1. Introduction

Mitochondrion, the cellular power factory with own heredity material and autonomic system of protein biosynthesis, is a important organelle in animal cells in which oxidative phosphorylation takes place. It is well known that mitochondrion has an outer membrane, which allows the passage of most small molecules and ions. Mitochondrion also has a highly folded inner membrane that contains high concentrations of various membrane protein complexes of the electron transport chain and ATP synthase [1, 2]. The constitution and conformation of these complexes in the mitochondria determine the function and bioactivity of mitochondria. It is important to know more about the molecular composition of mitochondria and their biochemical properties. Previous studies analyzed mitochondria composition and bioactivity of bulk samples using spectrophotometric methods, fluorescent probes and additional biochemical analysis techniques [1, 3]. These biochemical methods and protocols available to monitor the bioactivity of mitochondria are applied to a certain mitochondria population and not to single mitochondria.

Mitochondria are small organelles, with a size ranging from 0.5 to 5 μm in diameter [1]. It has been know that calcium ions overload cause a swelling of large magnitude of isolated mitochondria. The Ca²⁺, an important intracellular second-messenger, regulates energy metabolism under physiological conditions. But Ca²⁺ overload under pathologica conditions would do damage to cells and mitochondria. The in vivo introduction of Ca²⁺ leads to the accumulation of calcium diacetate in the mitochondrial matrix and subsequently induces the release of cytochrome C and other proteins from mitochondria intermembrane space into the cytosol, where it can activate caspases and lead to apoptosis [4]. For isolated mitochondria, high Ca²⁺ induced mitochondria swelling results in mitochondria damage that involves increased permeability of the inner membrane to small solutes, osmotic swelling of the mitochondrial matrix, and physical disruption of the outer membrane and translocation of cytochrome C from mitochondria to solution buffer [5]. Therefore, it is essential to study the response of individual isolated mitochondria to the exposure of Ca²⁺ solution in terms of its functions and composition. However, this process has not been observed by molecular vibration spectroscopy in real time for single mitochondria.

Raman spectroscopy is a vibration spectroscopic technique that can provide information about the molecular composition and structure of the samples. To date Raman spectroscopy has been widely used as a sensitive probe to analyze the subtle molecular changes of biological matters, including microorganism, tissues, and plant or mammalian cells [6–14]. This technique has been developed to identify the differences between cancer tissue and surrounding normal tissue [14]. And some investigators have explored the potential of Resonance Raman to study mitochondrial cytochrome C oxidase structures and their interaction with mitochondria [15, 16]. Recently, Toshiba et al also applied resonance Raman spectroscopy for probing the oxygen activation reaction in intact whole mitochondria and showed the high quality absorption and resonance Raman spectra of porcine heart mitochondria [17]. However, these results could not provide complete Raman spectrum of single mitochondria because that UV laser excitation can degrade biological samples due to strong absorption. For studies of living biological material, off-resonance Raman spectroscopy excited by near infrared laser sources has a great advantage in reducing potential photo damage and sample degradation, since biological samples usually have small absorption at near infrared wavelengths [18, 19]. Huang et al used time and space-resolved Raman spectroscopy to detect the molecular and structural information of single living fission yeast cells [20, 21]. They also recorded the Raman spectroscopy of GFP marked mitochondria in yeast cell. However, the mitochondria may move away from the excitation volume of the micro-beam within the acquisition time due to Brownian motion or mitochondria motility. Furthermore, these Raman spectra are not specific features of single intact mitochondria because of the background fluorescent interference and they can not directly measure the dynamical changes in biochemical properties of the single mitochondria during induced damage process.

Optical tweezers have become a powerful tool with which to capture and manipulate biological particles, including cells, bacteria, viruses, and chromosome [22–25]. In previous study, we developed a compact Laser Tweezers Raman Spectroscopy (LTRS) system that combined optical tweezers and near-infrared Raman spectroscopy for the manipulation and characterization of a biological material [26, 27]. In this system, optical trapping allows a mobile particle to be held at the focus of a near-infrared laser beam for long time observation and permits optimum excitation and collection for Raman scattering in a confocal configuration. We had demonstrated the application of this system for identification of single yeast cells [28], bacterial cells [29], spores [30] and human red blood cells [26, 31]. The average mitochondria, whether in a mammalian or in a lower eukaryotic cell (e.g., yeast cell), has approximately the same dimension as the bacterium Escherichia coli [1]. It is most commonly observed as an oval particle, 1-2 μm long and 0.5-1 μm wide. In our LTRS system, high power microscope objectives were used, focusing the laser beam into a spot approximately 0.5 μm in diameter. Mitochondria, therefore, can be effectively captured and manipulated by laser tweezers. In this study, we applied LRST system to investigate NIR Raman spectroscopy of single intact or swollen mitochondria trapped by laser tweezers. We demonstrated the Raman spectroscopic “molecular fingerprint” of single mitochondria is unique, highly reproducible and can be used to monitor composition and bioactivity changes of the intact and swelling mitochondria.

2. Materials and methods

2.1 Sample preparation

All animal experiments were approved by the Ethical Committee of the Institute of Zoology. Liver, kidney and heart Tissues were obtained from Sprague -Dawley rats that were 2–3 months old. Animals not showing macroscopic evidence of pathologies were sacrificed by decapitation after overnight fasting. Liver mitochondrial were isolated according to the method of Johnson and Lardy [32]. Briefly, the rat liver was rinsed free of blood and minced in ice-cold isolation medium (0.25 M sucrose, 0.3 M mannitol (Sigma), 10 mM HEPES and 0.1 mM EDTA, PH 7.4, in deionized water). The tissue was then homogenized with a loose-fitting glass homogenizer. The homogenate was then centrifuged at 1200 g for 10 min at 4.0 °C. And the resulting supernatant was then centrifuged at 10000 g for 10 min. The mitochondrial pellet was washed and centrifuged again in the isolation medium. The final pellet was resuspended in 10 ml isolation medium. The entire procedure was completed within 1.5 h. Kidney and heart muscle mitochondria were isolated using similar procedures according to references [32] and [33], respectively. For monitor the Raman spectral changes of swollen mitochondria, the freshly isolated mitochondria pellet were resuspended in KCL buffer (150 mM KCL, 5 mM Tris-HCL, 20 mM MOPS, 5 mM KH₂PO₄, PH 7.4) and then CaCl₂ stock solution was added to reach final concentration 100 μM. The mitochondrial medium was put in a microscopy sample plates which was made of a 4.0 mm thick glass slide and has a hole sealed with a cover slip. The plates with sample medium were placed directly under the microscope objective for measurements.

2.2 Raman instrumentation

Details of the Laser Tweezers Raman Spectroscopy have been published elsewhere [27]. Briefly, the LTRS instrument possesses a wavelength-stabilized, beam shape-circulated semiconductor diode laser beam at 785 nm that then is introduced into an inverted differential interference contrast (DIC) microscope (Nikon TE 2000) equipped with a high numerical aperture objective (100×, NA= 1.30) to form an optical trap. A mitochondrion in the isolation buffer can be trapped above the bottom cover-slip with the gradient force yielded by the focused beam. The same laser beam is used to excite Raman scattering of the trapped the mitochondrion. The scattering light from the mitochondrion is collected by the objective and coupled into a spectrograph through a 200 μm pinhole, which enables confocal detection and filtration of off-focusing Rayleigh scattering light. A holographic notch filter is used as a dichroic beam splitter that reflects the 785 nm excitation beam and transmits the Raman shifted light. A green-filtered illumination lamp and a video camera system are used to verify trapping and observe the image of the mitochondrion. The spectrum is obtained by a liquid-nitrogen-cooled charge-coupled detector (SPEC-10:100BR, Princeton, NJ). The spectral resolution of our Raman system is about 6 cm⁻¹ and the Raman spectra can be recorded in the “finger print” range from 600 to 1800 cm⁻¹.

2.3 Acquisition of Raman spectra of single optical trapped mitochondria

Before measurement, a polystyrene bead of 2 μm diameter suspended in water was used for the alignment and calibration of the LTRS system. The Raman spectrum of the bead was acquired with a 1 s exposure time and 15 mW excitation power. Then the bead was released from the trap and background spectrum was taken with the same acquisition time and power. Freshly isolated mitochondria were suspended into 10 ml of isolation buffer and diluted to single mitochondria. Then we loaded about 100 μl buffer into the hole of a temperature-controlled microscope sample holder which can keep at 4 °C. After loading the sample, a single mitochondrion in buffer was randomly trapped by the laser beam during the measurement. Raman spectra of the trapped mitochondrion were acquired with a 15 mW laser power and 90 s exposure time. After recording the trapped mitochondrion, the sample was released from the beam focus and the background spectrum of isolation buffer without the mitochondrion was also obtained with the same acquisition time and power. For monitoring the whole process of Ca²⁺ induced swelling of mitochondria, Raman spectra of single mitochondria isolated from rat liver was measured every 5 minutes after the sample was exposed to the high Ca²⁺ buffer in the presence of Pi. The whole measure time was up to 60 minutes and the above procedure was repeated for the observation of each individual mitochondrion. We recorded more than 30 single mitochondria and obtained the average of Raman spectra of them.

3. Results and discussion

Mitochondria are bacteria-sized organelles, which range in size from 0.5 to 5 μm in diameter. Therefore, it is difficult to probe their Raman spectra without marking and fixing because of Brownian motion and mitochondria motility. But fixing these biological samples may change their biological activity. Using our novel Raman spectroscopic techniques, the single mitochondria suspended in isolation buffer can be captured by optical tweezers. Both Raman spectra of mitochondria in the focus of optical spot and medium without mitochondria were recorded with the same acquisition time and power. Fig. 1 shows background subtracted Raman spectra of the intact mitochondria isolated from rat liver, heart muscle, and kidney which were suspended in isolation buffer. The insets of Fig. 1 display microphotographs of mitochondria trapped by optical tweezers. All of the mitochondria are composed of lipid, DNA and the associated proteins, and their Raman spectra contained the similar characteristic Raman peaks (Fig.1). From the average Raman spectra of mitochondria, we found that there were prominent spectral peaks at 1003, 1266, 1303, 1446, 1602, 1655 cm⁻¹. These spectra features arise from the molecular vibrations of mitochondria components, such as the lipids, nucleic acids and proteins. Table.1. shows the tentative assignment for the observed Raman signal of single mitochondria [21, 26–28]. The band at 1003 cm⁻¹ is assigned to ring breathing mode of phenylalanine. The signal at 1266 cm⁻¹ band can be undoubtedly assigned to amide □ of proteins. And the band at 1303 cm⁻¹ be assigned to the O-P-O⁻ of Amide and DNA indicates that the mitochondria have own heredity material nucleic acids. The broad band at 1446 cm⁻¹ band is very strong which can be assigned to the CH2 bending modes and CH3 deformation of lipids and proteins. Bands at 1602 cm⁻¹ and 1655 cm⁻¹ were assigned to tyrosine/phe/tryptophan and the c=c stretching vibration of the cis ‒CH=CH- linkage of the unsaturated lipid chains respectively. These bands were also observed by huang et al in the yeast cell [21]. They pointed that lipid Raman bands of mitochondria were in sharp contrast to the protein-dominant Raman spectra of cytoplasm and nuclei. And compared with the spectra of purified protein, these bands in the spectra of mitochondria were not detected in the cytochrome C. Therefore, these bands can be used as the Raman spectra “molecular finger” of mitochondria.

 

Fig. 1. NIR Raman spectra (averaged from more than 30 mitochondria) and images of single intact mitochondria. The acquisition time was 90.0 s with 15 mW excitation power at 785 nm. (a) the spectrum of heart mitochondria; (b) the spectrum of mitochondria isolated from rat kidney; (c) the spectrum of liver mitochondria, curve d, e, f, each is a difference between liver and heart muscle, liver and kidney, heart muscle and kidney, respectively. The insets show the images of single trapped mitochondria isolated from heart muscle (a), kidney (b) and liver (c). The scale bar is 2 μm.

Download Full Size | PDF

Table 1. Raman bands for mitochondrial (averaged over 30 mitochondria) and their tentative assignments.

View Table

Further examination of these spectra indicates that there were several subtle differences in Raman bands of mitochondria isolated from liver, heart muscle, and kidney, although most of spectra peaks positions were identical (such as those located at 1003, 1266, 1303, 1446, 1602 cm⁻¹ and 1655 cm⁻¹). The different spectrum (see fig.1 A, C) between the liver mitochondria and heart mitochondria might demonstrate that the lipid composition and the molecular weight of the proteins are dissimilar. The height of any given peak is based on the total intensity of the signal, which can vary from moment to moment. The relative intensity of 1446 cm⁻¹ band (which can be assigned to CH deformation of lipid and protein) of heart mitochondria is higher than that of liver mitochondria which indicate their different lipid content. Earlier studies proved that liver mitochondria contain relatively few cristae and less inner membrane surface than heart muscle mitochondria. The heart muscle mitochondria lack many of the enzymes found in liver mitochondria and their cristae are packed more densely. On the other word, there were different lipid compositions and percents of heart, kidney and liver mitochondria. And the detailed lipid analyses of bovine mitochondria have been done by Fleoischer et al [34]. Their results also demonstrate that the lipid content of liver mitochondria less than those of heart and kidney mitochondria. Here we firstly demonstrated their difference of vibration spectra which indicated the molecular composition specially lipid content difference of mitochondria from liver, heart and kidney. We recorded more than 30 single mitochondria and obtained the average of Raman spectra of them. To verify the reproducibility of the data, the measurements were repeated five times for mitochondria isolated from rats and obtained similar results.

To determinate whether the Raman spectra can reflect the bioactivity of individual mitochondria, we have measured the Raman spectra changes between intact and swelling mitochondria. In our experiment, the original isolation media were substituted with high ionic strength KCL buffer (without EDTA), and 100 μM Ca²⁺ was added to induce mitochondria swelling. NIR Raman spectra of mitochondria swelling by Ca²⁺ induced was periodically collected to monitor the dynamical changes in the structural and molecular composition in mitochondria. And the fresh control mitochondria were also collected before addition of Ca²⁺. After Ca²⁺ addition, the Raman spectra of mitochondria appeared significant changes. The results in Fig. 2 demonstrate significant changes of main spectra bands after Ca²⁺ addition. In particular, five minutes after exposure of liver mitochondria to Ca²⁺ led to a large decrease in the magnitudes of the 1602 cm⁻¹ band. This band became weaker as time going on and eventually disappeared after 30 minutes Ca² addition. In contrast, the magnitudes of 1000, 1260, 1320, 1450 cm⁻¹ in response to the Ca²⁺ induced mitochondria swelling did not appear significant changes in the initial period. After 30 minutes the 1602 cm⁻¹ band disappeared and other bands became to decrease. And after 1 h at room temperature, most of Raman bands of mitochondria disappeared because of Ca²⁺ induced swelling.

 

Fig. 2. Comparison of NIR Raman spectra of intact and Ca²⁺ induced swollen mitochondria. Mitochondria isolated from liver (1 mg of protein) were suspended in the same buffer. Raman spectra of single trapped mitochondria at different times after addition of 100 μM Ca²⁺ to mitochondrial suspension.

Download Full Size | PDF

Detailed analysis demonstrates that the intensity of the characteristic Raman bands at 1602 cm⁻¹, 1655 cm⁻¹ and 1605 cm⁻¹ decrease gradually with different ratio as time elapse, which is plotted as a function of time in Fig. 3. The control experiment showed the reduction in 1602 cm−1 and 1655 cm−1 is the mitochondria’s response to the Ca²⁺ solution, rather than the laser damage effect (see Fig. 3(b)). The intensity decrease in the 1602 cm⁻¹ band after Ca²⁺ addition is probably related to a vital change in the structure and molecular composition of mitochondria. In the single mitochondrion, Ca²⁺ overloading changes the ion balance of inermembrane of mitochondria, consequently causes the morphological alteration of the mitochondrion and the ensuing release cytochrome C and other caspase cofactors into the extra mitochondrial medium. Therefore, the intensity of 1602 cm⁻¹ band could be used as a measurement of bioactivity of single mitochondria. Recently, Huang et al reported the time and space resolved Raman spectroscopy of mitochondria in situ within yeast cell. They reported that there was no 1602 cm⁻¹ band in the Raman spectra of purified biomolecules from mitochondria and named this band with “the Raman spectroscopic signature of life” [20, 21]. Naito et al used this band signature to observe the mitochondrial metabolic activity of the budding yeast cells [35]. Here we demonstrated this band changed associated with the state of single mitochondria and confirmed it could be used as a marker to reflect the bioactivity of single mitochondrion.

 

Fig. 3. (a). The averaged intensities of the mitochondria Raman bands 1602, 1446, and 1655 cm−1 as a function of exposure time to Ca2+. (b) The average intensities of 1602, 1446 and 1655 cm⁻¹ bands as the function of time since single mitochondria are trapped in the laser beam in buffer without adding Ca²⁺ solution. The smooth curve is a forth-order polynomial functional fit.

Download Full Size | PDF

4. Conclusion

We have applied the LTRS system for the monitoring real time changes of structure and bioactivity of single mitochondria. In this paper, we demonstrated the feasibility of using the Raman spectroscopy to detect the physiological state of mitochondria isolated from liver tissue, heart muscle and kidney. Furthermore we showed the complete Raman bands of isolated fresh or swelling single mitochondria. It is noteworthy that the 1602 cm⁻¹ band intensities in the Raman spectra were significantly decrease related to the bioactivity of mitochondria that consistent with published results. These data indicate that infrared Raman spectroscopy can be used as a noninvasive means to reveal the activation status of single mitochondria. We showed the potential of applying Raman microspectroscopy for study real time changes in biomolecular composition and bioactivity of single cell organelles. To our knowledge, this is the first study of the composition and bioactivity of individual mitochondria using Raman spectroscopy in an optical trap. One of the potential applications of LTRS lies in studying the effects of drugs and toxins on the mitochondria of living cells. Further investigation is being carried out on the composition and bioactivity changes of mitochondria in tumor cells response to drugs and toxins.