We developed a multi-focus excitation coherent anti-Stokes Raman scattering (CARS) microscope using a microlens array scanner for real-time molecular imaging. Parallel exposure of a specimen with light from two highly controlled picosecond mode-locked lasers (jitter of 30 fs through an electronic low-pass filter with 150 Hz bandwidth, point-by-point wavelength scan within 300 ms) and parallel detection with an image sensor enabled real-time imaging. We demonstrated real-time CARS imaging of polystyrene beads (frame rate of 30 fps), a giant multi-lamellar vesicle of dipalmitoylphosphatidylcholine (frame rate of 10 fps), and living HeLa cells (frame rate of 10 fps).
©2009 Optical Society of America
Coherent anti-Stokes Raman scattering (CARS) microscopy is a powerful tool for chemical-sensitive three-dimensional imaging of biological specimens without staining [1, 2]. CARS measurement generally uses two highly synchronized picosecond mode-locked lasers of different wavelengths (ω 1 light and ω 2 light). When the frequency difference of the ω 1 and ω 2 beams coincides with a molecular vibration (Ω=ω 1−ω 2), a CARS signal (ωas=2ω 1−ω 2) is resonantly enhanced. CARS microscopy features high spatial resolution, label-free imaging, strong signals compared with conventional Raman scattering microscopy, the absence of Stokes shifted fluorescence, and the ability to detect molecular species. CARS microscopy has been applied to biological measurements. Applications include studies of lipid metabolism [3, 4], organelle transport , viral disease , dividing cell dynamics , membrane chemistry [8, 9], brain medicine , and so on.
In biological measurement, high-speed imaging is required for the observation of biological phenomena that change spatio-temporally. In the case of CARS microscopy, video-rate CARS imaging of tissue in vivo was realized by the use of a high-speed laser scanner with a polygon mirror and a galvanometer mirror . The mirrors quickly scanned a single focal point on a specimen, enabling the dwell time at each point to be shorten. Consequently, the intensity of the incident excitation laser light needs to be high to obtain a sufficiently high CARS signal for imaging. However, a high intensity might cause photo-damage of the specimen. High-speed imaging of a biological specimen therefore requires a high-speed laser scanning system, a high signal-to-noise ratio imaging system, and low-intensity excitation laser light.
In the present study, we realized real-time CARS imaging with low-intensity excitation of each spot by the use of a multi-focus excitation technique, and we demonstrated real-time imaging of polystyrene beads (30 fps), giant multi-lamellar vesicles (MLVs) of dipalmitoylphos-phatidylcholine (DPPC) lipids (10 fps), and living HeLa cells (10 fps).
2. Multi-focus CARS microscope
The intensity of a CARS signal I as from a spot on a specimen is written as
where χ(3), I 1, I 2, and τ ex are the third-order nonlinear susceptibility, the intensity of ω 1 light, the intensity of ω 2 light, and the dwell time at the spot. When performing high-speed CARS imaging with single beam scanning, a sufficient CARS signal is generally obtained by increasing the excitation laser intensity to compensate for the short dwell time. However, the laser intensity is limited to the range of a few milliwatts to a few tens of milliwatts. Because a CARS microscope generally uses near-infrared lasers (oscillating wavelength of 700–1000 nm) with a pulse duration of a few picoseconds or shorter, nonlinear photochemical damage is the major problem rather than photothermal damage because of the weak linear absorption of near-infrared light in most cells, except hemoglobin, chlorophyll, and other strong near-infrared absorbers [12, 13]. The photochemical damage is caused through a second- or higher-order photon process in which the photodamage efficiency is proportional to the second- or higher-order power of the intensity of the laser spot, and thus the peak power of the spot is limited for non-invasive imaging. In the case of multi-focus excitation, the dwell time is increased in proportion to the number of focal spots without reducing a frame rate. When the beams are split into 100 beamlets, the signal-to-noise ratio of the obtained CARS image is improved one hundredfold compared with single-beam scanning with the same spot intensity.
Unfortunately, the total laser intensity of all spots in multi-focus excitation is higher than that in single-beam excitation, and long exposure with high intensity might increase the temperature of the specimen. The temperature of the specimen, however, will not increase by more than a few degrees Kelvin during a <10 s exposure of near-infrared light with a power of 100–200 mW , and this temperature increase may not cause critical photothermal damage. Of course, in long-exposure imaging, such as high signal-to-noise ratio imaging of a weak Raman band or real-time imaging over a long duration, the temperature of the specimen might increase to the critical temperature for photothermal damage. However, in the case of a specimen in an aqueous culture medium, a temperature increase can be prevented by reflowing the culture medium because water is the major absorber of near-infrared light in a biological specimen.
We used a microlens array scanner for forming of multiple focal spots [15, 16]. The microlens array scanner has many microlenses on a rotating disk. An excitation laser beam incident on the microlens array scanner is split into multiple beamlets to expose multiple points on a specimen simultaneously. An image is obtained by rotating the microlens array disk. The multi-focus excitation technique with a microlens array scanner has been applied to fluorescence microscopy [17, 18] and second harmonic generation (SHG) microscopy , to realize real-time imaging. In this study, we have realized real-time CARS imaging by the use of a higher efficiency microlens array scanner and a higher sensitive image sensor than the previous study [15, 16].
3. Experimental setup
The optical setup of the multi-focus CARS microscopy system is shown in Fig. 1. The system consisted of two picosecond mode-locked Ti:sapphire lasers operating at different wavelengths (pulse duration of 5 ps, repetition rate of 80 MHz, Tsunami, Spectra-Physics), a high-precision pulse synchronization system , an automatic pulse duration minimizing system , a microlens array scanner (lens diameter of 0.58 mm, focal length of 11.6 mm, MLA1-DD, Nanophoton), a modified inverted microscope (TE-200, Nikon), and an electron-multiplying charge-coupled device camera (EM-CCD, DV-897, Andor).
The two lasers were synchronized by phase-locked loop (PLL) control using two types of error signals: the fundamental signal of the repetition frequency of the lasers and the signal of a balanced cross-correlator using two-photon detectors. The lasers were synchronized to within 1 ps of timing jitter by using the fundamental signal of the repetition frequency of 80 MHz beforehand. Then the error signal of the PLL controller was switched to the signal of the balanced-cross correlator, and the timing jitter between the two lasers was reduced to within 30 fs by using an electronic low-pass filter with 150 Hz bandwidth. A digital signal processor (DSP, TMS320C6713, Texas Instruments) with an analog-digital/digital-analog converter interface (DSK6713IF-A, Hiratsuka Engineering) was used as a controller for the PLL control. These synchronized laser beams were spatio-temporally overlapped and were made incident on a microlens array scanner to split them into multiple beamlets. The microlens array scanner was high efficiency for the CARS microscope, i.e., the larger diameter of the microlens and antireflection coating for near-infrared light to increase an excitation intensity on each spot compared with the previous study [15, 16]. The beamlets were collimated with relay lenses and were focused to multiple spots on the specimen with an objective lens (S Fluor, Nikon, x40, N.A.=0.85). The system produced seven focal spots on the specimen from laser beams 2 mm in diameter. CARS signals from the multiple focal spots on the specimen were collected by another objective lens (UPlanApo, Olympus, x40, N.A.=0.85), and the fundamental signals were cut with optical filters. The CARS signals were observed in parallel with the EM-CCD camera, and a CARS image was obtained by rotating the microlens array disk. The observed molecular vibration was tuned by adjusting the wavelength of the ω 2 light. The pulse duration of the ω 2 light was automatically minimized by the use of a two-photon absorption detector. The signal from the two-photon absorption detector was inversely proportional to the pulse duration; therefore, the group delay dispersion of the laser cavity was compensated with a Gires-Tournois interferometer to minimize the pulse duration by maximizing the signal of the two-photon absorption detector. The preparation time for CARS measurement, including wavelength tuning in short wavelength range, pulse duration optimization, and jitter control in the point-by-point wavelength scan, was within 300 ms.
4.1. Real-time spectral-imaging of polystyrene beads
The real-time CARS imaging capability was demonstrated with polystyrene beads. The polystyrene beads (3 µm diameter, Polybeads Microspheres 3.00µm, Polysciences) mixed with deionized water were dispersed on a slide glass and sealed by a coverslip with a 50-µm-thick silicone spacer.
Figure 2 shows the CARS spectra of the polystyrene beads and the background (water) at 1000 cm-1. The spectra were obtained from a pixel of the CARS images of the polystyrene beads in the water at each Raman shift in which the beads were not moved. The CARS images were obtained with an image acquisition time of 33 ms/image. The total intensities were 75.9 mW at 780 nm and 29.7 mW at 846 nm at the focal plane with 7 focal spots. The CARS spectrum of the polystyrene beads had a strong resonance at 1000 cm-1; however, the water did not have any peaks. As a result, we could selectively observe the polystyrene beads in a short exposure time with the developed microscope. The 1000 cm-1 resonance of the polystyrene beads is assigned to the phenyl ring breathing mode .
A video of real-time CARS images of moving polystyrene beads in water observed at 1000 cm-1 is shown in Fig 3 (Media 1). The images were obtained at a frame rate of 30 fps. The flowing polystyrene beads exhibiting Brownian motion in water were clearly observed.
We also demonstrated real-time CARS imaging of polystyrene beads with wavelength scanning. Figure 5 shows the temporal behaviors of the signal from the two-photon absorption detector that monitored the pulse duration of the ω 2 light, the observed Raman shift, and the CARS images during wavelength scanning. A video of the CARS images is shown in Fig 4 (Media 2). When the observed molecular vibration was changed from 998 cm-1 after about 10 s by tuning the oscillating wavelength of the ω 2 light by adjusting a birefringence filter driven by a computer controlled stepping motor, the two-photon signal decreased because the group-delay dispersion in the cavity was not optimized, and thus, the CARS image of the beads disappeared. However, the signal from the two-photon absorption detector spiked immediately, and the observed molecular vibration became 1004 cm-1 within 0.3 s. After about 45 s, at which the observed molecular vibration was changed from 1005 cm-1 to 1000 cm-1, the CARS image of the beads appeared within 0.3 s after adjusting the wavelength of the ω 2 light. The signal from the two-photon absorption detector showed the same behavior at other molecular vibrations. As a result, the pulse duration and the timing jitter of the two pulses were optimized within 0.3 s in the point-by-point wavelength scanning in short wavelength range. The system is thus capable of CARS spectral imaging of polystyrene beads with a frame rate of 1 fps, including information in 2–3 Raman shift signals.
4.2. Real-time imaging of a giant multi-lamellar vesicle
We also demonstrated real-time imaging of a DPPC MLV. The preparation of the DPPC MLV was as follows: First, DPPC lipids (P0763, Sigma) dissolved in a chloroform/methanol solution (98:2 vol/vol) were evaporated under a flow of nitrogen for 24 hours and in vacuo for at least 90 minutes in an Erlenmeyer flask. A HEPES buffer (10 mM, HEPES-NaOH, pH 7.0) was then added to the dried lipid films at room temperature, and the lipids formed MLVs. The MLVs were spread on a slide glass and sealed by a coverslip with a 50-µm-thick silicone spacer.
DPPC lipids exhibit CH2 deformation vibrations of the acyl chains at a Raman shift of 1442 cm-1 . CARS and Raman spectra of DPPC lipid powder are shown in Fig. 6. The CARS spectrum of the DPPC lipids was much broader than the Raman spectrum and exhibited a dispersive line shape, and so the peak of the CARS spectrum (1442 cm-1) was slightly shifted from that of the Raman spectrum (1445 cm-1). In this study, we visualized DPPC MLVs through the CH2 deformation vibrations.
Figure 7 shows a CARS image of a DPPC MLV observed at 1442 cm-1 and 1486 cm-1. The image acquisition time was 3 s/image. When the observed Raman shift was set at the CH2 deformation vibrations (1442 cm-1), a cross-sectional image of the DPPC MLV near the equator was clearly observed. In case of a nonresonant image (1486 cm-1), the CARS singnal was weakened. Because we used a two-dimensional image sensor in our developed system for simultaneous observation of multiple focal points, the spatial resolution in the xy plane was limited by the observation optics and was lower than in a conventional point-by-point one-dimensional CARS system . On the other hand, the spatial resolution along the z-axis depended on the excitation optics and the third-order nonlinearity of the CARS generation. The developed multi-focus CARS microscope thus had a high spatial resolution along the z-axis (~2 µm), and z-section CARS images of the DPPC MLVs were clearly visualized.
A video and snapshots of the real-time imaging of a DPPC MLV at various z-positions is shown in Fig. 8 (Media 3) and Fig. 9. The depth position of the DPPC MLV was set with a piezoelectric transducer stage (PZT stage, 17 ANC 001/MD, NanoBlock xyz Flexure Stage, Melles Griot). The observed molecular vibration was set at 1442 cm-1. Each cross-sectional CARS image of the DPPC MLV was observed within 100 ms. The images were obtained from the equator to the pole of the DPPC MLV, and the cross-sectional images of the DPPC MLV at each z-position were clearly observed.
Three-dimensional reconstruction of the CARS images of a DPPC MLV was also demonstrated (Fig. 10). The z-section CARS images were obtained by moving the specimen over 10 µm (200 nm/step) with the PZT stage. The observed molecular vibration was set at 1442 cm-1. The image acquisition time of each cross-sectional image was 100 ms/image. The three-dimensional CARS image of the DPPC MLV consisted of 50 slices and was obtained within 7 s, and a hollow-sphere-shaped DPPC MLV with a diameter of about 20 µm was visualized.
4.3. Real-time imaging of living cells
We also demonstrated real-time imaging of living HeLa cells. The HeLa cells were cultured on a slide glass bottom culture dish immersed in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum. The culture medium was replaced with a modified Tyrode’s solution before CARS imaging. A water immersion objective lens (NIR Apo, Nikon, x60, N.A.=1.0, water immersion) was used for collection of CARS signals.
Figure 11 shows a CARS image of the HeLa cells observed at 2840 cm-1. The image acquisition time was 1 s/image. Since the molecular vibration at 2840 cm-1 is assigned to CH2 stretching mode of lipids, lipid rich regions in the HeLa cells were clearly observed. A video of the CARS images of the HeLa cells with a frame rate of 10 fps is shown in Fig 12 (Media 4). As a result, we could also visualize the living cells without any treatment in a short exposure time with the developed microscope.
In conclusion, we have developed a multi-focus CARS microscopy system for real-time molecular imaging. With multi-focus excitation, the signal-to-noise ratio of a CARS image was improved in proportion to the number of focal spots without increasing the intensity of each focal spot. Real-time CARS spectral imaging of polystyrene beads, a DPPC MLV, and living HeLa cells was demonstrated. The system was capable of visualizing the polystyrene beads within 33 ms, and the DPPC MLV and the HeLa cells within 100 ms, respectively. The preparation time for point-by-point wavelength scanning was less than 300 ms.
For more detailed visualization of the behavior of biological specimens, analysis of CARS images at various Raman shifts in the fingerprint region is necessary. The CARS spectrum in the fingerprint region reflects structural changes of biological molecules, and analysis of CARS images at several Raman shifts gives us information about not only the distribution of the molecules, but also the activity of the molecules.
As the number of foci was seven in the present study, the multi-focus effect might be not so high. A large number of foci will provide a much higher multi-focus effect, however, a photo-damage risk to live cell or tissue caused by strong laser illumination might be increased. Therefore, more investigations will be required to show the advantage of multi-focus system.
This work was partially supported by a Grant-in-Aid for Scientific Research (A) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) and the program of Development of Systems and Technology for Advanced Measurement and Analysis (SENTAN) from the Japan Science and Technology Agency (JST). One of the authors (T.M.) acknowledges the support by Grant-in-Aid for JSPS Fellows from Japan Society for the Promotion of Science (JSPS).
References and links
1. A. Zumbusch, G. R. Holtom, and X. S. Xie, “Three-Dimensional vibrational imaging by coherent anti-Stokes Raman scattering,” Phys. Rev. Lett. 82, 4142–4145 (1999). [CrossRef]
2. M. Hashimoto, T. Araki, and S. Kawata, “Molecular vibration imaging in the fingerprint region by use of coherent anti-Stokes Raman scattering microscopy with a collinear configuration,” Opt. Lett. 25, 1768–1770 (2000). [CrossRef]
3. X. Nan, J. Cheng, and X. S. Xie, “Vibrational imaging of lipid droplets in live fibroblast cells with coherent anti-Stokes Raman scattering microscopy,” J. Lipid Res. 44, 2202–2208 (2003). [CrossRef] [PubMed]
5. X. Nan, E. O. Potma, and X. S. Xie, “Nonperturbative chemical imaging of organelle transport in living cells with coherent anti-Stokes Raman scattering microscopy,” Biophys. J. 91, 728–735 (2006). [CrossRef] [PubMed]
6. X. Nan, E. A. M. Tonary, A. Stolow, X. S. Xie, and J. P. Pezzacki, “Intracellular imaging of HCV RNA and cellular lipids by using simultaneous two-photon fluorescence and coherent anti-Stokes Raman scattering microscopies,” ChemBioChem 7, 1895–1897 (2006). [CrossRef] [PubMed]
8. E. O. Potma and X. S. Xie, “Detection of single lipid bilayers with coherent anti-Stokes Raman scattering (CARS) microscopy,” J. Raman Spectrosc. 34, 642–650 (2003). [CrossRef]
9. M. Müller and J. M. Schins, “Imaging the thermodynamic state of lipid membranes with multiplex CARS microscopy,” J. Phys. Chem. B 106, 3715–3723 (2002). [CrossRef]
10. C. L. Evans, X. Xu, S. Kesari, X. S. Xie, S. T. C. Wong, and G. S. Young, “Chemically-selective imaging of brain structures with CARS microscopy,” Opt. Express 15, 12076–12087 (2007). [CrossRef] [PubMed]
11. C. L. Evans, E. O. Potma, M. Puoris’haag, D. Côté, C. P. Lin, and X. S. Xie, “Chemical imaging of tissue in vivo with video-rate coherent anti-Stokes Raman scattering microscopy,” Proc. Natl. Aca. Sci. 102, 16807–16812 (2005). [CrossRef]
12. A. Schönle and S. W. Hell, “Heating by absorption in the focus of an objective lens,” Opt. Lett. 23, 325–327 (1998). [CrossRef]
13. K. König, T. W. Becker, P. Fisher, I. Riemann, and K. -J. Halbhuber, “Pulse-length dependence of cellular response to intense near-infrared laser pulses in multiphoton microscopes,” Opt. Lett. 24, 113–115 (1999). [CrossRef]
14. Y. Liu, D. K. Cheng, G. J. Sonek, M. W. Berns, C. F. Chapman, and B. J. Tromberg, “Evidence of localized cell heating induced by infrared optical tweezers,” Biophys. J. 68, 2137–2144 (1995). [CrossRef] [PubMed]
15. M. Hashimoto, T. Araki, and S. Kawata, “Multi-focus coherent anti-Stokes Raman scattering microscopy,” Microsc. Microanal. 9(Suppl 2), 1090–1091 (2003).
16. T. Ichimura, N. Hayazawa, M. Hashimoto, Y. Inouye, and S. Kawata, “Local enhancement of coherent anti-Stokes Raman scattering by isolated gold nanoparticles,” J. Raman Spectro. 34, 651–654 (2003). [CrossRef]
17. M. Straub and S. W. Hell, “Multifocal multiphoton microscopy,” Opt. Lett. 23, 655–657 (1998). [CrossRef]
18. T. Kaneko, K. Fujita, H. Tanaka, M. Oyamada, O. Nakamura, S. Kawata, and T. Takamatsu, “Real-time two-photon microscopy and its application for in situ imaging,” Acta Histochem. Cytochem. 34, 399–403 (2001). [CrossRef]
19. M. Kobayashi, K. Fujita, T. Kaneko, T. Takamatsu, O. Nakamura, and S. Kawata, “Second-harmonic-generation microscope with a microlens array scanner,” Opt. Lett. 27, 1324–1326 (2002). [CrossRef]
20. T. Minamikawa, N. Tanimoto, M. Hashimoto, T. Araki, M. Kobayashi, K. Fujita, and S. Kawata, “Jitter reduction of two synchronized picosecond mode-locked lasers using balanced cross-correlator with two-photon detectors,” Appl. Phys. Lett. 89, 191101 (2006). [CrossRef]
21. M. Hashimoto, T. Asada, T. Araki, Y. Inouye, and S. Kawata, “Automatic pulse duration control of picosecond laser using two-photon absorption detector,” Jpn. J. Appl. Phys. 44, 3958–3961 (2005). [CrossRef]
22. A. Fontes, K. Ajito, A. A. R. Neves, W. L. Moreira, A. A. de Thomaz, L. C. Barbosa, A. M. de Paula, and C. L. Cesar, “Raman, hyper-Raman, hyper-Rayleigh, two-photon luminescence and morphology-dependent resonance modes in a single optical tweezers system,” Phys. Rev. E 72, 012903 (2005). [CrossRef]
23. S. Straub and S. W. Hell, “Multifocal multiphoton microscopy: a fast and efficient tool for 3-D fluorescence imaging,” Bioimaging 88, 177–185 (1998). [CrossRef]
24. T. O’Leary and I. Levin, “Raman spectroscopic study of the melting behavior of anhydrous dipalmitoylphosphatidylcholine bilayers,” J. Phys. Chem. 88, 1790–1796 (1984). [CrossRef]