Portal de Periódicos da CAPES

The Production of ‘Cell Cortices’ for Light and Electron Microscopy

2000; Wiley; Volume: 1; Issue: 7 Linguagem: Inglês

10.1034/j.1600-0854.2000.010704.x

1600-0854

John Heuser,

Microtubule and mitosis dynamics

The inner surface of the cell membrane offers a wealth of structural information that is vital for understanding the behavior of the cell cortex in particular, and the whole cell in general. A quarter of a century ago, Mazia et al. [1], Clarke et al. [2], Vacquier [3], and their colleagues at Berkeley introduced a powerful method for viewing this surface by light and electron microscopy, which, in principal, remains unchanged to this day. They were the first to realize that polylysine (and other cationic polypeptides) adsorb strongly to various solid surfaces creating a high density of free cationic sites that strongly 'glue' cells down through electrostatic interactions with the negative charges that predominate on the cell surface. In their original report [1], these authors noted that cells thus attached to polylysine surfaces 'remain alive, but may flatten or spread themselves to a degree that is not normal but may be desirable for observational purposes.' More importantly, they introduced the technique that will be elaborated and brought up-to-date in the present 'Toolbox' with the simple statement [1]: 'One interesting application of the method is in the observation of the inner face of the cell surface. The cells are attached to the polylysine surface and subjected to shear in a medium in which the cytoplasm will disperse when the cell membrane is torn. The body of the cell is sheared away but the area of the cell surface which is glued by the polylysine remains attached, inner face up.' ( Fig. 1). Vacquier's original image of the cortex of a sea urchin egg attached to polylysine-coated glass and 'unroofed' by a squirt of buffer. This swept away the cytoplasm of the egg and left only the attached cortical granules that were attached to the inside of the plasma membrane and in position for exocytosis. Reprinted with permission from reference [3]. Soon after the introduction of this technique, Jacobson and colleagues adapted it for biochemistry in a highly useful manner, by developing a technique for coating glass beads with polylysine, attaching cells to the beads and then rupturing them, and thereby collecting sufficient quantities of plasma membranes for interesting biochemical analyses [4–9]. Unfortunately, this informative biochemical approach has fallen into relative disuse. In any case, the goal of the present report is to describe how the 'Mazia' technique for imaging the inner surface of the plasma membrane has evolved over the years [10–12], and to describe the procedures we currently consider to be optimal for attaching cells to polylysine-coated coverslips, for 'unroofing' them, and for imaging them in the electron microscope. Polylysine remains the 'glue' of choice in most circumstances; however, alternatives have come and gone over the years. Thus, Buechi and Bachi [13], virologists in Zurich wishing to see the inner aspect of the plasma membrane at sites of Sendai virus insertion and budding, described a procedure for cationizing glass with an amino-silane compound dissolved in acetone [see also [14]]. After covalently attaching this compound to the glass via its silane groups, these workers then exposed the glass to glutaraldehyde to convert the free amino groups into reactive aldehydes. This 'glued' cells to the glass covalently by the formation of aldehyde bonds, and was quite successful at creating a 'grip' that resisted cell rupture and subsequent preparations for electron microscopy (EM). Their amino-silane technique has been used occasionally since then [14] but has proved to be too tedious for routine use and, in our hands, offers no special advantages over the original polylysine technique. Nonetheless, when aldehyde fixation rather than electrostatic bonding of cells to glass is desirable, an adsorbed film of polylysine can also be reacted with glutaraldehyde without displacing it from the glass. This converts all the polylysine's free epsilon-amino groups to free aldehyde groups and renders the glass electroneutral, just like a glutaraldehyde-fixed amino-silane surface. In our experience, cells survive the attachment to such 'aldehyde monolayers' just as well as to poly-cationized surfaces, and sometimes survive and adhere even better! Until recently, this technique has generally been used with cell suspensions that are adsorbed acutely to the polylysine and 'unroofed' after just a minute or two of adhesion. Thus, it has been ideal for free-living cells like amoebae [2,15] or naturally suspended cells like erythrocytes [16] and leukocytes [10]. Likewise, it has also been used with tissue-cultured cells that have been acutely resuspended with trypsin-EDTA, or even with whole tissues that have been dissociated into individual cells with collagenases. In any case, it is important to realize that regardless of the source of cell-type used, the final cell suspension that is about to be glued down onto polylysine must always be washed carefully by repeated gentle centrifugation in protein-free saline. Furthermore, great care must be taken to avoid any cell rupture during these washes, or else cytoplasmic proteins will be liberated. The problem here is that any free protein outside the cells will adsorb to the polylysine-coated surface and promptly obliterate its stickiness, before the cells have had a chance to attach. Thus, this technique stands in stark contrast to the commonly used technique of pretreating glass or plastic with polylysine before tissue culturing. In this situation, full, serum-containing medium is always applied to the polylysine before or during the introduction of the cells and the high concentration of proteins in the medium – molecules such as fibrin, etc. – adhere to the polylysine and form a relatively 'natural' surface for the cells to grow on. In contrast, most tissue-cultured cells (and particularly very motile ones) do not even survive very long on 'bare' polylysine; the grip is just too tight and they tear themselves apart. The important point to stress here is the following: if adhesion of a cell suspension to polylysine fails, it is generally because [1] cell damage has occurred during the wash steps, [2] the cells are intrinsically too fragile, or [3] their exposure to the glass has gone on too long and they have begun to break open 'spontaneously'. We find that even after cells have been adsorbed to polylysine-treated glass, the release of proteins from a few broken ones will dislodge them all. The original approach of shearing open strongly adherant cells with a simple 'squirt' of buffer [1] has remained standard [11], but many variations have also been tried. Thus, the 'squirt' may come from a Pasteur pipette or from a needle-and-syringe; or as we prefer, it may come from the intense displacement of liquid caused by the standard type of 'horn'-type ultrasonic probe [17,18]. Our current protocol is to use one of the one-eighth inch diameter 'microprobes' originally designed to sonicate small samples in Eppendorf tubes, driven by a very gentle 10–25 W power supply running at only 10–20% of maximum power. This is dipped into ∼10 ml of 'breakage buffer' (basically, any K/Mg based isotonic buffer) in a 60 mm Petri dish, such that when power is applied, a jet of liquid and tiny cavatation-bubbles emerge straight down from the probe tip. A coverslip of polylysine-adherent cells is inserted into this dish of buffer at a 45° angle, positioned about 1/8th inch below the probe tip, and a momentary blast is delivered by a foot-control on the ultrasonic power supply. This is all done under strict visual control, with a bright 80× dissecting microscope viewing the sample from the side through a 45° mirror, since in that way it is easy to see the slight turbidity on the glass created by adsorbed cells and easy to see when they are partially 'blown away' ( Fig. 2). It is vital to stress that such partial cell removal is the desired goal, since only then can one be certain that some cells will be just barely unroofed and the maximum amount of cortical structure will remain on their inner surfaces! Direct visualization of this 'unroofing' step also shows when the jet of liquid has 'misfired' or has otherwise been inadequate for unroofing the cells in question, or when it has been too harsh and has totally denuded the coverslip. (In lieu of such a setup, one can simply 'unroof' several coverslips with different times or power settings of sonication and look for the best result. Cell cortices remain visible in standard tissue-culture type inverted light microscopes, albeit as nothing more than faint dark rings or 'smudges' on the glass where the cells used to be.) Anaglyph stereo view of our apparatus for 'unroofing' cells by exposing them to a gentle 'blast' from an ultrasonic probe immersed in a dish of buffer. For successful application of this technique, close visual monitoring must be established. This we accomplish by viewing the probe tip and sample through a high-magnification dissecting microscope projected through a mirror angled at 45°. [For discussion on the preparation of anaglyph stereo images see [41]]. The major alternative to 'unroofing' cells by a shearing fluid-flow has been to 'squash' them gently between two sticky coverslips [19,20] or between a sticky coverslip and an EM grid [21]. In our hands, this works quite well for cells that are so tough that they refuse to yield to a shearing force, or that adhere so weakly to polylysine that they are always completely detached without leaving behind their ventral surfaces. It is also useful for cells that are grown on non-adherent surfaces such as porous filters, as epithelial cells often are, especially when an EM view of the cell apex is desired [19–21]. By squashing cells between two coverslips or any two opposed surfaces, they can't get away! Additionally, their organelle 'guts' often are spilled out onto the opposed surfaces during the squash, and these are often informative, in and of themselves. Finally, such 'squashing' of course yields on the upper coverslip a sample consisting primarily of cell apices, whose structure can then be compared with the ventral surfaces normally obtained by the standard 'unroofing' procedure. The only problem with this 'sandwich' technique is that it generally yields much 'messier' preps on polylysine-coated glass surfaces. This is because it takes some time to separate the two coverslips after the squash, thus delaying the washing and fixation steps that are normally carried out immediately after 'unroofing'. Jumping ahead to the final visualization and interpretation steps of the techniques outlined here, we should note immediately that cells acutely attached to polylysine (or to other cationized surfaces or even to aldehyde-coated surfaces) in the absence of any exogenous proteins, display one huge artifact in their plasmalemmae! Their adhesion must be so tight that their plasma membranes become 'embossed' with an odd texture that is visible at high magnification in the EM. This we don't completely understand. Perhaps the texturing is a reflection of inhomogenities in the distribution of cations on the glass (or anions in the cell membrane), or perhaps it is actually proteins (maybe even the cells' own membrane proteins) trapped between the membrane and the glass. In any case, such 'embossing' is clearly an artifact, since it is entirely absent from regions of the plasma membrane that naturally curve inwards, such as nascent pinosomes or the bases of ruffles. Furthermore, it is entirely absent from cells that are 'unroofed'in the absence of polylysine, as will be described next. Perhaps the single most useful advance made over the original Mazia technique has been the realization that tissue-cultured cells need not be removed from their primary culture environments, resuspended and then acutely plated on polylysine: instead they can be grown directly on acid-cleaned glass and 'unroofed'while still on the original glass[22–25]. Often, the cells' grip to the glass is sufficient for at least some of their ventral surfaces to hang on throughout the 'unroofing' procedure. Importantly, these cells show none of the plasmalemmal embossing artifact described above. For 'unroofing' such long-term cultures, we have found that we can vastly increase the yield of properly exposed ventral cell bottoms by an elaboration of the original procedure. Specifically, we expose the cultures very briefly to a relatively low molecular weight polylysine, just long enough to allow it to diffuse under the outer edges of the cells and begin to glue their perimeters down onto the glass. This region again acquires the 'embossing' described above, and thus is ruined for high-resolution EM; nevertheless, it serves to hold down the more central regions of the cell cortex, which are not so embossed. To make this 'edge-gluing' trick work properly, the cell cultures must first be washed free of all protein, again by several rinses in warm PBS to remove their original culture medium. Only then can they be exposed to polylysine without fear of creating a precipitate of protein. To achieve the actual 'edge-gluing' effect, we generally use 0.1 mg/ml of polylysine dissolved in PBS, using the 30–60 kDa oligomers of polylysine from Sigma (St Louis, MO). Coverslips are swirled in this solution very briefly – for only 15 s – before being rinsed to arrest any further penetration of polylysine under them and to remove any excess polylysine that might contaminate their inner surfaces once they are 'unroofed'. Still, even this brief exposure to a relatively dilute polylysine solution creates a severe precipitate on the outsides of cells, which is readily visible on the cells that happen not to be 'unroofed' in the next step. One must hope that this precipitation doesn't have enough time to seriously alter cell physiology. One further aid to successful unroofing of cells that reside on their original glass coverslips is to swell them gently in the last moments before sonication. For this, we use a one-third osmotic strength solution. Finally, since the interior of the cell cortex is about to be exposed, it is only logical to immediately precede the 'unroofing' procedure with an appropriate change in salts from the usual Na+/Ca2+ base of the whole cells' external world, to the K+/Mg2+ base of their internal world. Putting all these considerations together, we unroof cells grown in long-term cultures on their own coverslips by the following protocol: 1) transfer them from full medium in the incubator to three washes in 37°C PBS over 15 min; 2) then pass them through 15 s in 0.1 mg/ml 'mid'-molecular weight polylysine in Ca-free PBS; 3) then immediately wash them in three changes over 30 s in one-third strength K+/Mg2+ buffer; and 4) finally transfer them to the sonication bath containing full-strength K+/Mg2+ buffer (±sucrose to stabilize internal organelles) and 'unroof' them by an approximately half second ultrasonic 'burst'. It is worth noting that 'unroofing' cells grown in culture of course yields only their original, ventral surfaces, whereas unroofing cells acutely plated on polylysine gives a random sampling of their ventral and apical surfaces (presuming that during their resuspension at 37°C, such regional differentiations were not lost). Having now successfully 'unroofed' cells by one means or another, the next question is how to prepare them for microscopy. We have found that their exposed ventral membranes will 'survive' in K/Mg buffer for ∼10 min before they become severely vesiculated. During this time, the actin filaments and microtubules that initially cling to them progressively depolarize, and the clathrin-coated pits on them round up, as if 'in preparation' for pinching off. Remarkably, however, their caveolae remain totally unperturbed and do not change their curvature. If desired, experiments like adding cytosol, etc., can be carried out on the 'unroofed' cells during these 10 min, or they can be briefly decorated with antibodies. More commonly, however, the aim will be to preserve the architecture of the cortex in as natural a state as possible, so chemical fixation or 'quick-freezing' will be mandated as soon after the 'unroofing' as possible. When fixatives are used here, the choice is straightforward: glutaraldehyde for the best structural preservation or formaldehyde for the best antigen preservation, but in either case, in the same K+/Mg2+ buffer as was used for the unroofing. Next comes the question of how to get such 'unroofed' cell preparations on polylysine-treated glass into the electron microscope. Again, Mazia et al. [1], Clarke et al. [2] and Vacquier [3], in their original studies, used standard OsO4 postfixation, ethanol dehydration, and critical-point drying before platinum-coating and imaging of their preparations with high-resolution SEM ( Fig. 1). So also did Nermut [16] and others in their extensive use of this technique [see also [5,15,17]] (c.f. Fig. 3). For decades we have advocated freeze-drying and TEM over critical-point drying and SEM, partly because it obviates the potentially damaging effects of osmification and dehydration, and partly because TEM is so much easier to do than high-resolution SEM [26–29]. Nermut's image of the inner surface of a cultured hepatocyte 'unroofed' by a squirt of buffer and then prepared for high-resolution TEM by fixation and critical-point drying. This clearly displayed the polygonal clathrin lattices that abound on this actively endocytic membrane surface. Reprinted with permission from reference [22]. In the end, the choice between critical-point drying and freeze-drying is largely a matter of personal preference, and becomes a decision based largely on whether the proper technology and expertise for decent freezing is locally available, or not. 'Quality' freeze-drying cannot be carried out if the freezing itself is not optimal; here we believe that only our liquid helium-cooled 'Cryopress' achieves truly optimal freezing [30]. Earlier studies of plasmalemmal lawns suffered technically exactly at this point. Buechi and Bachi [13,31], for example, froze their preparations by plunging them into liquid nitrogen – a clearly inadequate approach. Ohno achieved somewhat better freezing by plunging his preparations into a mixture of pentane and propane cooled with liquid nitrogen [14]. Following our earlier advice that methanol acts as a cryoprotectant, which will evaporate during freeze-drying [27], he also took the further step of washing his preps in 10% methanol before freezing. In any case, no cryo-'plunging' technique achieves the rates of cooling and minimizes ice crystal formation as much as does 'slam-freezing' against an ultra-cold metal block; that is why we advocate our original freezing procedures [30]. The Cryopress cushions the sample in such a way that controlled 'slamming' of glass coverslips onto ultracold metal does not risk breaking the glass. Quality freeze-drying also depends on the procedures followed after freezing. In general, this freeze-drying is done in a vacuum-evaporator, so that a metal replica can be applied immediately thereafter. But this requires transfer of the frozen sample onto a 'cryo-stage' inside the vacuum-evaporator where it can be kept frozen, and very few investigators have worried about the problems inherent in this transfer: namely, the transient warming of the sample and the accumulation of condensed atmospheric moisture or 'hoarfrost' on it. We avoid these 'perils' by capping the sample while it is still immersed in liquid nitrogen, and by not removing its cap until transfer into the evaporator is complete and the sample is firmly stabilized at −150° within a fully-evacuated chamber. Furthermore, few practitioners of freeze-drying have worried about the hazards of leaving samples too long in the vacuum after freeze-drying. Buechi and Bachi [13] left their samples to freeze-dry for 90 min at −40° and Ohno [14] freeze-dried for 60 min at −90°C, while we find that thin films of water on glass coverslips are completely freeze-dried in only 15 min at −100°C [18]. So we get the freeze-drying over with fast, and we watch its completion with a high-power dissecting microscope mounted in the evaporator; and, just as soon as possible, we create our platinum replica of the glass surface. All of these precautions serve to sustain the maximal surface architecture of our 'unroofed' cells and explain why they display such dramatic 3-D topology ( Fig. 4). Anaglyph stereo-view of the inner surface of a PC12 neuroendocrine cell. It was exposed by the 'unroofing' procedure described herein, displaying a number of chromaffin granules (secretory vesicles attached to the plasma membrane by long 'threads' and/or by actin filaments in the cortical cytoskeleton). We should add here that a 'shortcut' used by many labs to get such 'unroofed cell' preparations – or 'plasmalemmal lawns' as they are often called [32,33]– into the electron microscope without any replication at all is simply to briefly stain and air-dry them. This has been carried out most successfully in a number of immunocytochemical studies, where gold-labeled secondary antibodies showed up very nicely due to the thinness of the sample and the faintness of its staining [32–35]. Unfortunately, the quality of structural preservation is rather compromised by the air-drying used in these studies, so proper identification of structural details is often rendered somewhat ambiguous and nothing remains of the cells' natural 3-D topology. We strongly advocate such use of EM immunocytochemistry on 'unroofed' cells, particularly because such preparations offer perfect exposure to antibodies without the need for any detergent or formaldehyde fixation. However, we would suggest that it is perfectly possible to combine such EM immunocytochemistry with proper freeze-drying and platinum replication of 'unroofed' cells, and thus preserve their surface architecture as well. The gold particles that mark the secondary antibodies are readily visible underneath a platinum replica. This is equally true when indirect EM immunocytochemistry is performed on the residue of proteins that cling to the undersides of platinum replicas made from whole tissues [36]. The only problem with EM immunocytochemistry of 'unroofed' cells on polylysine-coated glass is that the gold particles tend to get 'eaten away' when the platinum replica is floated off the glass by angling it into a dish of hydrofluoric acid, as is standardly done. To overcome this problem, we have introduced one final 'wrinkle' in the original technique of Mazia et al. [1]. Specifically, we pre-coat the glass coverslips on which we grow cells (or to which we acutely glue-down suspended cells) with a thin layer of carbon. Having done so, the final platinum replica separates in hydrofluoric acid along the carbon/glass interface. Hence, the gold antibodies – now 'sandwiched' between the platinum replica above and the carbon beneath – remain totally unscathed ( Fig. 5). In fact, the entire ventral cortex of the 'unroofed' cells remains within this 'sandwich' as well. Anaglyph stereo-view of the inner surface of an 'unroofed' CHO cell subjected to indirect EM immunocytochemistry with anti-caveolin antibodies and 15 nm gold-tagged secondaries. Dome-shaped caveolae are clearly and specifically labeled with gold (yellow dots), in contrast to the unlabelled actin filaments and the one clathrin-coated pit in the field. Were this cortex any thicker, it might obscure detail or add confusing contrast to the platinum replica; however, it does not. Instead, this 'trapping' of the actual biological sample offers certain advantages. It allows other sorts of gold tracers, such as endosomal markers, to be introduced either before or after sonication and, more importantly, it also allows oxidized DAB to be visualized within an actual biological sample as well ( Fig. 6). Thus, HRP extracellular-tracers and HRP-labeled antibodies can be visualized in such freeze-dried and platinum-replicated samples, just as well as gold. Indeed, oxidized DAB created by photo-conversion of light microscopic fluorophores can also be trapped and visualized in such carbon-undercoated samples, thereby creating a wonderful link between 'deep-etch' EM and the world of light microscopy. HRP-labeled endosomes from a Dictyostelium amoeba 'squashed' on polylysine-coated glass. (Note the abundance of released ribosomes stuck to the glass in the background.) The DAB product of the HRP reaction looks white in such contrast-reversed anaglyph stereos. In the world of light microscopy as well as that of TEM and SEM, 'unroofed' cell cortices can also be highly informative, especially when imaged by high-contrast techniques like interference-reflection microscopy [15,24,37] or video-enhanced differential interference microscopy [39]. Moreover, they are intrinsically so thin that excellent fluorescence microscopy can be done on them, without the need for any form of 'confocal' imaging [38]. Finally, it is worth remembering that such substrate-attached cell cortices are not only great for all sorts of microscopy, but are also excellent preparations for a variety of in vitro reconstitution assays, providing ideal substrates for studying dynamic membrane processes such as actin polymerization, exocytosis [39], and endocytosis [20,32–35,40].