|Home | About | Journals | Submit | Contact Us | Français|
Adrenergic receptors (ARs) belong to a superfamily of the G-proteincoupled receptors and are categorized by their binding to endogenously occurring catecholamines, i.e., norepinephrine and epinephrine. Adrenergic receptors are classified into three groups (α1-, α2, and β-ARs), each of which is further divided into three subtypes. The α1-(α1A-, α1B-, and α1D-ARs) couple with Gq family of G-proteins (G11, G14, G15, and G16) and result in activation of phospholipase C-βs that liberate two second messengers, diacylglycerol and inositol-l,4,5-trisphosphate. The three subtypes of α2ARs are designated α2A-, α2B-, and α2CcAR. On binding with agonists, α2-AR inhibit adenylyl cyclase and calcium channels, but activate potassium channels through coupling to the Gi family of G-proteins (Gi1, Gi2, Gi3, and G0). Finally, the three groups of β-AR are designated β1-, β2-, and β3-AR: these increase the intracellular cAMP content by activating Gs, which is coupled to the enzyme, adenylyl cyclase (1).
Functions of the adrenergic receptors in vitro and in vivo have been analyzed mostly by administrating subtype-selective agonists or antagonists (2). However, although ligands specific for the three major adrenergic receptor types are available and have yielded much useful information, most of the ligands currently available do not exhibit sufficient specificity for discriminating among the subtypes (e.g., the A-subtype of α2-AR). Thus, an alternative approch for identifying the function of the subtypes has been to knock out the gene encoding for the particular receptor subtype. This approach has not always been met with success either, probably because other subtypes of catecholaminergic receptors compensate for the knockout. This is one reason many of the advances in our knowledge about the catecholaminergic receptor subtypes are derived from immunocytochemistry. For brain research, in particular, the immunocytochemical approach has been useful. This is because brain function depends critically on the connectivity formed among neurons. Thus, the effect of catecholamines within brain could differ greatly depending on the site of action of the neurotransmitter and on the receptor subtype located near the release of the neurotransmitter. The site of action of catecholamines can vary by region (e.g., visual vs auditory vs multimodal pathways), cell type within the region (e.g., neurons using excitatory transmitter for projecting long distances vs those using an inhibitory transmitter for local circuits or nonneuronal cells, such as astrocytes), and by the subcellular compartment (dendritic shafts, where primarily inhibitory inputs from other neurons are received, vs dendritic spines, where primarily excitatory inputs are received, or in axons, where outputs to other neurons are propagated and transmitted via release of neurotransmitters). For example, our study using an antiserum capable of selectively recognizing the A subtype of α2-ARs revealed that these occur presynaptically (Fig. 1A, B), some of which were positively identified as noradrenergic axon terminals (3). This result was expected, since earlier physiological studies had shown that α2-ARs operate as autoreceptors, inhibiting release of norepinephrine or epinephrine (4). However, we also observed that these receptors occur in noncatecholaminergic axon terminals, indicating that these may also operate as heteroreceptors regulating the release of transmitters other than norepinephrine and epinephrine. Furthermore, this receptor has been observed postsynaptically within the cerebral cortex (5,6; Fig. 1) and the hippocampus (7), even though electrophysiological studies have indicated a lack of α2-AR mediated postsynaptic effects in these forebrain structures (8). Differences in findings such as these indicate that α2-AR in these structures, unlike those in the brainstem, may activate intracellular second messenger cascades without activating potassium channels.
Similarly, a series of studies using antisera directed against distinct domains of β-ARs have revealed interesting differences in the receptor’s conformation across developmental states and cell types within intact cerebral cortical tissue. The first polyclonal antiserum that became available for ultrastructural studies was raised by Joh, using the antigen harvested from frog erythrocyte membranes by Strader and colleagues (9). This antiserum yielded immunolabeling of various portions of neurons, including perikarya and axons, but primarily distal dendrites. Astrocytic processes also were immunolabeled using this antiserum (9,10). In sharp contrast to this result, it was observed that polyclonal antisera and monoclonal antibodies (MAbs) directed against the third intracellular loop region yielded immunolabeling primarily of perikaryal regions of neurons (11), although distal dendrites, spines (11), and axons including presynaptic portions of axons (12), also were immunoreactive. Finally, another polyclonal antiserum directed against the C-terminus of β-ARs recognized primarily astrocytic processes in adulthood (13-16; Fig. 2) but also immunolabeled the earliest-formed synapses within neonatal cortices (17; Fig. 3). Each of these immunolabeling patterns was confirmed to be specific by showing abolishment of antigenicity following preadsorption of the primary antibodies (see Notes 1 and 2). Future studies that examine the relationship of β-ARs to the molecules known to interact with them, such as β-arrestin, β-AR kinase, and Gs proteins, under physiologically specified conditions promise to provide detailed knowledge required for understanding the dynamic regulation of cell physiology by epinephrine and norepinephrine.
Immunocytochemistry has also been useful for studying the subcellular compartmentation of different receptor subtypes that are coexpressed within single cells. For example, adipocytes and heart cells express both the βl- and β3-AR-subtypes of β-ARs. Interestingly, however, such coexpression does not allow for functional compensation, even though both subtypes are coupled to Gs and adenylyl cyclase (18,19). It is possible that differential localization of the receptors within each cell, i.e., compartmentation, influences the response, since signaling molecules and second messengers are not expected to diffuse freely within cells. In another example, Jurevicus and Fischmeister (20) reported the functional compartment of the β1-AR-mediated cAMP accumulation that is important for increase of calcium current through L-type calcium channel in heart, indicating that the β1-AR, but not the β3-AR, was closely associated with the effector molecule. In another example, it has been reported that the α2C- and α1A-ARs mainly localize intracellularly (21,22). The compartmentation can facilitate efficient signal transduction from the receptor to cellular response and avoid unfavorable responses. In cells as polarized as neurons and glia, precise knowledge about receptor localization becomes ever more important, as diffusion of second messengers become restricted to single dendritic spines, dendritic shafts, or axon varicosities.
The cellular mechanism by which receptors, G-proteins, and effector molecules become properly localized is yet unknown. Clearly, elucidation of such cellular mechanism requires precise knowledge about the subcellular localization of the receptor and related elements, for which specific antibodies that recognize the receptors are required.
Much about the signal transduction mechanism has been learned also by expressing the receptors in physiologically “irrelevant” cell lines, i.e., those that are derived from cell types that, when within intact tissue, are devoid of the particular receptor. However, the level of expression of the exogenously transfected receptors tends to be high (above pmol/mg protein) compared to the level for endogenously expressed receptors (typically 10-200 fmol/mg protein). Although this difference serves as an advantage for measuring signal one still needs to be prudent about checking the behavior of the, receptors in native and intact cells, where fine-tuning of receptor-mediated effects could depend critically on the exact location of the receptors in relation to other modulating and competing biochemical pathways.
Antibodies that are capable of recognizing the posttranslationally modified receptors promise to be powerful tools for analyzing the physiological conditions and consequences of posttranslational modification. The types of posttranslational modification include phosphorylation of specific intracellular domains in association with receptor desensitization, glycosylation of extracellular domains, and the addition of palmitoyl and myristoyl groups during intracellular trafficking.
Electron microscopic immunocytochemistry (EM-ICC) is very useful for determining existence and coexistence within fine processes of particular receptors,receptor subunits, neurotransmitters, or enzymes involved in generation of second messengers. The visualization of fine organelles is especially useful for distinguishing fine processes as astrocytic, axonal, or spinous, many of which often are less than a micrometer in diameter. Most importantly, electron microscopy (EM) is essential for identifying the morphological characteristics of synaptic junctions. Conversely, EM has been useful for identifying the presence at catecholaminergic receptors at sites lacking morphological characteristics of synaptic junctions, thereby providing strong support for the idea that catecholaminergic neuromodulation can occur by volume transmission, in addition to the more conventional transmission whereby transmitters released from axon terminals remain within the junctional cleft (23). For this reason, the methods chosen for EM-ICC aim for optimal tissue preservation while also avoiding loss of antigenicity brought about by excessive tissue preservation.
In this chapter, we will describe various techniques available for visualizing antibody–antigen complexes for EM-ICC, as well as their advantages and limitations. This discussion will be limited to immunolabeling intact brain tissue for which the authors have direct experience and have obtained useful results. Problems encountered with producing polyclonal antisera and possible solutions to these problems also are included under Subheading 4. (see Notes 1–7).
The sources listed here have been used by the authors and shown to yield useful data. On the other hand, other sources are also likely to provide reagents of sufficient purity or specificity.
Brain tissue can be obtained from a variety of animals. In our hands, vertebrates ranging from amphibians (e.g., frogs) to nonhuman primates have yielded useful data when using polyclonal antisera. Ultrastructural analysis is facilitated when tissue is preserved by rapid transcardial perfusion of fixatives, as detailed under Subheading 3.
All fixatives used for EM are volatile and highly reactive with tissue. Thus, these materials and particularly the solutions must be handled under a wellworking hood. Some fixatives may also need to be collected after use. One should check with the local administrator regarding hazardous waste disposal.
The following sources have been used and yielded good preservation of the ultrastructure:
Heavy metals, osmium tetroxide, and Lowicryl can be purchased from EM Sciences (Fort Washington, PA), EM Corp., or from Ted Pella (Redding, CA).
For optimal preservation of cellular morphology and of the distribution of molecules within cells, transcardial perfusion with fixatives is required. However, there sometimes are needs to analyze the ultrastructure of tissue that has not undergone transcardial perfusion. One example is the need to analyze the ultrastructure of biopsy samples or blocks of tissue obtained postmortem. Even brains. that have undergone transcardial perfusion with fixative may need to be postflxed by immersiOn for further improvement of structure. Under such Circumstances, tissue may be fixed by immersion. Since penetration of fixatives through tissue is a slow process, relative to the rate of ultrastructural deterioration owing to anoxia, immersion fixation necessarily results in suboptimal conditions for ultrastructural analysis, particularly within portions of tissue removed from tissue surface. On the other hand, the surface-most portions of such tissue may be usable, since fixatives reach these portions with minimal delay. It is not advisable to use tissue that has undergone freezing prior to fixation: such tissue exhibits gross destruction of membranes, resulting from expansion of water during ice formatiOn. Even when freezing follows fixation, destruction of the plasma membrane is not entirely avoidable during the freezethaw process. The problem with damaged plasma membranes is that identification of the boundaries of individual cellular processes by EM becomes difficult. This limitation, in turn, prevents analysis of the subcellular distribution of antigens.
Paraformaldehyde, used most widely for light microscopy, and acetone, used more for cultured cells, are not sufficient for ultrastructural preservation, since these fixatives do not preserve membranes of intracellular organelles or of the plasma membrane adequately for analysis. The most widely used fixative for EM is glutaraldehyde. This aldehyde has been used by electron microscopists at concentrations ranging from 0.05 up to 5%. Although the preservation of the ultrastructure is improved with increasing concentrations of glutaraldehyde, concentrations >0.1 % have led to marked reduction of immunoreactivity for catecholaminergic receptors (unpublished observations). On the other hand, a brief (<7 min) exposure of tissue to another highly reactive, small aldehyde, i.e., acrolein (24), at concentrations ranging from 3.0 to 3.75% has permitted good preservation of the ultrastructure as well antigenicity of catecholaminergic receptors (3,5-7,10-17, Figs.Figs.11–3). Thus, authors of this chapter and others using antisera directed against catecholaminergic receptors have often used the following combination of fixatives: 0.05 or 0.1 % glutaraldehyde in combination with 4% paraformaldehyde or 3–3.75% acrolein in combination with 2–4% paraformaldehyde.
Transcardial perfusion with fixatives is one of the most critical steps for successful ultrastructural preservation of tissue. The aim of transcardial perfusion is to achieve ultrastructural preservation before morphological (and presumably chemical) alterations of tissue are triggered by anoxia. In order to minimize artifactual alterations of tissue, anoxia may be minimized by maintaining artificial ventilation during transcardial perfusion. The other key to success is speed, i.e., minimizing the number of seconds that lapses from the onset of anoxia (which begins the moment the diaphragm is cut for gaining access to the heart) up to tissue fixation (i.e., which must be preceded by steps whereby fixatives diffuse out of the blood vessel lumen and into the surrounding neuropil). A number of factors determine the speed. These include the rate of diffusion of the fixative, the efficiency with which one gains entry to the heart by dissection, and the rate of flow of the fixative through the cardiovascular system. For maximizing the rate of diffusion of the fixative within tissue, we recommend the use of small, highly reactive aldehydes, such as acrolein. Regarding swift entry into heart, one simply needs to practice the dissection procedure to gain expertise. The rate of flow of the fixative is best controlled using a peristaltic pump. This assures that the rate is maintained at a high level, but not overly high to cause rupture of blood vessels. For adult brains of most mammals, a flow rate setting of 70 mL/min is recommended. In order to avoid blockage of blood vessels by coagulated and aldehyde-fixed blood cells, one should flush the cardiovascular heart with heparinized saline (100–1000 U/mL of Heparin, added to 0.9% NaCI) prior to perfusion with fixatives. This saline flush, however, should be kept to a minimum in order to minimize delay of perfusion with fixatives.
For preparation of tissue for the immunocytochemical detechon of catecholamine receptors, we and others have found the following aldehydes to be suitable, both for retention of antigenicity and ultrastructural preservation: a mixture of 3% acrolein and 4% paraformaldehyde, buffered with 0.1 M phosphate buffer (pH 7.4), perfused over a period of 3–7 min, followed by perfusion with 4% paraformaldehyde in phosphate buffer without acrolein (3,5–7,10–17). Alternatively, a mixture of 0.1 % glutaraldehyde with 4% paraformaldehyde perfused over a period of 30 min has been useful (10).
Specifically, the following steps are recommended for transcardial perfusion:
By far the most favorable sectioning procedure for EM-ICC detection of antigens is to use a vibratome. This procedure avoids freeze-thawing of tissue, which can, in turn, cause morphological damage owing to formation of large ice crystals. Vibratomes can readily generate sections as thin as 30 μm from moderately fixed brains. The stronger the fixation, the thinner the sections can be Conversely, weakly fixed tissue, such as early postnatal tissue or those fixed with low concentrations or minimal volume of fixatives (e.g., 2% paraformaldehyde), need to be sectioned at greater thicknesses, e.g., 100 μm for postnatal d 3 rat brain sections, with greater vibration amplitude and with slower blade strokes.
Alternative choices for tissue sectioning include using the freezing microtome or a cryostat. However, these alternatives are less desirable because of unavoidable tissue damage caused by the freeze-thaw steps, even when precaution is taken to cryoprotect the tissue. When sectioning in a frozen state is unavoidable, one must take every precaution to avoid formation of large ice crystals that damage membranes: this is best managed by immersing the smallest possible block in a cryoprotectant, such as a mixture of sucrose (25 %) and glycerol (10%), buffered with 0.05 M phosphate buffer, and then freezing rapidly using Freon or isobutyl alcohol chilled to a temperature colder than −70°C by using liquid nitrogen. Further details of cryoultramicrotomy can be found in manuscripts by Tokuyasu et ai., Liou et al., and Sitte (25,26) since discussion of this technique is beyond the scope of this chapter.
For obtaining specific immunolabeling, it is desirable to control the termination as well as the initiation of fixation. Tissue fixed using highly reactive aldehydes, such as glutaraldehyde and acrolein, continue to form covalent bonds with primary amine groups of proteins, even after tissue has been sectioned and all excess aldehydes have been removed by rinsing. In order to terminate the aldehydes’ crosslinking actions, one needs to treat sections with reducing agents, such as sodium borohydride (27), that render the aldehyde groups nonreactive by converting them to alcohol groups or by treating with excess of primary amine groups. Acrolein- and glutaraldehyde-fixed sections of about 40-μm thickness require immersion for 30 min in a solution of 1 % sodium borohydride, buffered with 0.1 M phosphate buffer. This solution must be made immediately prior to use. Following the 30-min incubation period, sections should be rinsed in 0.1 M phosphate buffer until bubbles cease to emerge.
Tissue permeabilization is a step taken to increase penetration of immunoreagents, particularly antibodies, into tissue. For antigenic sites embedded within organelles, such as within vesicles, this step appears to be essential. For antigens that are soluble, such as those in the cytosol and for intracellular domains of membranous proteins, including the adrenergic receptors, permeabilization may be kept to a minimum. For EM-ICC, the permeabilization methods involving extraction of lipids from the plasma membrane, such as incubation in nonionic detergents (e.g., Triton X-lOO), interferes with ultrastructural analysis. Thus, detergent-treatment should be avoided. In cases where tissue penetration is required, methods compatible with EM include the following three:
Sections can be stored at 4–6°C for several months with minimal loss of ultrastructural details or antigenicity. The recommended storage buffer is PBS-azide.
The synaptic molecules are most readily detected using the enzymatically amplified method, i.e., the avidin–biotin–horseradish peroxidase complex, with DAB as substrate (HRP-DAB). Our previous experience with this label indicates that when used judiciously (i.e., with minimal peroxidase reaction period), HRP-DAB provides subcellular localization of antigens precise enoug to differentiate labeled from unlabeled portions of dendrites (see Figs. 1C1C and and3A):3A): For. example, we (3,5,6,10–17) and others (7,32) have used this label to distinguish immunolabeled from unlabeled synaptic membranesthat are positioned immediately adjacent to one another.
The advantage of using HRP-DAB is that this label is compatible with conventional resins and strong fixative for membranes, such as osmium tetroxide. These reagents provide excellent preservation of tissue, and this factor is useful, although not absolutely necessary, for identifying various types of synapses, such as nascent synapses within developing tissue and symmetric synaptic Junctions. We have noted that catecholaminergic synaptic junctions, identified by the presence of catecholaminergic receptors, differ from glutamatergic synapses in that the former frequently lack the conventional morphological features of synapses. This conclusion could not have been made if the analyzed specimens were fraught with difficulty caused by suboptimal preservation of the ultrastructure, such as the loss (rather than absence) of conventional morphological features of synapses.
HRP-DAB is compatible with pre-embedding silver-intensified gold immunolabeling (Fig. 2, detailed below), thereby allowing for identification of two antigens within single fine processes. Moreover, HRP-DAB allows for light microscopic inspection prior to EM processing, thereby allowing for assessment about the specificity of immunolabeling by comparing with, expected (previously reported) distribution patterns of immunoreactivity across cell types and brain regions.
However, the HRP-DAB label is not free of limitations. Although the enzymatic amplification afforded by this method allows for excellent detection of antigens, diffusion of HRP-DAB labels precludes identification of antigenic sites as membranous vs cytosolic, nor are quantitative measurements such as the concentration of antigens within single immunolabeled profiles, possible. For questions requiring this level of resolution and quantification, the pre- and postembedding gold methods, respectively, are recommended (see below). Additionally, the HRP-DAB label, when weak, is difficult to distinguish from the conventional counterstain, lead citrate. Thus, one may wish to omit the lead citrate counterstain to optimize detection of HRP-DAB labels.
An alternative method with which receptors can be labeled is the SIG method (see Note 8). Penetration of the secondary antibody is optimized by using small sizes of conjugated colloidal gold. One-nanometer colloidal gold particles are commercially available. In our experience, this smallest size colloidal gold is necessary for detection of antigens within fine processes, such as dendritic spines and axons. Since 1 nm is below the limit of resolution of electron microscopes, these colloidal gold particles will need to be silver-intensified for EM as well as light microscopic detection of immunolabels. For detection of antigens in larger profiles, such as cell bodies and dendrites, colloidal gold of larger sizes, such as the 5- and 10-nm variety, can be used without silver intensification for EM. However, these larger sizes of colloidal gold will still require silver intensification for light microscopy.
These colloidal gold labels will offer greater subcellular localization than HRP-DAB for questions, such as the membranous vs cytosolic localization, since the label is not diffusible (Fig. 4). As with HRP-DAB, sections immunolabeled with SIG can be examined by both light and EM, thus allowing for assessment of specific immunolabeling based on the cellular and areal distribution pattern. Moreover, the SIG label is compatible with osmium fixation of membranes and thus can be combined with HRP-DAB (for which osmium treatment is required to render the DAB reaction product electron-dense ).
The shortcomings of the SIG method are that the labels are not enzymatically amplified. For this reason, the procedure using SIG label is less sensitive than those using HRP-DAB labels, judging from the antibody concentration required to attain equivalent immunolabeling (estimated to be one-tenth). Others have also noted that immunogold labeling rarely occurs directly over postsynaptic densities, even for the presumed synaptic molecules, such as receptors. Instead, immunolabeling tends to occur at the edges of synaptic specializations (33) owing, possibly, to steric hindrance caused by colloidal gold particles even for cases where the smallest available size (1 nm) is used. Nevertheless, the pre-embedding SIG procedure continues to be an excellent label for combining with HRP-DAB and for discriminating localization of antigens to cytosol or over membranes.
Finally, the PEG procedure would be useful for questions requiring the most precise localization of antigens, such as the potential coexistence of two molecules within single PSDs or of their coexistence along the same patch of nonsynaptic plasma membrane or cytoplasmically. For example, it has been demonstrated that β-ARs regulate N-methyl-d-aspartate (NMDA) receptors, suggesting that the two receptors coexist with single dendrites, thereby allowing for their interaction following near-synchronous depolarization of noradrenergic and glutamatergic fibers (34). The PEG method, combined with a pre-embed label, could readily determine whether the two receptors, indeed, coexist within single fine processes (Fig. 5). Yet another useful application of PEG is for comparing the concentration of antigens across PSDs of different synapse types (e.g., noradrenergic synapses formed on pyramidal neurons vs those formed on inhibitory interneurons).
Successful dual localization of the two receptor subunits to single PSDs and at sites away from synapses has been achieved by using two sizes of colloidal gold for PEG labels and/or by combining PEGs with SIG (35). These studies indicate that PEG is applicable for the precise localization of neurotransmItter receptors. Moreover, the PEG procedure can, in some cases, be applied to osmium-fixed tissue for combining with HRP-DAB (HRP-DAB requires osmium treatment for rendering the label electron-dense for EM) (Erisir et a1., unpublished observations). Results from these studies indicate that not only double, but also triple, EM-ICC is a possibility by combining SIG, HRP-DAB, and one or two PEG labels.
Should the molecule of interest not withstand the pre-embed osmium fixation, then PEG will need to be performed on tissue in which the osmium fixation of membranes is substituted by a protocol using tannic acid in combination with uranyl acetate and iridium tetrabromide (35–39). The preservation of membranes is not as complete as with osmium tetroxide but is still useful for yielding information regarding the subcellular distribution of receptors. This Point is evident by comparing Figs. 1–3, which used osmium tetroxide, with Figs. 44 and and5,5, which were preserved Without the use of osmium tetroxide.
Should the antigen of interest not withstand the heat needed to polymerize conventional resins (60°C for Epon and Epon-Spurr), an alternative resin, Lowicryl, can be used, since this resin can be polymerized at temperatures below −40°c. In recent years, results obtained using this resin have revealed exquisite, highly localized distribution of glutamate receptors (38,39). Tissue to be embedded using Lowicryl cannot be fixed by osmium, since osmium interferes with UV irradiation required for Lowicryl polymerization.
In short, by combining multiple EM-ICC techniques, one can maximally analyze the cellular and molecular details of synapses while also compensating for the well-known short-comings of each method.
As noted above, the most sensitive method currently available for EM-ICC uses DAB as substrate for HRP, which, in turn, is attached to antibody–antigen complexes via avidin–biotin links (ABC). The alternative peroxidase-based EM-ICC procedure, peroxidase-antiperoxidase (PAP), has been described in detail and is also applicable for the detection of adrenergic receptors (10). However, this procedure will not be described here, since it is less sensitive than the ABC-DAB procedure.
The development of HRP-DAB immunolabels by the ABC method involves linking of biotinylated secondary antibodies (antirabbit IgG) to biotinylated HRP, usig. the four binding sites on avidin as bridges. Specifically, the following procedure is recommended, as described by the manufacturer (Vector Labs):
Secondary antibodies are available conjugated to varying sizes of colloidal gold. For our purposes, which were to visualize immunoreactivity within fine processes (i.e., <1 μm in diameter), we have opted to use col1oidal gold of 1–1.4 nm in diameter. Although this size of colloidal gold requires silver intensification for EM visualization, the extra steps required for silver intensification are well worth the trouble, since these allow for detection of antigens within dendritic spines as well as axons. In contrast, secondary antibodies conjugated to larger sizes (>5 nm) of colloidal gold do not require silver intensification for EM visualization, but preclude analysis of small profiles, such as axons, spines, or astrocytic processes.
Tissue to be used for silver-intensified colloidal gold labeling should have the endogenous zinc chelated, since zinc, together with colloidal gold, becomes silver-intensified, yielding particles indistinguishable from silver-intensified colloidal gold particles (40). This is achieved by injecting sodium diethyl. dithiocarbamate, ip (1 g/kg), 15 min prior to transcardial perfusion of the animal.
The protocol that has yielded useful data is as follows, which applies osmium tetroxide as the fixative of membranes. Alternatively, sections can be processed for EM without treatment with osmium tetroxide, in order to avoid any loss of silver-intensified gold particles. Procedure for the osmium-free treatment of sections for EM is outlined in Subheading 3.6., step 2.
PEG is achieved by applying primary and secondary antisera directly on ultrathin sections prepared from resin-embedded tissue. Thus, considerations must be made for retaining antigenicity of the molecules throughout the procedure required for embedding sections in resins (which typically involve dehydration and irradiation with heat [ca. 60°C] or UV for a few days). Furthermore, one must expect further loss of antigenicity during the steps taken to incubate ultrathin sections in buffers for PEG labeling. For this reason, the PEG procedure should be performed soon after preparing ultrathin sectioning. Moreover, it is advantageous to use the strongest fixative possible for transcardial perfusion, such as a high concentration of glutaraldehyde, in order to minimize leaching of antigens out of ultrathin sections. Of course, the choice of fixatives is likely to be constrained further by the potential loss of antigenicity owing to denaturation of the molecule caused by strong fixations.
A few choices for embedding resins are available. For studies involving epitopes of molecules or antigens that are not heat-sensitive, EMBED 812 (Epon) or Epon-Spurr would be the resin of choice. Of the two, Epon-Spurr is more hydrophilic, thereby allowing greater penetration of ultrathin sections by immunoreagents. Should the antigen be heat-sensitive, then one will need to resort to using embedding resins that do not require heat for polymerization. Lowicryl is one such resin, for it can be polymerized at freezing temperatures by UV irradiation. One additional advantage of Lowicryl is that it is a hydrophilic resin, allowing for excellent penetration of immunoreagents through the thickness of ultrathin sections. However, UV radiation also may cause loss of antigenicity.
In general, all solutions listed below, except antisera, should be filtered using O.22-μm Millipore filters. Small-size filters that fit on the tip of syringes are useful for this purpose. Grids are incubated by submerging them, face-up, in droplets. It is best to submerge grids by sliding them sideways into droplets, formed on the surface of parafilm or silicone mats for grids. The specific steps are as follows, based on a procedure optimized by Phend et al. (37).
The procedure described below involves postfixation, flat-embedding, capsule-embedding, and then preparation of ultrathin sections. Two procedures are outlined: one for HRP-DAB and SIG (steps 1,3-7), and another—osmium free (2–7)—that may need to be followed for SIG and PEG immunolabels, depending on the susceptibility of the antigen to denaturation caused by chemical and heat treatments. The osmium-free method is as outlined previously by Phend et al. (37).