Dynamic Confocal Imaging of Living Brain. Advantages and risks of multiphoton microscopy in physiology

Transcription

1 Dynamic Confocal Imaging of Living Brain Advantages and risks of multiphoton microscopy in physiology Confocal laser scanning microscopy In conventional optical microscopy focused and out-offocus light is detected. The confocal principle suppresses all structures that fall outside the focal plane of the microscope objective, that is, outside of the focus. This is achieved by means of an aperture (detection pinhole, Fig. 1), which is placed in the beam path in an optically conjugated plane to the point light source. Therefore, using confocal laser scanning microscopy optical sectioning is possible. For acquisition of an image in the focal plane, the sample is scanned with a point laser beam in the x and y direction. Moving the sample along the optical axis (z direction) facilitates spatial imaging. A basic confocal laser scanning microscope consists of a microscope stand, the scan head and the lasers coupled into it. The scan head contains, among others, the following four elements. (1) The illumination unit with the main beam splitter for coupling the light from the laser (the excitation) into the beampath (Fig. 1). The beam splitter also separates the excitation from the returning light of the sample (the emitted epifluorescence). (2) The scanning unit with galvanometer-driven mirrors for moving the point laser beam over the fixed specimen in one optical plane, that is in the x and y directions. Fluorescence emitted from the sample is de-scanned Ulrike Tauer Leica Microsystems Heidelberg GmbH, Am Friedensplatz 3, D Mannheim, Germany (Manuscript received 6 August 2002; accepted 23 September 2002) Multiphoton microscopy is based on the simultaneous absorption of two photons emitted by a pulsed infrared laser source. In this technique, the excitation is restricted to a very small focus and thus results in optical sectioning a priori without the need of a confocal aperture. Multiphoton microscopy was introduced in live cell imaging as an alternative to confocal microscopy due to its superior qualities, such as the deep penetration depth, the reduced photodamage and the lack of out-of-focus bleaching. However, during the past years, examinations revealed severe limitations to the initial expectations. In the focal plane, photodamage and photobleaching can be worse than in single photon microscopy. However, studies showed that with low excitation intensity and by special technical adaptations photodamage could be avoided successfully. For functional biological imaging, multiphoton excitation provides an excellent tool such as the release of caged compounds in a diffraction-limited volume combined with multiphoton or confocal imaging. Experimental Physiology (2002) 87.6, through the scanning unit before it is guided to the pinhole and the detector unit. (3) The detection pinhole for suppression of the out-of-focus light (Fig. 1). The diameter of the pinhole is variable. A combination of the diameter of the pinhole, the wavelength and the numerical aperture of the objective determines the axial dimension of the optical section. (4) The detection unit for wavelengthspecific detection of the light emitted from the sample (Fig. 1). The different emissions from the individual fluorochromes are optically separated and fed to separate photomultipliers. This separation is achieved either by means of secondary beam splitters, a prism or a grating, depending on the design of the scan head. Transmission images can also be recorded in laser scanning microscopy. In this case, the fluorescence light that is transmitted through the specimen contributes to the imaging process. These detectors are named nondescanned detectors, because the light does not have to pass back through the scanning unit or the other optical components of the scan head for detection. In the strictest sense, this does not involve confocal images since the beam path no longer passes through the confocal pinhole. The design of confocal laser scanning microscopes and their different optical components are described in detail in the handbook of confocal microscopy (Pawley, 1995). Presented at a workshop held at the University of Bristol on 30 June 2001 supported by The Physiological Society, Leica Microsystems and Molecular Probes. Publication of The Physiological Society ulrike.tauer@leica-microsystems.com 2464

2 710 U. Tauer Exp. Physiol Fluorescence in single photon versus multiphoton excitation Fluorescence is a phenomenon in which absorption of light by a fluorophore is followed by the emission of the light of another wavelength. In traditional fluorescence, a single photon is used to excite a fluorophore from its ground state to an upper energy state, the excited state. When returning to the ground state, a photon of a lower energy (and thus longer wavelength) is emitted which can be detected as fluorescence light. The same fluorophore can be repeatedly excited and detected. The principle of multiphoton fluorescence is based on the simultaneous absorption of two or multiple long wavelength photons by a fluorophore to reach the excited state. Each Figure 1 The confocal principle with spot illumination and spot observation for fluorescence light. The excitation light (produced by the laser, dark grey) is focused onto a small aperture, the excitation pinhole, by a first lens. The beam splitter directs the light emerging from this pinhole towards the sample. The objective focuses the light onto the specimen. The fluorescent light emitted by the sample (light grey) passes through the beam splitter due to the fact that its wavelength is longer than of the excitation light. The emission is focused onto the detection pinhole. Only light that passes the aperture finally reaches the detector. This light is derived only from the focal plane. All out-of-focus light is suppressed by the detection pinhole. multiple absorption induces a molecular excitation of a magnitude equivalent to the sum of the absorbed photon energies. In the case of two-photon excitation, the sample is illuminated with a wavelength around twice that of the absorption peak. For example, the simultaneous absorption of two red photons (each 980 nm) is equivalent to that of a single blue photon (490 nm). The emission spectrum remains unchanged. Details are described in the handbook of confocal microscopy (Pawley, 1995). Multiphoton excitation and its integration into laser scanning microscopes The phenomenon of multiphoton excitation occurs only in a very restricted spatial focus (Denk et al. 1990). Therefore, fluorescent light originates only from this spot. Due to that fact, a confocal pinhole is not necessary for multiphoton microscopy. To use all emitted photons for detection, the confocal pinhole can be opened to its largest diameter. All photons including those photons scattered when passing through the sample contribute to the image. Because a confocal pinhole aperture is no longer needed, non-descanned detectors can be used (Denk, 1996). Fluorescent photons, which have been scattered, can be collected by such detectors and therefore significantly increase the efficiency of the yield (Fig. 2b and c). They are placed close to the sample and the light does not have to pass through all the elements of the scan head. The transmitted light detectors are located behind the condenser of the microscope stand similar to the widefield detection in light microscopy. In addition, reflected light detectors, which can be used only in multiphoton microscopy, are placed behind the objective to collect all epifluorescent light from the specimen. Reflected light detectors are especially advantageous if thick samples such as whole organism are examined. In this case, the fluorescent photons emitted from the upper tissue layers do not have to pass through the entire tissue on their way to the transmitted light detector. Therefore, the detection of the reflected epifluorescent light is significantly higher in thick samples (Fig. 2c). Multiphoton microscopes are equipped with a pulsed infrared laser, which is usually a titanium sapphire laser with a maximum tunable range between 700 and 1000 nm. These lasers have a very high peak power but low average power. Pulses have only a very short duration of either picoseconds (ps) or femtoseconds (fs). The short pulses (fs) are generally of higher peak intensity; however, the average power is greater for the longer pulses (ps). It is important to note that, especially with short pulses, the microscope optics themselves also cause pulse broadening when not compensated for. Thus a 100 fs pulse can be broadened to fs, while a 1.2 ps pulse is only broadened to about 1.3 ps, that is, the difference in the pulse duration at the sample is only a factor of 4 6 in multiphoton microscopy. To maintain the image quality, the average power has to be adjusted by a factor of (the fluorescence efficiency follows a P 2 /t dependence where P is the average power, and t is the pulse width; see König et al. 1999; Koester et al.

3 Exp. Physiol Multiphoton microscopy 711 Figure 2 Eye of a zebrafish embryo stained with DAPI to visualize the nuclei. A comparison between UV and multiphoton excitation and between detection with internal detectors of the Leica TCS SP 2 scan head and external, non-descanned detectors. With multiphoton microscopy, light penetrates deep into the tissue. The detection by external non-descanned detectors is more efficient than with internal detectors. Size of each image (a, b, c) is 125 mm w 125 mm. a, excitation with UV light and detection with an internal detector. b, multiphoton excitation at 750 nm and detection with an internal detector. The detector settings are the same detection range as for a. c, multiphoton excitation at 750 nm and detection with non-descanned detector for reflected fluorescence emission. Settings of the laser power are the same as for b and c. Images kindly supplied by S. Liebe of Leica Microsystems. Figure 3 Uncaging of nitro phenyl EGTA (npegta, a calcium chelator), which rapidly delivers calcium upon photolysis, in a cardiac myocyte loaded with Fluo-3 for calcium imaging. Uncaging was carried out at 800 nm with multiphoton excitation by means of a regionof-interest scan 10 s after starting the acquisition (the multiphoton light flash within the indicated area had a duration of ~6 s). During the experiment, the calcium indicator dye Fluo-3 was illuminated by visible light (488 nm). Five hundred images are taken within 30 s (duration for each image 62 ms). Here, only seven images are shown with 2.5 s between each of them. According to the colour scale, white and red colours represent high calcium concentrations, green represents medium and blue represents low concentrations. The experiment was carried out with the Leica TCS SP2 RS system, which is a laser scanning microscope with a resonant scanner for fast image acquisition. Images kindly supplied by D. A. Eisner, M. E. Díaz & A. W. Trafford of the University of Manchester.

4 712 U. Tauer Exp. Physiol ). The technology supporting picosecond pulses is significantly less complex and therefore less problematic than for femtosecond pulses. The repetition rates are 80 MHz, independent of the pulse duration. Advantages of multiphoton microscopy Multiphoton laser scanning microscopy is a promising imaging technique that has significant advantages for certain applications. (1) At the longer wavelengths used in two photon excitation, the scattering coefficient is lower than for the Figure 4 A pyramidal neuron filled with Oregon Green 488 BAPTA-1 using the patch clamp technique in an acute cortical slice. Multiphoton imaging at 880 nm excitation with the resonant laser scanning microscope Leica TCS SP2 RS. Figure kindly supplied by T. Nevian of the Max-Planck-Institut für Medizinische Forschung. wavelengths used in single photon excitation. The result is a reduction in loss of light, thus allowing imaging from structures deep within a tissue block (Fig. 2a and b). The latter is enhanced by the use of near-infrared excitation. In our experience, penetration up to 500 mm can be achieved. The penetration depth depends strongly on the tissue characteristics of the sample. (2) Excitation can only occur in the focal plane. Therefore, out-of-focus bleaching is avoided (Denk et al. 1990). In confocal microscopy excitation can take place also in the layers above and below the focal plane and can cause greater bleaching. (3) The ability of photo-uncaging of caged compounds (Fig. 3) in a diffraction-limited volume (Denk, 1996). One example is molecules that change their biological activity via a photochemical reaction. Such caged molecules have become readily available during the last two decades and are now widely used to deliver effector molecules. (4) Lower phototoxicity due to the longer wavelengths used (Denk et al. 1990). Therefore multiphoton microscopy provides longer imaging times. Risks of multiphoton microscopy In various applications with multiphoton microscopy, the following risks and disadvantages of this technique are now topics of contemporary discussion. (1) Photobleaching and photodamage. According to the most recent literature, the photodamage and photobleaching observed in multiphoton excitation can be greater than in single-photon excitation at least in the focal plane (for detailed summary see Drummond et al. 2002). Photodamage can lead to protein denaturation, oxidative stress or DNA damage. In cell cultures, cell viability is decreased (König et al. 1999). In calcium imaging of neuronal cells, photodamage can cause pathological and morphological changes (Koester et al. 1999). (2) Influence of pulse duration. Cell damage exhibits a dependence on P 2 /t, similar to the efficiency of twophoton fluorescence generation (König et al. 1999; Koester et al. 1999). This behaviour indicated that cell destruction is probably based on a two-photon excitation process rather than a one- or a three-photon event. According to the formula, as the pulse width becomes smaller the rate of cell damage increases. In addition, as the average power increases, the rate of cell damage increases. Both, fs and ps pulses provide the same optical window for safe twophoton fluorescence microscopy (König et al. 1999). With ps pulses it is possible to safely apply higher average powers to samples before cell damage occurs (König et al. 1999). However, it is also necessary to apply higher laser power to generate fluorescence. With fs pulses, cells can only withstand low average powers; however, the two-photon fluorescence generation is also higher. (3) Three-photon excitation. Since the operation in the fs mode requires the use of higher peak powers, one runs the risk of causing cell damage due to three-photon excitation

5 Exp. Physiol Multiphoton microscopy 713 (for example, three-photon excitation due to intense illumination at 780 nm results in excitation in the range of ~260 nm which can cause direct DNA damage). In addition, the high peak powers can result in plasma generation (i.e. the generation of a high number of free electrons), which causes cell damage accompanied by localized intense white luminescence and eventual carbonization. (4) Heating. In pigmented samples, single-photon absorbers such as haemoglobin, melanin or chlorophyll can cause severe physical cellular damage through heating. Measures for safe operation of multiphoton microscopy To carry out multiphoton microscopy successfully, several measures can be taken into account to minimize the overall exposure of the sample. (1) Using low excitation rates, photodamage caused by exposing the sample to intense light can be reduced. The excitation level should be kept as low as possible, consistent with obtaining a sufficiently large signal. (2) To obtain a sufficiently large signal, the focal volume can be increased. This can be achieved by underillumination of the back focal plane of the objective (Hopt & Neher, 2001). Underillumination means that the diameter of the laser beam is smaller than that of the back focal plane of an objective. In that case, the focus is enlarged in the axial direction and therefore increased in volume. Thus, a higher number of photons can reach the specimen for multiphoton excitation. As a result, more fluorescent signals are emitted. With Leica laser scanning microscopes (Leica TCS SP2 MP or Leica MP RS), the so-called variable beam expander can be used to easily adjust the diameter of the laser beam for underillumination of the objective. (3) External, non-descanned detectors placed in the reflection and transmission beam path can significantly increase the efficiency of the yield. In the case of transmitted light, a condenser is located between the specimen and the transmitted light detector. Standard microscopes are equipped with a condenser with a numerical aperture of 0.9. To further increase the detection efficiency, a condenser with a high numerical aperture should be used to capture more light. Oil immersion condensers with a numerical aperture of 1.4 are commercially available. If living tissue is being imaged then oil immersion medium is inappropriate so a water-dipping objective with a high numerical aperture can replace the condenser lens. APPENDIX Excitation wavelengths for various fluorescent dyes The fluorescent dyes are excited with pulsed infrared light of approximately twice the wavelength of the absorption maximum in the visible range (Table 1). For some fluorochromes, a certain blue shift in the shorter wavelength range is needed. The so-called cross-section describes how efficiently the fluorophore converts light Table 1. Single-photon and two-photon excitation range of several dyes * Fluorescent Single-photon Two-photon dyes excitation excitation Indo-1 UV nm Fluo nm nm DAPI *, Hoechst * UV nm FITC 488 nm nm Cy5 633 nm nm Rhodamin nm or 514 nm nm EBFP UV or 405 nm nm ECFP 430 nm or 458 nm nm Yellow Cameleon (FRET pair CFP/EYFP) 430 nm or 458 nm nm Fura-2 UV nm TRITC 543 nm nm Oregon Green 488 nm 840 nm EGFP 476 nm or 488 nm nm EYFP 514 nm nm * A three-photon excitation can be achieved especially with femtosecond laser pulses at higher wavelengths. EBFP, Enhanced Blue Fluorescent Protein; ECFP, Enhanced Cyan Fluorescent Protein; EGFP, Enhanced Green Fluorescent Protein; EYFP, Enhanced Yellow Fluorescent Protein. (excitation) to fluorescence (emission) depending on the excitation wavelength (Xu & Webb, 1996). Commercial multiphoton lasers have a maximum tunable wavelength range of nm. This range covers the multiphoton excitation of many fluorescent dyes used in live cell imaging including some of the dyes normally employed for UV microscopy (Fig. 2a). For example, using multiphoton excitation for the calcium indicators Fura-2 and Indo-1, it is possible to avoid the relatively phototoxic ultraviolet radiation. However, ratio imaging is not possible for Fura-2 because fast switching of two excitation laser lines cannot be achieved for technical reasons. In the author s experience, ratio imaging with Indo- 1 is not efficient with the standard setup of a multiphoton microscope. To detect a significant change in the ratio of the calcium-saturated form (detection range, ~400 nm) and the calcium-free form (detection range, ~475 nm), the multiphoton excitation wavelength should be tuned to a wavelength below 700 nm (Xu & Webb, 1996). A conversion of the standard mirror set of the cavity of the laser to gain shorter wavelengths leads to restrictions for longer wavelengths such as those required for Green Fluorescent Protein (GFP) (> 900 nm). Nevertheless, an excitation above 700 nm is sufficient for non-ratiometric imaging with Indo-1. For calcium imaging, Fluo-3 (Fig. 3), Oregon Green BAPTA (Fig. 4) and the Yellow Cameleon (a constructed protein with CFP and YFP for ratiometric measurements based on fluorescence resonance energy transfer, FRET; Miyawaki et al. 1997) are suitable and often-used fluorescent indicators (see Table 1 for the author s own observations of the given ranges of the two-photon excitation).

6 714 U. Tauer Exp. Physiol DENK, W. (1996). Two-photon excitation in functional biological imaging. Journal of Biomedical Optics 1, DENK, W., STRICKLER, J. H. & WEBB, W. W. (1990). Two-photon laser scanning fluorescence microscopy. Science 248, DRUMMOND, D. R., CARTER, N. & CROSS, R. A. (2002). Multiphoton versus confocal high resolution z-sectioning of enhanced green fluorescent microtubules: increased multiphoton photobleaching within the focal plane can be compensated using a Pockels cell and dual widefield detectors. Journal of Microscopy 206, HOPT, A. & NEHER, E. (2001). Highly nonlinear photodamage in two-photon fluorescence microscopy. Biophysical Journal 80, KOESTER, H. J., BAUR, D., UHL, R. & HELL, W. S. (1999). Ca 2+ fluorescence imaging with pico- and femtosecond two-photon excitation: signal and photodamage. Biophysical Journal 77, KÖNIG, H., BECKER, T. W., FISCHER, P., RIEMANN, I. & HALBHUBER, K.- J. (1999). Pulse length dependence of cellular response to intense near-infrared laser pulses in multiphoton microscopes. Optics Letters 15, MIYAWAKI, A., LLOPIS, J., HEIM, R., MCCAFFERY, J. M., ADAMS, J. A., IKURA, M. & TSIEN, R. Y. (1997). Fluorescent indicators for Ca 2+ based on green fluorescent proteins and calmodulin. Nature 388, PAWLEY, J. B. (1995). Handbook of Biological Confocal Microscopy, 2nd edn. Plenum Publishing Corporation, New York. XU, C. & WEBB, W. W. (1996). Measurement of two-photon excitation cross sections of molecular fluorophores with data from 690 to 1050 nm. Journal of the Optical Society of America B 13, Acknowledgements I wish to thank the following persons for the images: Dr S. Liebe at Leica Microsystems Bensheim (Fig. 2), Professor D. Eisner, Dr M. Díaz and Dr A. Trafford at the Unit of Cardiac Physiology, The University of Manchester (Fig. 3), and Dr Thomas Nevian from the Max-Planck-Institut für medizinische Forschung, Heidelberg (Fig. 4).