power at the sample was ~12 mW; the Stokes wavelength was fixed at 1040 nm, and the power at the sample was ~90 mW. Each image contains 200 × 200 pixels, the pixel dwell time is 30 μs.
Spontaneous Raman spectroscopy
Bacteria in solution were deposited on a coverglass for spontaneous Raman measurement. Spontaneous Raman spectra of bacteria were acquired with an inverted Raman spectrometer (LabRAM HR evolution, Horiba scientific) with 532 nm laser source. The laser power at the sample was ~12 mW after 40x air objective, the acqusition time was 10 s. The grating was 600 l/mm.
Broth dilution method
Bacteria were cultured in D2O containing LB medium in a 96-well plate. Antibiotics, using triplicate samples, were added to the plate and serially diluted. After about 20 h incubation at 37 °C, plates were visually inspected, and the MIC was categorized as the concentration at which no visible growth of bacteria was observed.
Results and Discussion
D2O induces negligible toxicity to bacteria
We first tested whether D2O has toxicity to bacteria by measuring their growth in D2O containing medium. Three types of bacteria (E. coli, S. aureus, and P. aeruginosa) were cultured at different concentrated D2O containing LB medium, and their growth were monitored with optical density (OD) measurement at 600 nm wavelength. We found that D2O concentration up to 100% did not show significant toxicity to the growth of E. coli and S. aureus, as indicated by the growth curve in D2O media of various concentrations (Figure 1a and 1b). The growth of P. aeruginosa was initially slowed down in medium with D2O concentration of 70% and up, but eventually restored to normal growth after about 18 h for D2O concentration of 70% and 80%, and about 22 h for D2O concentration of 100% (Figure 1c). Therefore, D2O concentration of 70 % or lower in the medium does not induce significant toxicity to the bacteria.
Imaging metabolic incorporation of D2O in a single bacterium
We chose 70% D2O containing LB medium to cultivate bacteria, and used P. aeruginosa to test whether the D2O metabolic incorporation in a single bacterium can be monitored by our SRS microscope. Spontaneous Raman spectra of bacteria showed a broad peak (2070 2250 cm−1) at C-D vibrational region for bacteria cultivated in the D2O containing medium for 2 h (Figure 2a), indicating D2O had been successfully utilized for biomolecule synthesis. For control, bacteria cultivated in normal medium did not have this peak at this region (Figure 2a). To image single bacterium, bacteria were further diluted and deposited on an agarose gel pad. By tuning the Raman shift to C-D region (~2162 cm−1), a strong signal was observed for individual bacterium cultivated in D2O containing medium (Figure 2b, right). As a control, no C-D signal was observed for bacteria cultured in normal medium (Figure 2b, left). The results were confirmed by SRS spectra (Figure 2c) obtained through temporal tuning of chirped pump and Stokes femtosecond pulses. Similar results were obtained on E. coli (Supporting Figure 1) and S. aureus (Supporting Figure 2).
In order to shorten the D2O culture time, we tested whether non-chirped femtosecond pulses can enhance the signal over the chirped picosecond pulses. Because the C-D vibration band is relatively broad with a width of 180 cm−1 (Figure 2c), we hypothesized that femtosecond SRS without chirping could significantly increase the signal to noise ratio (SNR). To test this, we cultivated P. aeruginosa in 70% D2O containing LB medium for 30 minutes, and imaged them at ~2162 cm−1 by chirped picosecond pulses and non-chirped femtosecond pulses, respectively (Figure 3a and 3c). The pump and Stokes powers were adjusted to make sure the same average pump and Stokes powers were used. The SNR of individual bacterium with picosecond and femtosecond SRS was 1.43 and 7.81, respectively, indicating ~5.5 times SNR improvement with femtosecond SRS over picosecond SRS (Figure 3b and 3d). This improvement is attributed to two factors. First, the C-D vibrational band is broad, and the femtosecond SRS can detect broader band signal than the chirped picosecond SRS. Second, the pulse chirping reduced the peak power of the pulses. Although the same average power was used at the sample, the reduced peak power decreased the signal level due to the nonlinear effect of SRS.
Time lapse measurement of D2O metabolic incorporation in a single bacterium
Next, we studied the time lapse of D2O metabolic activity in single P. aeruginosa with femtosecond SRS. The P. aeruginosa was cultivated in 70% D2O containing LB medium for up to 3 h. Figure 4a shows the SRS images of single P. aeruginosa at ~2162 cm−1, C-D signal in individual P. aeruginosa can be observed after culture as short as 10 minutes. Statistical analysis showed that the average C-D signal intensity in individual bacterium increases with time, and saturates at ~1.5 h (Figure 4c), which is about three generations since the generation time of P. aeruginosa cultivated in LB medium is 24-27 minutes28. To view individual bacterium more clearly, the images in Figure 4a were further zoomed in (Figure 4b). Interestingly, in the 10 minutes result, a stronger signal was observed in the cell periphery of bacterium (Figure 4b), as indicated by the intensity plot over the bacteria (Figure 4d). In contrast, in and after 30 minutes, the signal intensity is stronger in the intracellular area (Figure 4b), as indicated by the intensity plot over bacteria in the 30 minutes result (Figure 4e). Collectively, these results suggest that D2O is initially used to synthesize cell membrane and/or cell wall in P. aeruginosa.
Rapid AST
To examine how antibiotics affect the metabolic incorporation of D2O in bacteria, and whether it can be used for rapid AST through SRS imaging, P. aeruginosa were cultiva ted in 70 % D2O containing LB medium, with the addition of 20 μg/ml gentamicin or cefotaxime. The susceptibility of P. aeruginosa were pre-determined to be susceptible to gentamicin and resistant to cefotaxime at this concentration by the conventional culture-based microdilution method. SRS imaging at ~2162 cm−1 showed that the C-D signal was significantly reduced after cultivation in 20 μg/ml gentamic in (Figure 5a), indicating that the metabolic activity of D2O in P. aeruginosa was inhibited by gentamicin. On contrary, P. aeruginosa cultivated in cefotaxime can be observed at ~2162 cm−1 SRS imaging at all time points, indicating active metabolic incorporation of D2O in P. aeruginosa when cultivated in cefotaxime. We observed that P. aeruginosa tends to form long rods when cultivated in cefotaxime (Figure 5b). This filamentary formation, which happens when Gram-negative bacteria are treated with β-lactam antibiotics, was also observed for P. aeruginosa treated with ceftazidime10.
To examine whether the D2O metabolic activity of bacteria can be used to rapidly differentiate the antibiotic susceptibility of bacteria, the average C-D signal intensity of bacteria was compared between three groups (Figure 5c - 5h), the control without antibiotics treatment (Figure 4a), treated with gentamicin (Figure 5a), and treated with cefotaxime (Figure 5b). To separate the susceptible and resistant group, we determined a 65% line threshold, which is 65% the average C-D intensity of bacteria in control, in all plots from 10 minutes to 3 h results (Figure 5c - 5h). This threshold can clearly divide the susceptible and resistant groups, the group treated with gentamicin was always below the threshold, and the group treated with cefotaxime was always above the threshold. Therefore, the susceptibility of P. aeruginosa to gentamicin and cefotaxime can be determined in as short as 10 minutes.
Metabolic activity based MIC determination
To test whether SRS metabolic imaging can quantitate the minima l inhibitory concentration (MIC) of antibiotics to bacteria, we cultivated P. aeruginosa in 70% D2O containing LB medium for 1 hour with the addition of gentamicin with serial diluted concentration. SRS imaging at ~2162 cm−1 showed that the D2O metabolic activity of P. aeruginosa was inhibited at 8 μg/ml or higher concentrated gentamicin cultivation (Figure 6a). For control, no C-D signal was observed for P. aeruginosa cultivated in normal LB medium. The average intensity of P. aeruginosa C-D signal was plotted and compared. With the 65% intensity threshold, the metabolic activity based MIC was determined to be 8 μg/ml by the SRS-based D2O metabolic imaging method (Figure 6b). This value is consistent with the MIC determined by the broth dilution method.
Conclusion
We demonstrated rapid determination of the susceptibility of bacteria in 10 minutes by SRS imaging of the D2O metabolic incorporation in bacteria. Femtosecond SRS imaging can monitor D2O metabolism in bacteria at the single cell level with high signal to noise ratio. Antibiotics can inhibit the metabolic activity of D2O when the bacteria are susceptible to this antibiotic, and this inhibition can be observed after culture in D2O containing medium for 10 minutes. The metabolic activity based MIC can be quantitated by our method. Because water is an essential molecule for biosynthesis in bacteria, the SRS-based D2O metabolic imaging method has the potential to be generalized for rapid AST in various species including clinical samples.
AUTHOR INFORMATION
Author Contributions
J.X.C, and W.H. conceived the idea. W.H. designed the experiment. W.H. and L.L. conducted the experiment. W.H. analyzed the data. W.H., and J.X.C. co-wrote the manuscript. All authors have given approval to the final version of the manuscript.
Notes
The authors declare no competing financial interests.
ACKNOWLEDGMENT
This work was supported by Keck Foundation Science & Engineering Grant and R01GM118471 to JXC.