1. Introduction

Optical trapping is a powerful tool that uses optical forces to hold a micrometer-sized particle near the focus of a single tightly focused laser beam in three dimensions [1]. This technique has been used to capture and manipulate biological particles such as cells, bacteria, virus, and chromosomes without obvious damage using near-infrared laser beams [2, 3]. Biomechanical properties of single cells and single biomolecules have been extensively studied with optical tweezers [4–6]. The trapped particles can, for example, be set into rotation by controlling the angular momentum of the laser beam to form micromachines [7, 8]. The combination of Raman spectroscopy with optical trapping offers a new degree of direct analysis of chemical constituents of microscopic particles [9–12]. In this paper, we report on real-time Raman spectroscopic studies of single optically trapped living cells or single organelles for the purpose of the analysis and identification, without the need of mechanical immobilization or introducing biochemical stains or tags. With the introduction of vibrational spectroscopy [13–17], the constitution and conformation of macromolecules inside the trapped cells can thus be identified and studied. The essence of this paper is to demonstrate that this technique is an effective method to the study in real-time changes in biochemical properties of a single cell (or organelle) as the function of the changes in the temperature of the medium and the technique could also be used as a sensor for rapid identification of microbes in a fluid.

It is known that biological cells contain a very complex mixture of organic, inorganic and biochemical components enclosed in cell membrane. A large number of biological molecules such as proteins, nucleic acids, polysaccharides, and lipids etc. and a variety of electrolytes populate the interior of the cell. The constitution and conformation of these biomolecules in a living cell depend on the nature of its physiological state and thus may change as a function of time in the presence or absence of external chemicals or physical agents such as temperature and/or light. It is important to notice that the nature of the biochemical properties and processes may vary from one cell species (or strain) to another. Confocal Raman spectroscopy has been applied for the analysis of chemical constituents of single living cells in different physiological stages [13–17, 21, 22] and for identification of some microorganisms [18, 19]. The main drawback of the conventional method is that the living cells under study must be immobilized either on a glass cover slip or in a solid culture medium in order to avoid Brownian motion or cell motility when suspended in a liquid medium. Cell immobilization using a cover slip changes the chemical and physical micro-environment of this living microorganism and may yield among others changes in the electrochemical potentials across the cell membrane which is known to affect cellular functions. In order to avoid this problem and consequent artefacts, we used the LTRS system that combines optical trapping and confocal Raman spectroscopy. In our scheme, optical trapping allows a mobile particle to be held at the focus of a laser beam for long-time observation and moved away from the coverslip up into solution. Furthermore, levitation of the particle in the bathing solution causes a reduction in fluorescence and stray scattering interference from the coverslip thus undermining the weak Raman signals from the cell. In addition, optical trapping permits optimum excitation and collection for Raman scattering in a confocal configuration because the particle is maintained in the focus of the laser beam and within the focal plane of the objective. Although some degree of photodamage to the living cells may be caused by the trapping beam, this effect can be decreased substantially by using low power and selecting near-infrared laser wavelength [2].

2. Experimental setup

The configuration of our laser tweezers and Raman spectroscopy (LTRS) instrument is shown in Fig. 1. A laser beam at 785 nm from a wavelength-stabilized, beam shape-circulated semiconductor diode laser is introduced into an inverted microscope (Nikon 2000S) through a high numerical aperture objective (100×, N.A.=1.30) to form an optical trap. The same laser beam can be used to excite Raman scattering of the trapped particle. The scattering light from the particle is collected by the same objective and coupled into the spectrograph through a 200-µm pinhole, which enables confocal detection and rejection 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 cell. The spectrograph is equipped with a liquid-nitrogen-cooled charged-coupled detector (CCD). The spectral resolution of our confocal Raman system is about 6 cm^-1. To reduce the effective laser power illuminating on the trapped cell, the near-IR diode laser can be operated in a power-switching scheme that sets a low power for trapping (~2 mW at the sample) and a high power (~15 mW) for a short period when Raman measurement is obtained. After the spectrum is recorded, the laser returns to low-power operation for trapping only. The laser spot size at the focus is approximately ~2 µm in diameter and the depth resolution is about 2–3 µm, determined by the pinhole. The cell is levitated approximately 15 µm above the coverslip to reduce the fluorescence background from the substrate. A major portion of a rod-like bacterium is exposed to the excitation beam. To increase the sampling area for a large cell, the excitation beam (at 785 nm) can be steered up to a 10-µm size in the lateral dimension by rapidly rotating a slightly tilted mirror, while an infrared trapping beam at 1064 nm from a continuous-wave diode-pumped YAG laser can be used to hold the cell.

 

Fig. 1. Experimental setup of the combined laser tweezers and Raman spectroscopy system. Abbreviations: M, mirror; L, lens; DM, dichroic mirror; PH, pinhole; HNF, holograph notch filter.

Download Full Size | PDF

3. Results and discussions

The LTRS system allows manipulating and identifying single organelles within a living cell. Fig. 2(a) shows Raman spectrum of the nucleus of a pine plant cell that was cultured to a stationary phase. The medium for pine cells suspension is 0.5% saline solution and the buffer for bacteria is a standard chemically defined medium (5g-glucose, 1g-NH₄H₂PO₄, 5g-NaCl, 0.2g-MgSO₄, and 1g-K₂HPO₄ in 1 liter water). The nucleus of the pine cell can be captured and moved within the cell interior. We accumulated the Raman spectra of nucleus and vacuoles. The Raman spectrum of the nucleus can be easily identified from that of other organelles such as vacuoles, which have very different spectra (not shown).

The LTRS system allows for rapid identification of single bacterium on the basis of the relative abundance and the structure signature of its biomolecules. This is done without the need for performing time-consuming laboratory incubation procedures in order to obtain large number of cells to which certain stains are introduced for each chemical analysis. Fig. 2(b–d) depicts the Raman spectra for three types of bacterial cells, Bacillus sporeformer, Streptococcus salivarius, and E. coli cultured to stationary phase in a LB medium. The samples were obtained from the America type culture collection (ATCC, Manassas, VA). The dimension of rod-like bacteria E. coli and bacillus sporeformer is about 2 µm in length and 0.5 µm in diameter, and the sphere-shape S. salivarius is about 0.5–1.0 µm in diameter.

 

Fig. 2. NIR Raman spectra of (a) a single trapped nucleus of a pine cell, (b) Bacillus sporeformer, (c) Streptococcus salivarius, and (d) E. coli bacterium suspended in a liquid medium. The pine nucleus is moved to the left from near the center of the cell, as indicated by the arrow in the image (a). The background spectrum measured without the cell in the trap has been subtracted for each spectrum. Spectrum (a) is average result of eight pine cells and spectrum (b–d) is the average result of 20 bacterial cells, respectively. Experimental conditions: trapping power 5 mW, Raman excitation power 15 mW; acquisition time 30 s for curve (a) and 60 s for curves (b, c, d). Raman line assignments are based on assignments in [13, 15, 16, 18]. Abbreviations: Phe, phenylalanine; Tyr, tyrosine; A, adenine; C, cytosine; G, guanine; BK, backbone; p, protein; def, deformation; str, stretching.

Download Full Size | PDF

When a small bacterium is trapped in the laser beam, the entire bacterial cell will locate inside the focused laser beam generates Raman signal. It is well known that the majority of the bacteria of interest move rapidly in the medium so that Raman spectrum from an individual bacterium cannot be obtained without optical trapping. A comparison between these spectra shows that the ratio between the heights of some specific bands is different for different types of bacterial. For example, the height ratio between 785 cm^-1 band (assigned to nucleic acids) and 1004 cm^-1 band (assigned to phenylalanine) is approximately 1.35:1.0 for Bacillus sporeformer, 1.45:0.72 for Streptococcus salivarius, and 0.72:1.09 for E. coli, respectively. The discrimination of the spectra could be used to distinguish these bacteria.

The LTRS system can be used to study the dynamic change in conformation or structure of biomolecules inside a trapped cell due to changes in the ambient temperature. The use of extreme temperature to control the growth of microbes is widely employed in microbiology. Generally, when the temperature is increased, proteins inside the living cells could be denatured and cell death will occur eventually at the elevated temperature. However, this irreversible process has not been observed by molecular vibration spectroscopy in real time for an individual cell. In our experiment, a living yeast cell is held in the beam focus while the medium’s temperature is changed slowly at a rate of 5 °C per 5 min by a temperature-controlled microscope stage, which took about 2 min to reach the set temperature. The temperature is first increased from 25 °C to 80 °C with a step of 5 °C and then decreased back to 25 °C. During the temperature change process, Raman spectrum from the target cell is periodically collected to monitor the dynamical change in the structure or conformation of molecules inside the cell. The measurements were repeated and averaged for 15 cells. Intensity scale for all spectra is the same and the background has been subtracted. The yeast cells (Fleischamann’s, active dry, Hubbard Scientific) were incubated in a liquid culture medium (2% galactose, 2% glucose, and 2% peptone) for more than 4 hours at 32 °C before measurement. The yeast suspension was diluted with the same medium for measurement.

 

Fig. 3. Raman spectra of a single yeast cell at different temperatures. A living yeast cell is held in the beam focus while the medium’s temperature is changed slowly. Experimental conditions: trapping power 2 mW, Raman excitation power 16 mW; acquisition time 20 s. Raman band assignments are based on the assignments in [13, 15, 16, 18].

Download Full Size | PDF

Figure 3 shows the Raman spectra of the trapped yeast cells as the temperature is changed. We found that as the temperature was slowly increased from 25 °C to 50 °C, Raman spectrum of the same cell appeared with no obvious change. However, as the temperature was increased up to 70 °C, significant changes in spectra were observed. For example, the magnitudes of the 1004 cm^-1 band (assigned to the aromatic ring of phenylalanine) and that of the 1602 cm^-1 band (assigned to the C=C bond of phenylalanine and tyrosine [13]) were obviously increased, while the magnitudes of 1443cm^-1 (assigned to the CH bond in lipid and proteins) and 1656 cm^-1 (assigned to amide I of proteins) bands experienced little change. As the temperature was decreased from 70 °C down to 25 °C, the magnitudes of 1004 cm^-1, 1602 cm^-1, and 1443 cm^-1 band keep high while the magnitude of 1656 cm-1 remained almost the same.

Figure 4 plots the dependence of the signal intensity of the 1004 cm^-1 band on the ambient temperature. The results demonstrate that the magnitude of the 1004 cm^-1 band shows no obvious change below 60 °C and changes rapidly above this critical temperature.

 

Fig. 4. Signal intensity of the 1004 cm^-1 band as the function of temperature. The smooth curve is an exponential function fit.

Download Full Size | PDF

The intensity increase in the 1004 cm^-1 and 1602 cm^-1 band at the elevated temperature is probably related to a vital change in the environment of the Phe/Tyr side-chain of yeast proteins due to heat-denaturation [20]. In a living cell, the yeast protein folds in an ordered structure and the Phe/Tyr side chain in the yeast protein is “buried” or “masked” so that the intensity of the excited vibration of the Phe/Tyr is low. When heated, the protein unfolds since hydrogen bonds are disrupted and, therefore, the Phe/Tyr side chain in the yeast proteins is “exposed”. The exposed Phe/Tyr yields a large Raman intensity at 1004 cm^-1 nd 1602 cm^-1 vibration bands. In addition, after heated, some Phe/Tyr enriched micro-organelles may have more chance to accumulate in the focused laser beam and generate large Raman intensities. The fact that the 1004 cm^-1 and 1602 cm^-1 band intensity did not decrease as the temperature was decreased from 70 °C down to 25 °C indicates that the heat-denaturation process of proteins in a yeast cell is irreversible. Therefore, the 1004 cm^-1 and 1602 cm^-1 band intensity could be a measure of protein denaturing activity in a living yeast cell. Recently, Huang et al. reported the time- and space-resolved Raman spectroscopy of single living yeast cells [21, 22]. The 1602 cm^-1 band intensity was found to be high for certain stages of yeast life and may reflect the respiratory activity in mitochondria.

4. Conclusion

We have applied the LTRS system for the study of real-time or in-vitro Raman spectroscopy of optically trapped living cells and interior organelles. We have measured the Raman spectra of the nuclei within living pine cells, as well as Bacillus sporeformer, Streptococcus salivarius and E. coli bacteria cultured to stationary phase and suspended in solution. This study showed the potential of using the LTRS system as a sensor for rapid identification of microbial cells in a fluid. In a heating experiment, we have measured the real-time Raman spectra of living yeast cells as the function of the elevated temperature of the surrounding medium to investigate the process of the cell life to death at the molecular level. We found that the 1004 cm^-1 and 1602 cm^-1 band intensities in the Raman spectra were significantly increased as the temperature was slowly increased from 25 °C to 80 °C, but not changed back to the original level as the temperature was decreased back to 25 °C. These band intensities could be used as a measure of irreversible heat-denaturing process of proteins in a living yeast cell. Our experiments show the feasibility of applying the LTRS system in rapid identification and perhaps separation of microorganisms, in-vitro analysis and manipulation of organelles within living cells, and the study of dynamic biochemical processes on a single cell level.