Specimen Preparation and Imaging
The procedures for preparing and imaging specimens in the confocal microscope are largely derived from those that have been developed over many years for use with the conventional wide field microscope. The best approach in developing a new protocol for a specimen to be imaged with the confocal microscope is to begin with one known to be appropriate for conventional microscopy, and to modify it as necessary.
Regardless of the specimen preparation protocol employed, a primary benefit of the manner in which confocal microscopy is carried out is the flexibility in image display and analysis that results from the simultaneous collection of multiple images, in digital form, into a computer. This is discussed in more detail below, but one elegant example of the image display possibilities is presented in Figure 1, a triple-labeled Drosophila embryo at the cellular blastoderm stage. The specimen was immunofluorescently labeled with antibodies to three different proteins. After three corresponding images were collected in the red, green, and blue channels of the confocal system, the images could be rearranged by copying them to different channels. By evaluating the image resulting from merging the three, the most effective color-to-channel assignment for illustrating the various protein domains was chosen. Figure 1 presents the merged three-channel image (combined red, green, and blue channels).
Most of the methods that have proven successful in preparing specimens for the conventional wide field optical microscope are specifically aimed at reducing the amount of out-of-focus fluorescence, since this produces flare in the image that greatly reduces the resolution of the features of interest. Due to the optical sectioning achieved by the confocal approach, the confocal microscope undersamples the fluorescence in a thick specimen as compared to the conventional epifluorescence microscope. The result is that samples may require increased staining times or stain concentrations for confocal analysis, and if evaluated in the conventional microscope, may appear to be over stained.
Although the illumination in the typical laser scanning confocal system appears to be extremely bright, the average illumination at any given point on the specimen is relatively moderate, due to the fact that many points are scanned per second. At a typical scan speed of one point per 1.6 microseconds, the actual illumination at any given point is generally less than in a conventional wide field epifluorescence microscope. It is usually advisable to use the lowest laser power that is practical for imaging in order to protect the fluorophore. Although many protocols include an antibleaching agent to prevent fading of the fluorescent species, such additives may not be required with many of the more modern confocal instruments.
The primary advantage and application of the confocal microscope is in improved imaging ability of thicker specimens, although the success can be limited by the specific properties of the specimen. Certain minimum physical requirements for the specimen apply; it must fit on the microscope stage, and the area of interest must be able to be placed within the working distance range of the objective lens. In some cases resolution may have to be compromised in order to accommodate a specimen, and to avoid damage to it or the objective lens. For example, a high resolution lens such as a 60x having a numerical aperture of 1.4 may have a working distance of 170 micrometers, whereas a 20x (having a typical numerical aperture of 0.75) might offer a relatively large 660 micrometer working distance, with the ability to access more restricted areas of a specimen without physical interference.
Specimens that have three-dimensional structure that is to be studied with the confocal microscope, have to be mounted in such a way as to preserve the structure. Some sort of spacer, such as fishing line or a piece of coverslip, is commonly placed between the slide and the coverslip to avoid deforming the specimen. When living samples are to be studied, it is usually necessary to mount them in a chamber that provides all of the necessary requirements for life, and that will also allow sufficient access by the objective lens to image the desired area.
Objective Lens Parameters and Optical Section Thickness
Specimen properties that affect light transmission, such as opacity and turbidity, can greatly influence the depth of penetration of the laser beam into the specimen, and consequently the structures that can be imaged. Unfixed and unstained corneal epithelium of the eye, for example, is relatively transparent and a laser beam will penetrate it to a depth of about 200 micrometers. In contrast, unfixed skin is relatively opaque and scatters more light, limiting the laser penetration to about 10 micrometers. Many fixation protocols include some form of clearing agent intended to increase the transparency of the tissue.
If sufficient laser penetration cannot be achieved with a whole mount specimen, thick sections can be cut using a microtome. Fixed tissue is usually used for sectioning, but tissue (such as living brain) has been cut by vibratome and successfully imaged. To gain access to deeper parts of a section mount, it is possible to remove the specimen from the slide, invert it, and remount it, but this is usually not very successful. Images from a somewhat deeper part of a specimen can be obtained by using dyes that are excited at longer wavelengths (such as cyanine 5), as opposed to those that require shorter wavelength excitation. The use of longer wavelength illumination will, however, slightly reduce the maximum resolution that can be achieved in comparison to images acquired at shorter wavelengths. For similar reasons, multiple-photon imaging techniques allow images to be collected from deeper levels within a specimen (due to the use of red light for excitation).
The Objective Lens
For confocal microscope studies, the choice of the objective lens used is extremely important, as the light-collecting ability of the lens, measured as its numerical aperture, is a determinant of both resolution and optical section thickness. Holding other microscope variables constant, the higher the objective numerical aperture is, the thinner the optical section will be. As examples for one particular instrument, the optical section thickness using a 60x (numerical aperture of 1.4) objective with the pinhole diameter set at 1mm is approximately 0.4 micrometer, and with a 16x (numerical aperture of 0.5) lens, with the same 1-millimeter pinhole, the section thickness is on the order of 1.8 millimeters. By opening the pinhole to a larger diameter, the optical section thickness can be increased. Table 1 gives optical section thickness values (in micrometers) for various objective lenses at two different pinhole diameters for one model of LSCM. The image resolution is always poorer vertically than it is horizontally. For example, using the 60x, 1.4 numerical aperture, objective lens the horizontal resolution is approximately 0.2 micrometer, and the vertical resolution about 0.5 micrometer. Flatness of field and chromatic aberration are additional lens characteristics to be considered when choosing an objective lens. The degree of chromatic aberration correction is particularly important when imaging multilabeled specimens at different wavelengths.
Objective lenses that are capable of the highest resolution generally are those with the highest magnification and the highest numerical aperture. They are also the most expensive, so a compromise is often made between the area of the specimen that is scanned and the maximum resolution that can be achieved for that area. If imaging insect embryos and imaginal disks, for example, a 4x lens might be used to locate the specimen on the slide, a 16x (numerical aperture of 0.5) lens for imaging whole embryos, and a 40x (numerical aperture of 1.2) or 60x (numerical aperture of 1.4) lens for resolving individual cell nuclei within embryos or imaginal disks. For imaging larger tissues, such as butterfly imaginal disks, the 4x lens would be useful for whole wing disks, and the 40x or 60x for resolution of individual cells. Figure 2(a) illustrates use of a 4x objective to obtain an overall view of an entire butterfly fifth instar wing imaginal disk, and the additional nuclear detail provided by a 16x lens (Figure 2(b)). One way in which high resolution and large image fields of view can be combined is to acquire many images from adjoining areas and to combine them digitally into montages. Some microscopes have automated x-y stages that can be set up to move around the specimen and collect multiple images into a large-area montage.
One of the more useful features that is characteristic of most LSCMs is the ability to zoom an image using the same objective lens, with no loss of resolution. This capability is achieved simply by decreasing the area of the specimen scanned by the laser, by control of the scanning mirrors, while maintaining the same image display size or memory storage array size, effectively increasing the magnification. In this way several magnifications are achieved with a single lens without disturbing the specimen or losing track of reference points in the field of view. Whenever possible, however, a lens of higher numerical aperture should be used to maximize resolution, rather than zooming using a lens of lower numerical aperture. The capability of zooming with a single lens (40x) is illustrated in Figure 2(c-f). Panel (c) of the figure shows the additional nuclear detail provided by the 40x objective (compared to the 16x view of panel (b)). Panels (d) through (f) are images obtained by zooming the same 40x lens by progressive increments, accomplished by reducing the area scanned on the specimen.
A number of confocal instrument designs have an adjustable pinhole that limits the out-of-focus light that reaches the detector. Opening the pinhole to a larger diameter produces a thicker optical section and reduced resolution, but is often necessary to include more specimen detail or to increase the light striking the detector. As the pinhole is closed (diameter reduced) the optical section thickness and brightness decrease. Resolution increases until a certain minimum pinhole diameter is reached, beyond which resolution does not increase but brightness continues to decrease. The pinhole diameter at which this condition is reached is different for each objective lens.
Probes for Confocal Imaging
The development of confocal instrumentation continues to both influence and be influenced by the synthesis of novel fluorescent probes that improve immunofluorescence localization. Fluorochromes are being introduced that have excitation and emission spectra more closely matched to the wavelengths produced by the lasers supplied with most commercial LSCMs. Improved probes that can be conjugated to antibodies of current research interest are continually developed. As one example group of dyes, the cyanines have developed as alternatives to other long-established dyes, with cyanine 3 as a brighter alternative to rhodamine, and cyanine 5 finding increased use in triple-label strategies.
Fluorescence in-situ hybridization (FISH) has advantages in resolution and sensitivity of probe detection that are further enhanced when coupled with the LSCM, and is a valuable approach for imaging the distribution of fluorescently labeled DNA and RNA sequences in cells. In addition, brighter fluorescent probes are currently available for LSCM imaging of total DNA in both nuclei and isolated chromosomes.
A large number of fluorescent probes are available that, when incorporated in relatively simple protocols, specifically stain certain cellular organelles and structures. Among the plethora of available probes are dyes that label nuclei, the Golgi apparatus, the endoplasmic reticulum, and mitochondria, and also dyes such as fluorescently labeled phalloidins that target polymerized actin in cells. Such dyes are very useful in multiple labeling approaches to locate antigens of interest having specific compartments in the cell. For example, Figure 3 presents the employment of a combination of phalloidin and a nuclear dye (ToPro) with the appropriate antigen in a triple labeling scheme applied to whole mounts of butterfly pupal wing imaginal disks. As illustrated in Figure 3(a), phalloidin can be used to accentuate cell outlines in developing tissues, with the peripheral actin meshwork being labeled as bright fluorescent rings. Panel (b) of the figure illustrates the dramatic specificity of the nuclear dye in labeling just that one cellular component. In addition to this cellular compartment labeling strategy, antibodies to proteins of known distribution or function in cells (such as antitubulin) can be usefully included in multilabel studies.
If living cells are being imaged, it is critical to be aware of the effects of adding fluorochromes to the system. These probes can be toxic to living cells, especially when they are excited with the laser. Toxic affects are reduced in some preparation protocols by the addition of ascorbic acid to the cell medium. The particular cellular component that is labeled can affect the viability of the cells during imaging. For example, stains for the cellular nucleus tend to have more deleterious effects than do cytoplasmic stains. Probes are available that distinguish between living and dead cells (among these are acridine orange), and that can be used in assays of cell viability during imaging. Most such assays are based upon the premise that the membranes of dead cells are permeable to many materials, such as the dyes, that cannot penetrate them in the living state.
Fluo-3 and rhod-2 are examples of dyes that have been synthesized to change their fluorescence characteristics in the presence of certain ions such as calcium. New probes have been developed for imaging gene expression, including for example, the jellyfish green fluorescent protein (GFP), which enables gene expression and protein localization to be observed in vivo. The use of GFP has enabled the monitoring of gene expression in a number of different cell types including living Drosophila oocytes, mammalian cells, and plants (using the 488nm line of the excitation laser of the LSCM). Mutants of GFP with spectral variations are available for use in multilabeling experiments, and these have also found use for avoiding interference from autofluorescence in living tissues.
Autofluorescence of tissues occurs naturally in many cell types, and can be a major source of background interference during imaging. For example, chlorophyll in yeast and plant cells fluoresces in the red part of the spectrum. Certain reagents, such as glutaraldehyde fixative, are sources of autofluorescence, which can be reduced by treatment with borohydride. Autofluorescence can sometimes be avoided by using fluorophores that can be excited at wavelengths that are out of the range of natural autofluorescence. Cyanine 5 is often chosen since it is excited at a longer wavelength that avoids the shorter-wavelength autofluorescence.
Although it is most often considered a problem, tissue autofluorescence can be utilized for imaging overall cell morphology as a part of multilabeling studies. The contribution from autofluorescence to the total fluorescence can be assessed by viewing an unstained specimen at different wavelengths and noting the laser power and PMT settings of black level and gain. Autofluorescence can often be bleached out by brief exposure to the laser at high power, or by flooding the specimen with light from a mercury lamp. More sophisticated approaches to dealing with autofluorescence include using time resolved imaging, or removal using digital image processing techniques such as image subtraction.
Beginning users of confocal microscopes can gain experience in several ways. The microscope manual provided by the manufacturer usually includes a series of simple procedures necessary for getting started. In most multi-user facilities, the person primarily responsible for operating the instrument may provide orientation sessions, or the facility manager may require a short training session and demonstration of a certain competence level before solo use of the instrument is allowed. Particular attention should be paid to the house rules of the facility. Information and training can also be gotten from training courses conducted by the microscope companies, from workshops on microscopy, and from a variety of publications.
Before work is done with experimental specimens, it is essential to be familiar with the basic operation of the imaging system. It is usually beneficial for the novice to begin trial imaging with a relatively easy specimen rather than a more difficult experimental one. Some better test samples include paper soaked in one or more fluorescent dyes or a preparation of fluorescent beads. Both types of specimen are brightly fluorescent and relatively easy to image with a confocal system. Another excellent sample is a slide of mixed pollen grains that exhibit autofluorescence at many different wavelengths. These can be easily prepared from pollen collected from garden plants, or can be obtained from commercial suppliers of biological specimens. The pollen grain images in Figure 4 were collected simultaneously with the same PMT black level and gain, and pinhole diameter settings, but reveal three types of pollen that each fluoresce at different excitation wavelengths. These specimens are valuable as test subjects because they not only have some interesting surface details, but also maintain their properties relatively well when exposed to the laser beam. For trials with living tissue, specimens prepared from onion epithelium or the water plant Elodea sp. are reliable, using either autofluorescence or staining with DiOC6.
Before attempting imaging, the confocal instrument should be set up to give the best possible performance. This requires optimal alignment, especially when one of the older confocal microscopes is being used. The alignment routine used is highly specific to the particular instrument, and is usually best done by the person who has overall responsibility for maintaining it. In no case should alignment be attempted without proper training and permission from the owner of the microscope. Improper procedures used to attempt alignment can result in complete loss of the beam and can, in the case of some instruments, require a service visit to rectify.
The confocal system is based on a conventional optical instrument, and the fundamental procedures and practices of optical microscopy should be followed at all times. It is extremely important that all glass surfaces in the optical path be clean because dust, oil, or grease on slides, coverslips, and objective lenses is a primary cause of poor images. The refractive index between the objective lens and the specimen must be appropriate to the lens in use. For example, the correct immersion oil must be used for a given objective numerical aperture, and the specimen must be mounted to be within the working distance of the lens. Coverslip thickness must be correct for the lenses used, especially for the higher power objectives, which require a No. 1 or No. 1.5 coverslip instead of a No. 2. The coverslip must be sealed to the slide using an appropriate medium, and mounted flat. Nail polish can be used for fixed specimens if care is taken to ensure that it is dry before imaging. A mixture of petroleum jelly, beeswax, and lanolin, or some other nontoxic sealing material must be used with living specimens. Following strict basic cleanliness procedures at the specimen preparation stage can save much time and effort later.
In preparation for confocal mode imaging, a region of interest is located using either brightfield or conventional epifluorescence microscopy. It is preferable to do this survey using the microscope of the confocal system, but it can be extremely difficult for the novice to find the correct focal plane using the confocal imaging mode alone. If conventional imaging modes are not available on the confocal system, then structures of interest can be located using a separate fluorescence microscope, and their positions marked using a diamond marker on the microscope, a marking pen, or by recording the position coordinates from the microscope stage. It is especially useful to be able to preview specimens with the actual microscope of the confocal system when attempting to image a rare phenomenon such as a gene expressed at a particular stage of development in a specimen containing perhaps hundreds of embryos of different ages. A great deal of time can be saved in this way, over having to scan many specimens using the confocal mode. Confocal instruments commonly have a low-resolution rapid scanning mode that makes the preliminary scanning more efficient. The best approach when searching for rarely occurring events, however, is to scan the slides using a conventional microscopy mode and then to immediately switch to confocal mode on the same microscope to collect the images.
Successful confocal imaging relies on the "secret" of mastering the interplay between objective lens numerical aperture, pinhole size, and image brightness, and using the lowest laser power possible to achieve the best image. The novice user should experiment with varying these parameters using test specimens and several objective lenses of different magnifications and numerical apertures to gain a sense of the capabilities of the microscope before attempting imaging on experimental specimens. A comparison should be made of images acquired using the zoom function of the confocal system with those obtained using objectives having higher numerical aperture. The particular specimen and features being imaged will determine which lenses and methods are most appropriate. Two examples of the many objectives suitable for confocal microscopy are illustrated in Figure 5. The figure includes a 60x planapochromat oil-immersion lens, and a 20x planfluorite. The latter objective has an adjustment collar that allows it to be utilized with oil, glycerin, or water as an immersion medium.
Specific microscope settings appropriate to the specimen should be set up away from the primary region of interest to avoid photobleaching of the fluorescent species in valuable regions of the specimen. Usually this requires setting the gain and black levels of the photomultiplier detectors together with the pinhole size to obtain the best balance between acceptable resolution and adequate contrast, using the lowest laser power possible to minimize photobleaching. Many instruments utilize color lookup tables designed to aid in setting the correct dynamic range for the image. Such tables are designed so that the darkest pixels, having brightness values around zero, are arbitrarily displayed as green (for example), and the brightest pixels, with brightness values near 255 in an 8-bit system, are displayed as red. The microscope parameters such as gain and black level, and the pinhole diameter, are adjusted so that there are only a few green and red pixels in the image, ensuring that the full dynamic range from 0 to 255 is utilized, but with little cutoff at either end of the brightness range. Although these adjustments can be made by eye, the use of pseudocolor at the extremes of the dynamic range of the imaging system makes the adjustment much less subjective. In some cases images must be collected at less than the full dynamic range of the system because less than optimum laser power must be used or the specimen has uneven fluorescence, causing a bright region to obscure a dimmer region that is of interest in the frame.
During scanning of the specimen an image averaging routine is usually employed to reduce random noise from the detection system and to enhance the constant (nonrandom) features in the image. An image equalization algorithm can be applied after collection of images to scale them to the full dynamic range of the display. Care should be taken not to apply this type of routine if measurements of fluorescence intensity are to be made unless a control image is included in the same frame as the experimental images before the equalization routine is applied. In using any type of image processing routine, it is a good strategy to save raw unprocessed images in addition to any processed ones.
The usual strategy in image collection is to save images onto the hard disk of the confocal system's computer, and later to back them up onto another mass storage device. In general it is always advisable to collect as many images as possible during a microscope session, and to discard unneeded ones during later review. Many images that seem unnecessary at first consideration become highly valuable at a later date after further review (especially with one's peers). If it seems wasteful to take seemingly superfluous images, consider that it much harder to prepare another specimen, and harder still to reproduce the exact conditions of an experiment or even to exactly duplicate the specimen preparation protocol.
A strategy for labeling image files in an informative way should be developed before imaging is begun. During imaging many notes should be taken or added to the image file along with the image if this capability is available on the system used. Tests should be done to ensure that any saved information is accessible after saving the images, keeping in mind that text and other information related to an image may be lost when it is subsequently transferred to image editing programs such as NIH Image or Adobe Photoshop on other computers. A well-organized notebook or laptop computer file may be preferable over other means of recording imaging session details, and should include filenames, comments, and details of the objective and any zoom factor used to allow calculating scale bars at a later date. Most confocal systems do not automatically record the objective lens used, and this information is important for calculating field widths and scale bars for later publication. Many modern systems utilize an image database that organizes file names and locations of the files, and that usually will display thumbnail files of the images. Care must be taken to follow image naming restrictions imposed by the system, such as the number of characters allowed in a file name and whether characters such as periods or spaces might be misinterpreted by the software.
In any experimental discipline, a protocol that has produced good results will sometimes inexplicably cease to work. When this occurs with confocal imaging experiments, there may be an initial reflex to blame the instrument rather than the specimen, but tests should be conducted to confirm that the specimen is not at fault before any troubleshooting is begun on the instrument. A good first test is to view the specimen on a conventional epifluorescence microscope. If some fluorescence is visible by eye, the signal should be very bright on a properly functioning confocal system. Having confirmed that fluorescence is present in the specimen, some tests of the confocal system should then be done using a known test specimen rather than the experimental one. For reference purposes, the confocal system should have a digital file of an image of the test specimen accessible to users of the microscope, including all parameters of its collection, such as laser power, pinhole diameter, objective lens and zoom value, and gain and black level of the detector.
If initial tests do not provide a clear solution, it is advisable to seek help from an expert who may have experienced the problem before. As a rule, if a user is not sure of something, it is best that they step back and ask for help before attempting any remedies. All of the companies that supply confocal microscopes have telephone help lines and websites that may be accessed for additional help.
Problems that are traced to the preparation protocols are usually caused by degradation of reagents, and this should be checked by performing a series of diagnostic tests. It is usually advisable for the person doing the experiments to make up their own reagents, or at least to obtain them from trusted co-workers. Antibodies should be allocated from frozen stock in small batches, then stored under refrigeration, and should not be reused unless absolutely necessary. Sometimes this is unavoidable with rare or expensive reagents, and often does not present a problem.
In experiments with multilabeled specimens, bleed-through from one channel into another can occur as the result of properties of the specimen itself, or due to problems with the microscope. Published reviews in the literature should be consulted for the details of causes of bleed-through and possible remedies. A good test of the instrument itself involves imaging of a test specimen with known bleed-through properties, using both multiple-label and single-label settings. Images of the test specimen should be stored along with records of all pertinent microscope settings, so that when problems do occur the test specimen can be re-imaged with the same instrument conditions and the images compared with the stored ones recorded when the instrument was known to be operating optimally.
Other tests that may be done when problems arise include a visual inspection of the color of the laser illumination and a check of the anode voltage of the laser. If, for example, the beam from a krypton/argon laser appears blue instead of white when scanning on a multiple-label setting then this may indicate that the red line is weak. In this case, the anode voltage will probably be too high, and can usually be reduced to an acceptable level by adjusting the mirrors of the laser. Such adjustments should be done by, or supervised by, the person responsible for maintaining the confocal system. If the voltage cannot be brought into an acceptable range, a replacement for the laser may be required.
Another problem that may be encountered is that antibody probes may have degraded or need to be repurified or otherwise cleaned. Specimens that have been prepared for some time may develop increased background fluorescence and bleed-through caused by the fluorochrome separating from the secondary antibody and diffusing into the surrounding tissue. If at all possible, imaging should be carried out on freshly prepared specimens. Sometimes changing the concentration and/or the distribution of the fluorochromes will help alleviate problems. As one example, fluorescein may bleed into the rhodamine channel, and can be switched so that rhodamine is on the stronger channel. The rationale for this is that the fluorescein excitation spectrum has a tail that overlaps with and is excited in the rhodamine wavelength range. In subsequent experiments, the concentration of the secondaries can be reduced.
Image Processing and Publication
Images acquired with the confocal microscope are usually saved as digital computer files in a format that allows them to be easily manipulated using the proprietary software provided as part of the confocal system. One of the most dramatically improved capabilities of current LSCMs is their display of confocal images. This is of great importance because the improvements in imaging using the confocal microscope are of little value if there is no means to effectively display the images or reproduce them as hard copy.
As recently as 5 or so years ago most laboratories were still using traditional photographic darkrooms and chemical processing of film and paper for their final hard copy of images. There was particular difficulty in reproducing color images, because they were usually printed by independent printers who often had little idea of what constituted correct color balance in a micrograph, and the cost of quality prints was high. To obtain hard copy of images now, the image files can be exported to a slide printer, a color laser printer, or to a dye sublimation printer for publication quality prints, with direct control of the image characteristics being maintained by the person who acquired the images. Photographs for prints or slides can be taken directly from the video monitor screen, and movie sequences can be published on the Web.
Most journals are now able to accept digital image files for publication, and this has resulted in a dramatic improvement in the quality of published images. The image quality achieved by the confocal imaging system can now be more faithfully reproduced in published articles. In some cases journals also make their articles available on CD ROM, which means that readers can have access to published images exactly as they appeared when collected on the confocal systems of those doing the research. Not surprisingly the technological advances in image acquisition, display, and publication are especially beneficial in the case of color images in that journals can now accurately reproduce images with their original resolution and color balance, and theoretically, at much lower cost to the author.
Stephen W. Paddock - Laboratory of Molecular Biology, Howard Hughes Medical Institute, University of Wisconsin, Madison, Wisconsin 53706.
Thomas J. Fellers and Michael W. Davidson - National High Magnetic Field Laboratory, 1800 East Paul Dirac Dr., The Florida State University, Tallahassee, Florida, 32310.