There are two fundamentally different approaches to imaging biochemical events over time. One strategy is to collect images from a series of fixed and stained tissues at different developmental ages. Each animal represents a single time point in the experiment. Alternatively, the same cells and tissues can be imaged in the living state during development. Here, the events of interest are captured directly. The second approach, imaging living cells and tissues, is technically more challenging than the first because great care must be taken to maintain cells in a healthy state during the time course of an experiment, with checks at the end of the experiment for cell damage.
Avoidance of Artefacts
The only way to eliminate artefacts from specimen preparation is to view the specimen in the living state. Many living specimens are sensitive to light, especially those labelled with fluorescent dyes. This is because the excitation of fluorophores can release cytotoxic free radicals into the cell. Moreover, some wavelengths are more damaging than others. Generally, the shorter wavelengths (higher-energy light) are more harmful than the longer ones (lower-energy light) and, therefore, near-infrared light rather than ultraviolet light is preferred for imaging (Figure 1).
Fig1. The visible spectrum – the spectrum of white light visible to the human eye. Our eyes are able to detect colour in the visible wavelengths of the spectrum; usually in the region between 400 nm (violet) and 750 nm (red). Most modern electronic detectors are sensitive beyond the visible spectrum of the human eye.
The intensity of light used for imaging must not compromise the cells. This is achieved using extremely low intensities of light, using relatively bright fluorescent dyes and extremely sensitive photodetectors. The viability of cells may also depend upon the cellular compartment that has been labelled with the fluorophore. For example, imaging the nucleus with a dye that is excited with a short wavelength will cause more cellular damage than imaging in the cytoplasm with a dye that is excited in the far red. For these reasons, it is necessary, when possible, to control for artefacts introduced by the imaging system itself.
Great care has to be taken in order to maintain the tissue in the living state on the microscope stage. A live cell chamber is usually required for mounting the specimen on the microscope stage. This is basically a modified slide and coverslip arrangement that allows access to the specimen by the objective and condenser lenses. It also supports the cells in a constant environment, and, depending on the cell type of interest, the chamber may have to provide a constant temperature, humidity, pH, carbon dioxide and/or oxygen levels. Many chambers have the facility for introducing fluids or perfusing the preparation with drugs for experimental treatments.
Time-Lapse Imaging
Time-lapse imaging continues to be used for the study of cellular dynamics. Here, images are collected at pre-determined time intervals (Figure 2). Usually, a shutter is placed in the light path so that the the specimen is only exposed to light when an image is collected and the amount of energy impacting the cells is reduced to an absolute minimum. When the images are played back in real time, a movie of the process of interest is produced.
Fig2. Time-lapse imaging of Caenorhabditis elegans development. Z-series were collected every 90 seconds of a developing C. elegans embryo genetically labelled with GFP- histone (nuclear material) and GFP- α tubulin (microtubules – cytoskeleton), and imaged with a spinning-disc confocal microscope. Each column consists of six optical sections collected 2 μm apart, and the columns are separated by 90 second increments of time. (Image kindly provided by Kevin O’Connell, National Institutes of Health, USA.)
Time-lapse is used to study cell behaviour in tissues and embryos, and the dynamics of macromolecules within single cells ( Figure 3c,d). The event of interest and also the amount of light energy absorbed and tolerated by the cells govern the time interval used. For example, a cell in tissue culture moves relatively slowly and a time interval of 30 seconds between images might be used. Stability of the specimen and of the microscope is extremely important for successful time-lapse imaging; importantly, the focus should not drift during the experiment.
Fig3. Contrast methods in the light microscope. (a) and (b) A comparison of bright-fi eld (a) and dark-fi eld images (b). Here, the bristles on the surface of the fl y appear dark on a white background in the bright-fi eld image (a) and white on a black background in a dark-fi eld image (b). The dark colour in the larger sensory bristles in (a) and (b) is produced. (c) and (d) Phase contrast view of cells growing in tissue culture (c) and (d). Two images extracted from a time-lapse video sequence (time between each frame is fi ve minutes). The sequence shows the movement of a mouse 3T3 fibrosarcoma cell (top) and a chick heart fibroblast. Note the bright ‘ phase halo’ around the cells. (e) and (f) Differential interference contrast (DIC) image of two focal planes of the multicellular alga Volvox (e) and (f). (Images (e) and (f) courtesy of the late Michael Davidson, Florida State University, USA.)
Whereas phase contrast is the traditional choice for imaging cell movement and behaviour of cells growing in tissue culture, differential interference contrast (DIC) or fluorescence microscopy is generally chosen for imaging the development of eggs and embryos. Computer imaging methods can be used in conjunction with DIC to improve resolution. Such enhancement is based on subtraction of a background image from each time-lapse frame in order to improve the contrast electronically. Such an approach has enabled the motility assays for motor proteins (for example, kinesin and dynein) whereby microtubules have been visualised on glass, assembled in vitro from tubulin in the presence of microtubule-associated proteins.
Photo Animation
The problems of presenting time-lapse series in a publication have been largely solved by the ability to publish QuickTime (a multimedia framework by Apple Inc.) movie fi les on the web pages of various journals. The software Photoshop (Adobe) also pro vides a bridge to additional image processing. For example, sequences of confocal images of different stages of development have been manipulated using Photoshop, and subsequently transferred to a commercially available animation program such as Morpheus (Morpheus Development LLC), and processed into short animated sequences of development. These sequences can be further edited and compiled using video editing software such as Final Cut Pro (now by Apple Inc.), and viewed as a digital movie using a video player directly on the computer or exported to DVD for presentation purposes.
Fluorescent Stains of Living Cells
Relatively few cells possess any inherent fluorescence (autofluorescence) although some endogenous molecules are fluorescent and can be used for imaging, for example, NAD(P)H. However, small fluorescent molecules can be loaded into living cells using many different methods, including diffusion, microinjection, bead loading, virus entry and electroporation. Larger molecules such as fluorescently labelled proteins are usually injected into cells.
Many reporter molecules are available for recording the expression of specific genes in living cells using fluorescence microscopy, as well as viewing whole trans genic animals using fluorescence stereomicroscopes (Table 1). In particular, the green fluorescent protein (GFP) is a very convenient and valuable reporter of gene expression, because it is directly visible in the living cell using epifluorescence light microscopy with standard filter sets.
Table1. Commonly used dyes in fluorescence microscopy
The GFP gene can be linked to another gene of interest so that its expression is accompanied by GFP fluorescence in the living cell. No fixation, substrates, coenzymes or cell-loading techniques are required. The fluorescence of GFP is extremely bright, and variants with enhanced brightness continue to be isolated or designed/ bio-engineered. Moreover, spectral variants of GFP and additional reporters such as DsRed are available for multiple labelling of living cells. These probes have revolutionised the ability to image living cells and tissues using light microscopy ( Figures 2 and 4).
Fig4. Multiple labelling in the living brain using the Brainbow technique. Unique colour combinations in individual neurons are achieved by the relative levels of three or more fluorescent proteins (XFPs – spectral variants of GFP). The images are collected using a multi-channel laser-scanning confocal microscope. Up to 90 different colours (neurons) can be distinguished using this technique. (a) Hippocampus. (Image courtesy of Jeff Lichtman, Harvard University, USA.) (b) Multiple labelling in living zebrafish skin using the Fishbow technique – developed from the Brainbow technique. Every cell on the surface of a zebrafish glows in a slightly different colour hue when viewed in a fluorescence light microscope. This method effectively gives each cell a unique identity, like a living bar code, that allows tracing over the eight-day lifespan. It is thus possible to simultaneously track hundreds of cells during the development and regeneration of fish tissues. (Chen-Hui Chen and Kenneth Poss, Department of Cell Biology, Duke University Medical Center, Durham, NC.)
Multi-Dimensional Imaging
The collection of Z-series over time is called four-dimensional (4D) imaging, where individual optical sections (x - and y -dimensions) are collected at different depths in the specimen (z -dimension) at different times (the fourth dimension), thus resulting in one time and three space dimensions (Figure 5). Moreover, multiple wavelength images can also be collected over time. This approach has been called 5D imaging. Software is available for the analysis and display of such 4D and 5D datasets. For example, the movement of a structure through the consecutive stacks of images can be traced, changes in volume of a structure can be measured, and the 4D data sets can be displayed as series of Z-projections or stereo movies. Multi-dimensional experiments can present problems with respect to handling large amounts of data, since it is not unusual to acquire gigabytes of information for a single 4D imaging experiment.
Fig5. Multi-dimensional imaging. (a) Single wavelength excitation over time or time-lapse X , Y imaging; (b) Z-series or X , Y , Z imaging. The combination of (a) and (b) results in 4D imaging. (c) Multiple wavelength imaging. The combination of (a), (b) and (c) comprises a 5D imaging experiment.
Scanned-Light-Sheet Microscopy
This method uses a thin sheet of laser light for optical sectioning with an objective lens and CCD camera detector system oriented perpendicular to it. The technique was developed to improve the penetration of living specimens and enables the imaging of live samples from many different angles at a cellular resolution. This approach is used by selective plane illumination microscopy (SPIM) where the specimen itself is rotated in the beam. Advantages of the technique include extremely low photo dam age and high acquisition speed. It has been used to image every nucleus in zebrafish embryos over 24 hours of development at stunning resolution.
Super-Resolution Methods
Several advanced methods are now breaking the theoretical resolution limit of the light microscope, as first proposed by Abbe in the early 1800s. Up until recently, the dogma was that the limit of resolution of the light microscope was dependent on the wavelength of light used, and was fixed at around 0.5 μm. Better resolution could only be achieved using electron microscopy, and, consequently, only fixed/non-living specimens could be imaged at higher resolution.
New super-resolution light microscopes are able to achieve resolutions down to 0.1 μm in the lateral dimension and 0.6 μm in the axial direction, and in living cells. Such instruments include fluorescence photoactivation localisation microscopes (FPALMs), stimulated emission depletion microscopes (STEDs), stochastic optical reconstruction microscopes (STORMs) and 3D structured illumination microscopes (SIMs). All of these methodologies make use of the excitation and emission properties of fluorescent probes, together with imaginative instrumentation and computer programs to break the diffraction limit. They provide valuable information on macromolecular dynamics in living cells that are not available from conventional microscopy techniques (see, for example, Figure 6).
Fig6. Comparison of images obtained by confocal and super-resolution microscopy. Tetrahymena cells labelled with antitubulin polyglycylation antibody and rhodamine secondary antibody. The confocal microscopy image (left) shows a uniform signal throughout the cilia, whereas the super-resolution microscopy image (right) shows that poly-glycylated tubulin is present in the outer doublets of the cilia only and not in the central pair microtubules. The cilia show two lines of poly-glycylated tubulin, which are the doublet microtubules. The central space occupied by the central singlet microtubule lacks the red poly-glycylated tubulin signal. (Image kindly provided by Mayukh Guha and Jacek Gaertig, University of Georgia, USA).
Whole-Animal Methods
Various instruments have been designed over the years for imaging cells in living animals. There are two main approaches; mini microscopes that can be mounted on an animal for long-term observations or hand-held probes that can be pressed against an animal for immediate diagnostic imaging. This continues to be an area of ongoing efforts, mainly resting on the development of new lenses for efficient light capturing and fibre-based endoscopes that can capture the signal in vivo .
