1. Introduction

Raman microspectroscopy is one of the most powerful and nondestructive methods in order to elucidate intra-cellular structure and its dynamics in vivo with three-dimensional sectioning capability [1–8]. Based on the time- and space-resolved molecular specific information obtained by confocal Raman microspectroscopy, we have recently investigated the mitochondrial metabolic activity and an initial death process of a living yeast cell [5–8].The “Raman spectroscopic signature of life” has been found in living fission [5–7] and budding [8] yeast cells, enabling quantitative analysis of cellular bioactivity at the molecular level. In order to trace the detailed dynamical behavior, however, spontaneous Raman microspectroscopy may not be suitable because of its low efficiency; it often takes several minutes to obtain one spectrum. This low efficiency originates from the small scattering cross section of the spontaneous Raman process. An alternative approach to obtain vibrational images with high speed is coherent Raman microspectroscopy. Among them, coherent anti-Stokes Raman scattering (CARS) microscopy has been widely exploited [9–15]. In particular, multiplex CARS microspectroscopy is promising because of its capability to obtain vibrational spectra efficiently [12, 16–19]. The multiplex CARS process requires two laser sources, namely, a narrow band pump laser (ω₁) and a broadband Stokes laser (ω₂). The multiple vibrational coherences are created because of the wide spectral range of the frequency difference, ω₁-ω₂. If we can prepare ultrashort laser pulses, an impulsive Raman excitation and a subsequent narrow-band probe can also generate a multiplex CARS spectrum [20, 21]. One of the most prominent features of multiplex CARS microspectroscopy lies in the fact that it can easily distinguish the concentration change of a particular molecule from the structural change through the spectral analysis. It should be emphasized that a single-wavenumber CARS detection, which is widely adopted in CARS microscopy, cannot discriminate these two phenomena. Although there were several restrictions on the spectral coverage of multiplex CARS microspectroscopy mainly due to the bandwidth of the laser emission [12, 16–19], the spectral coverage has been significantly broadened using the supercontinuum light source generated from a photonic crystal fiber [22–25] or a tapered fiber [26]. Recently, the spectral coverage of the multiplex CARS microspectroscopy has been extended to be more than 2800 cm^-1 [20], which is ranging from 360 nm to 3210 cm^-1. In view of electronic spectroscopy, the supercontinuum light source can also be used as an excitation light source for the two-photon excitation fluorescence (TPEF) [27–29]. Owing to the broadband spectral profile of the supercontinuum, the two-photon allowed electronic state can be excited efficiently in comparison with conventional TPEF microscopy using a Ti:Sapphire oscillator. In the present study, we have combined our multiplex CARS setup with the TPEF detection. Both CARS and TPEF signals have been successfully obtained simultaneously within short data-acquisition time such as 100ms.

2. Experimental

The supercontinuum-based nonlinear optical microspectroscopy system has been described elsewhere [24]. Briefly, an unamplified femtosecond mode-locked Ti:Sapphire oscillator (Coherent, Vitesse-800) was used as a laser source. Typical duration, pulse energy, peak wavelength, and repetition rate were 100 fs, 12 nJ, 800 nm, and 80 MHz, respectively. About 20 % of the output from the oscillator was used for a seed laser to generate a supercontinuum in the PCF (Crystal Fibre, NL-PM-750). The fundamental of the Ti:Sapphire laser and the supercontinuum were used for the pump (ω₁) and Stokes (ω₂) lasers, respectively. In order to obtain CARS spectrum with high frequency resolution, the pump laser pulses were spectrally filtered using a narrow band pass filter. The bandwidth of the pump laser is about 20cm^-1. The pulse energy of the pump and Stokes lasers were 200 and 170 pJ, respectively. Two laser pulses were superimposed collinearly using an 800-nm Notch filter, and then tightly focused onto the sample with a microscope objective (x40). We have modified an inverted microscope (Nikon, TE2000-S). The forward-propagating CARS signal was collected with another microscope objective (x40) in an opposed configuration. After passing through an 800-nm Notch and short wavelength pass filters, the CARS signal was spectrally dispersed by a polychromator (Acton, SpectraPro-300i) and detected by a CCD camera (Roper Scientific, Spec^-10:400BR/XTE). The sample was scanned by a piezo stage (MadCity, Nano-LP^-100). An exposure time for each point was 100 ms. The spatial resolution was estimated to be 0.47 ± 0.01μm for the lateral and 1.51± 0.02μm for the axial directions, respectively [30]. We used fission yeast Schizosaccharomyces pombe (S. pombe) as a sample [5–7]. The nuclei of yeast cells were labeled by green fluorescent protein (GFP). Yeast cells in water were spread on a slide-glass and sandwiched with a cover-glass. Because of a small quantity of the sample, yeast cells were immobilized between a slide-glass and a cover-glass. All measurements were performed at room temperature.

3. Results and Discussion

Figure 1(a) shows typical spectral profiles of the CARS signal of a living yeast cell (red) and surrounding water (blue). As clearly shown, a strong signal is observed inside of the yeast cell at the Raman shift of 2840 cm^-1. This band originates from CH₂ stretching vibrational modes, which shows a slightly dispersive lineshape due to an interference with the nonresonant background. In a previous study, we have extracted only the vibrationally resonant signal using a differentiation method [25]. In the present study, however, the vibrational signal is dominantly observed probably due to the improved spatial resolution and spatial overlap between the pump and Stokes laser pulses. On the basis of our previous spontaneous Raman [5–7] and CARS [25] studies, the signal at the Raman shift of 2840 cm^-1 is found especially in mitochondria, because mitochondrion is an organelle containing a high concentration of phospholipids. Figure 1(b) shows a CARS image of living cells at the Raman shift of 2840 cm^-1. The CARS spectra of a yeast cell (red) and water background (blue) in Fig. 1(a) are obtained at (x, y)=(5.65 μm, -3.05 μm) and (6.71 μm, 1.22 μm), which are indicated as black and white crosses in Fig. 1(b), respectively. Figure 1(b) indicates various yeast cells at different stages of the cell division cycle. Especially, a septum is visualized in the yeast cell around the center of Fig. 1(b). The septum is composed of carbohydrates such as saccharide, which is also rich in CH₂ bonds. The CARS signal at the edge of the yeast cell decreases mainly due to the imperfect focusing caused by the refractive index mismatch between the yeast cell and surrounding water.

 

Fig. 1. (a) Spectral profiles of the CARS signal of a living yeast cell (red) and surrounding water (blue) obtained in 100 ms exposure time; (b) CARS image of living yeast cell at a Raman shift of 2840 cm^-1. The CARS signals at the positions of black and white crosses correspond to the red and blue curves in Fig. 1(a), respectively.

Download Full Size | PDF

Thanks to the three-dimensional sectioning capability, CARS microscopy enables us to obtain not only a lateral but also an axial slice of a living yeast cell. Figures 2 (a) and 2(b) show a lateral and an axial CARS images of a yeast cell, respectively. Figure 2(b) corresponds to the vertical slice of the yeast cell at the position of y=0. The CARS signal is weaker at the top part rather than the bottom part. It is due to the imperfect focusing of the two laser beams because of the spatially heterogeneous refractive index inside of the cell. Figure 2(c) shows axial slices of the same yeast cell at each depth position. Three-dimensional distribution of mitochondoria is clearly observed in Fig. 2(c) with the high three-dimensional spatial resolution.

 

Fig. 2. Lateral(a) and axial(b) CARS images of a yeast cell at a Raman shift of 2840 cm^-1; (c) Axial slices of the yeast cell at different depth positions, which are indicated at the top in a micrometer scale.

Download Full Size | PDF

Next we shall focus on the two-photon excitation fluorescence (TPEF). As described in the previous section, the nuclei of the yeast cells in the present study are tagged by GFP. Figure 3(a) shows the CARS and the TPEF spectra obtained in 100 ms exposure time. The weak but distinguishable peak is observed around 506 nm. Taking into account of the spectral profile of this signal, it is assigned to the TPEF signal due to GFP. Owing to the high quantum yield of GFP, the fluorescence intensity is enough to be detected with such short data-acquisition time. Figures 3(b) and 3(d) show CARS and TPEF images at zero delay time between the pump and Stokes pulses. It should be emphasized that both images of multiplex CARS and TPEF signals are obtained simultaneously. Since there is no overlap between the CARS and TPEF signals in spectral domain, it can be easily differentiated using the spectrometer. Although dual imaging of the CARS and the TPEF signals have been reported using two synchronized Ti:Sapphire oscillators [31], full spectral information is obtained for the first time in the present study. The CARS and TPEF signals indicated as red and green curves in Fig. 3(a) are obtained at (x, y)=(0.31 μm, -2.29 μm) and (4.27 μm, 1.53 μm), respectively. A yeast cell around the center of Figs. 3(b) and (d) shows two spots due to GFP in Fig. 3(d). It means that the yeast cell is in the M phase. Compared with Fig 1(b), septum is not clearly visualized by CH₂ stretching vibrational mode. From our previous study [5], Raman spectrum of a matured septum is different from that of a primary septum. From this result, the CARS signal intensity due to CH₂ stretching vibrational mode can be used for monitoring the evolution of the septum. Figures 3(c) and 3(e) show CARS and TPEF images at -4-ps delay time between the pump and Stokes pulses. Since the CARS process occurs through a virtual electronic transition, no signal is observed when there is no temporal overlap between the pump and Stokes pulses at such a large delay. On the other hand, TPEF signal is still observed even without temporal overlap. It should be noted, however, that the signal intensity is drastically decreased by a factor of 3.4. The signal is, therefore, ascribed to degenerate 2ω₁- and/or 2ω₂-photon processes. In the present experimental condition, the TPEF signal due to the 2ω₂-photon process was negligible. In order to evaluate the signal intensity only due to the non-degenerated (ω₁+ω₂)-photon process, we have subtracted the TPEF signal intensity in Fig. 3(e) from that in Fig. 3(d). The result is depicted in Fig. 3(f). It is clear that the TPEF signal due to the (ω₁+ω₂)-photon process is larger than that due to the 2ω₁-photon processes. It can be explained by the following two reasons. First, the TPEF signal intensity in Fig. 3(f) is proportional to the intensity both of the ω₁ and the ω₂ pulses. On the other hand, the TPEF signal intensity in Fig. 3(e) depends quadratically on the intensity of the ω₁ pulses. Therefore, the threshold of the laser power in the former is lower than that in the latter. Second, the wide spectral bandwidth of the Stokes pulse works as an efficient twophoton excitation light source. On the other hand, the degenerate two-photon process provides a bandwidth much narrower than that of the non-degenerated (ω₁+ω₂)-photon process. Furthermore, a spectrally filtered or a temporally chirped supercontiuum with a narrow-band pump pulse can be used for an ideal light source to obtain fluorescence excitation spectrum of a two-photon allowed electronic state.

Figure 4 shows multi-nonlinear optical imaging of the CARS (red) and TPEF (green) signals, which is obtained in Fig. 3. As described, the yeast cell at the center of Fig. 4 is in the M phase. It is also found that the mitochondria exist around the nucleus.

 

Fig. 3. Spectral profile of the CARS and TPEF signals of a living yeast cell; CARS lateral images of living yeast cells at the Raman shift of 2840 cm^-1 at the delay time of zero (b) and -4 ps (c), respectively; TPEF lateral images of the same system at 506 nm at the delay time of zero (d) and -4 ps (e), respectively; TPEF image only due to the non-degenerated (ω₁+ω₂)- photon process.

Download Full Size | PDF

 

Fig. 4. In-vivo multi-nonlinear optical imaging of yeast cells. The CARS and TPEF images are indicated as red and green, respectively.

Download Full Size | PDF

4. Conclusion

In conclusion, multi-nonlinear optical imaging of a living yeast cell has been performed by supercontinuum-based microspectroscopy. Owing to the broadband feature of the supercontinuum, the spectral profile of the multiplex CARS signal can be elucidated in detail. In addition to the ultrabroadband multiplex CARS detection, the efficient two-photon excitation can also be performed in resonance with the two-photon allowed electronic state. Since the Raman signal due to a nucleus is weak to be detected, dual imaging of the CARS and the TPEF signals provides useful information on the dynamical behavior of the living system with high speed. The supercontinuum enables us to perform unique multi-nonlinear optical imaging through various nonlinear optical processes.