(Extract from an email from Thurston Lacalli to an enquiring Japanese student)
Almost without exception, my larval EM work over the past 25 years has been done using a fixation method known as semisimultaneous fixation. This is the best, and perhaps only method that deals adequately with delicate larvae, many of which are filled with mucus cells that burst with conventional treatments. Conventional fixation typically involves treating with glutaraldehyde for a prolonged period before adding osmium. I decided it would be better to treat first with glutaraldehyde just long enough for it to penetrate, and then hit the tissue immediately with the stronger osmium fixative to preserve the membranes intact and prevent bursting. It worked so well that we subsequently used this fixative for any small larvae where penetration was not a problem, and it usually works the first time. This is good for field work, where you do not have the opportunity to test fixative methods by repeated trial, e.g. you may want to fix something that you only find once. It has the added advantage that the tissue is fixed so thoroughly that it can be stored in alcohol and embedded later, sometimes years later, without significant degradation. This is not possible with conventional fixation. Also, the tissue will tolerate uranyl acetate staining at 65 degrees C, which is very important for serial work. And finally, we find it differentially affects neurons depending on their functional state, so different classes of cells look different in the microscope (some are pale, some dense, etc.). In contrast, the goal of conventional fixation is usually to make all the cells look the same, which does not help when you want to distinguish different types. In short, there are lots of advantages if the semisimultaneous method works on your specimens.
Nevertheless, when starting with a new type of specimen, you should ideally test several different fixation methods, including conventional ones (for larvae, papers by Cloney describe reasonably good conventional techniques. These usually require careful balancing of osmotic concentrations, temperature etc., because the tissue has to survive 1 hour before it is transferred to osmium). For our method, these minor adjustments don't seem to matter. The problem with ours, however, is that it doesn't work when there is a significant barrier to penetration. It doesn't work on arthropods, for example, even tiny larvae, because they have a cuticle, and it doesn't work well on post larval stages where the juvenile tissues are present and getting thicker. It's best on delicate things with thin epithelia (and it works well on yolk in some cases). So the question with non-larval subjects is whether they basically fix like larvae or not. For juvenile and adult organisms, I would expect it to work well only on very small species where the tissues are not too thick.Method (step by step):
(1) For Fixative 1 (the primary fixative) I use sealed 2 ml vials of 70% glutaraldehyde (bulk glutaraldehyde in bottles tends to go acidic after a while, in sealed vials it is OK). I usually make 50 mls of solution, which is enough for many fixations of small larvae. First dissolve 2.14 sodium cacodylate (the buffer) in 50 ml distilled water; add the 2 ml of 70% glutaraldehyde, and any cloudiness should eventually clear. Then add 3 gm sucrose and let it dissolve. Store in the refrigerator, and it should keep for weeks or longer. I usually test the pH with pH paper, and it should be 7 7.6. Neutral is good, a little alkaline probably doesn't matter, but acidic is bad. The overall concentration should be about 2.8% glutaraldehyde in 0.2M buffer with 6% sucrose for osmotic balance.
(2) For fixative 2 (osmium) I use sealed vials of 4% aqueous osmium tetraoxide, which come in 2, 5 or 10 ml sizes. You must always handle this carefully in a hood to avoid fumes. >>Be careful with it - experienced colleagues can tell you the precautions to take.<< My old supervisory professor once was blind for a day because of fumes the osmium fixed the cornea of his eyes. So, open the vial in a hood, then I put it in a plastic vial with a good top that seals. I use it in small amounts (for larvae, 5 or 10 ml will do lots of fixations), and then store what is left by freezing. Refrigeration is not enough, because the fumes will fill the fridge and ruin it. However, if you seal the vial in a larger vial, also with a good seal, and freeze it, the osmium will keep for a long time. It will not oxidize and go black, and fumes do not seem to form when it is frozen. This is good, because it means you can use a little bit each time and not waste it, so long as you re freeze it quickly.
(3) Fixing procedure. To fix, I take a few larvae in a small volume (a few mls) of seawater in a small plastic dish (tilted up so it doesn't have to be full), and then double the volume by adding isotonic Magnesium chloride to relax the larvae. This takes only a few moments. Then I add Fixative 1 dropwise until I have doubled the volume again (so the gluaraldehyde is about 1.4%; keeping everything as cold as the specimens will tolerate probably helps also at this stage). I leave this for about 30 seconds to 2 or 3 minutes, depending on the delicacy of the larvae (20 30 sec for very delicate, several minutes for amphioxus, but after about 2 minutes you will start getting tissue damage. As soon as the time has passed, I start adding Fixative 2 (osmium) dropwise until I have about 2% osmium concentration in total. Then I leave the mixture for 30 min. If the larvae are nicely black, I leave for another 30 minutes, but if they are not blackening, I will sometimes add more osmium or move the larvae to fresh osmium solution. During this time a flaky precipitate may form, but this does not matter, so long as you can still find the specimens at the end.
After the hour in osmium, I pipette the larvae into a vial and wash them with several changes of distilled water, allowing 15 30 minutes for the osmium to diffuse away. Getting the osmium out is important, because it will otherwise precipitate inside the tissue during the next step. Leaving the specimens in the water for a while will not harm them, but it is good to get to the staining step without too much delay. This involves leaving them in 2% aqueous uranyl acetate overnight (8 - 10 hours, usually in a small vial) at 60 - 65 degrees. The temperature is critical; below 60 will not work, but 70 is OK, so it's better to be a little high than too low. This may seem like a brutal treatment, but in fact it further fixes the larvae as well as staining them, and then no subsequent staining of the sections is required. This is the key step that makes easy serial sectioning possible. The greatest loss of sections occurs when you try to stain the sections on the grids, and stain deposits will often contaminate the section anyway. We don't have these problems because we never have to stain anything so long as the tissue has been pre stained. The uraynl acetate generally has to be freshly made, though it will keep for a few days or weeks in the refrigerator, and it has to be kept away from the light. Any precipitate in the bottle means you need to make a new batch, and it takes a day or so to dissolve (also in the dark). The stain may precipitate around the larvae while you are staining, but this cannot be avoided. An alternative is to make the uranyl acetate up in 55% ethanol or methanol, and then the staining can be done at room temperature, but think the hot method is better. After uranyl acetate treatment, wash in several changes of distilled water, then dehydrate in a series of steps (only a few minutes in each step is needed for small specimens) to 70% alcohol, and then they can be stored. There is no need to rush to embed like in other methods, though rapid embedding may help a little. But we have kept well fixed specimens treated in this way for years before embedding and they were fine. Conventionally fixed specimens usually have to be embedded right away, and the staining step doesn't always work either, because the heat may damage the tissue.
This may sound like a lot of effort, but it really is very fast. I can take the osmium out of the freezer, warm it just enough to thaw it, fix in a few minutes, leave for an hour, decant and into uranyl acetate in a few minutes, into the oven overnight, and a few minutes to dehydrate in the morning. So we can do dozens of separate fixations of different larvae in a day.
We always use Spurr's resin (medium hardness, which is the middle of the 3 options the manufacturer usually lists). Spurr is very odd. When you trim the specimen with a razor blade, it appears to be very brittle, far more so than Epon. However, at the micro level, Spurr cuts like butter the sections peal off easily, even with glass knives. The disadvantage is that the sections are weak, and collapse under the electron beam, so you either have to use tiny mesh grids (200 at most, which is too small for good medium power photos) or grids with Formvar support films. Since we need to have support films anyway, this is not a problem, but for beginning your work, it might be easier to start with very small mesh just for test subjects. You could try embedding in Epon, which is stronger, if you have colleagues familiar with it, but it is much more difficult to cut. It does give you good training in correctly tilting the block, however, because the tilt has to be perfect to cut Epon. In any case, for Spurr, you go from 70% alcohol to several changes of pure alcohol, then to a mix of alcohol and propylene oxide, then pure propylene oxide, then a mix of resin and propylene oxide, and finally several changes of Spurr (I forget the details here, but it's basically standard procedure). The only thing to watch is the polymerization: check that there are no hot spots in the oven. Spurr typically gets hardened for 8 - 12 hours at 70 degrees, but if you use a shallow metal dish for the specimens as we do, you have to keep it from contacting the metal of the oven. Otherwise the bottom of the resin hardens first, so the hardness is uneven, and the block is difficult to cut smoothly.
Now you are ready to try sectioning. Practice with glass knives, and try making Formvar films as well, and your technique with both will improve. For serial work, there are special tricks for getting good films and picking up sections on coated grids (this is harder than with uncoated grids), but I will tell you those later when the need arises. There are two potential problems you may have that we do not, due to humidity. Humidity is very low in Saskatchewan, and both Spurr and Formvar are sensitive to moisture. We use our solutions for long times, and they get better with time rather than worse. But other people in other cities have more problems, and Kyoto is probably pretty humid in summer. So sometimes the Spurr is better if it is not too old, the Formvar has to be made up more frequently, the solvents (propylene oxide for the pre resin steps and dichloroethane for Forvar) have to be kept dry, usually with water absorbing pellets), and the caps have to be kept on tight.
Here are some tips now on the sectioning procedure itself. If you are just beginning, you will try different methods of holding the grids and picking up sections, and perhaps wonder how one can ever get serial sections at all. So here are a few biginning tips.
To begin with, to see all of each section, you need to use slot grid so that no part of the specimen is covered by grid bars. This adds a difficulty, because sections are much easier to pick up on grids with bars, and most people will start with those when they learn. The advatage is that the excess water will flow through the grid, so the sections are not carried away by the water current. With a slot grid, you have to make a support film (for Spurr resin, you will need a support film anyway, so this problem cannot be avoided), so the water cannot escape through the grid. Instead it has to flow over the edges, and the sections will spill off also. So the main problem after the sections are cut is to get the ribbon onto the grid without having them flow away and break apart. If that happens, then you have to corral them like ducks on a pond and try again.
There are basically three ways to pick up sections.
(1) Using an empty slot: this is the most reliable for technicians with less training, but I don't like it. It is what is done in some of the medical labs that have tried to master the technique. Basically you pick up the ribbon from underneath, coming up through the water from below with the grid as flat as possible, but you use a a grid without a support film. The sections will be caught in a bubble of water held in the slot by surface forces and will float on top. Then you bring this first grid down up side down on to a second grid with a support film so that the slots aligned. The sections with a bubble of water will jump to the second grid, and now the sections will be sandwiched between the water and the support film. The grid can then be set aside to dry. Usually the second part of the procedure can be done with the grids held in adjustable clamps, so there is no human error in bringing them together. In some labs it is done by two people, one to do the first step, the second to transfer to grid number two. So it is automated to a degree.
(2) Using a support film grid from above: in this method you hold the grid flat with forceps and come down on the freshly cut ribbon from above. Surface tension will cause the water with sections to leap up to meet the grid and stick to it. If the grid is aligned properly, the section will adhere directly and firmly to the support film, and the water can be drained off with a piece of filter paper without dislodging them. This is by far the easiest way to pick up section without errors in my view. The problem is that when they leap, the sections do not remain fully spread as they should, and folds often develop. The bigger the section, the more this is a problem like trying to keep a blanket spread out. For a long ribbon of very small sections, this might be the best method.
(3) Using a support film grid from below: this is the method we use, but it also requires the greatest manual ability and steadiness. Coming from underneath, and because of the angle you must hold the forceps, it is almost impossible to keep the grid flat enough that the sections are not trying to flow off the sides. Different coatings may have different attractive properties, but in our experience, formvar is usually a bit hydrophobic, which increases the tendency of the water and sections to flow away. There are three solutions. (1) You can hold the ribbon in place with a probe (we use an eyelash hair glued to a glass needle, also useful for moving sections around and cleaning the knife edge) in your other hand. This risks puncturing the film, and it is very difficult to hold both hands steady enough the make this method work every time. Plus, it puts stress on the ribbon, which may break. (2) Using the surface tension in your favor (this is the method used by my most successful technician). First you need to have a good ribbon and to get one end to touch firmly against the support film just where you want the ribbon to start sticking. You do this with the long axis of the slot perpendicular to the water surface. The edge of the ribbon will usually adhere quite firmly where it touches. Then, if the ribbon is correctly aligned, you can pull the grid out of the water keeping the long axis in the same orientation, even at a fairly steep angle, and the ribbon will lay down along it in an orderly fashion. This is because the excess water is running off to either side, so there is no force on the ribbon itself to disturb it, and there is also no excess water to be removed at the point of contact. (3) A modification of method (2) was developed by my most recent technician, and this takes less skill. That is to turn the grid 90 degrees and pull it out of the water with the long axis of the slot parallel with the water surface. This works if you furst push the ribbon up sideways against one side of the slot and force it to touch all along the lateral edge. Then, again, the ribbon will just lay down on the film as you pull the grid out because there is no excess water to eliminate between the ribbon and the support film where they touch.
(1) You may have to change the water level every time you pick up the sections. This is because a certain shape of water meniscus is required in order to see the sections while they are being cut, in order to judge their thickness. But a convex meniscus is better for picking up and getting the right surface tension. You have to experiment to get the right combination.
(2) The water in the boat has to be cleaned repeatedly when you're trying to get a good series. Grease builds up rapidly and makes a mess of the support film, so the boat has to be emptied, rinsed and refilled about every 2 3 ribbons. In a good microtome, this can be done without disturbing the position of the knife, so it will start cutting right away again if you're careful.
(3) Getting good ribbons can be difficult. The block faces have to be clean, and we brush across the knife edge with a hair to clean it during cutting. The front and back edges of the block need to be parallel, and the side of the block face are usually trimmed at opposing angles (to make a rhombus) so the front side that you cut into is longer than the side you come out of. The wider you can make the block, the more distance there is for the sections to adhere, so that can help. Sometimes we make the sections much bigger than we would otherwise need as a way of maximizing the distance for sticking. Usually once a block is giving good ribbons, it will keep doing so. If you're using a diamond knife and alternating sections are crumpling or tearing away from the knife edge, it is a sign to adjust the knife angle. With glass knives, it may just be a sign that the knife is dull.
(4) Spurr compresses when it's cut, so once you get a ribbon, waving a stick immersed in xylene close to it will make it spread out, often by increasing the area by 20 30%. This helps the ribbon stay together because the sections are expanding against each other, and they lock. If you have scattered sections, you can herd them together and spread them, and this again will help them lock together in a bunch.
(5) Section thickness. EM people usually like to have silver or pale silver sections for high power work. They become irregular and tend to tear if you get any thinner, and we like thicker ones anyway, since they're easier to handle. Mostly we cut sections that are gold when cut, turning to light gold or silver when spread. Brown or purple sections are too thick. The support film itself will also ideally be light gold, i.e. about the same thickness, so you won't get the contrast needed for really good high power work.
(6) Formvar films. We use standard recipes, which I think are for 4% formvar in dichloroethane. Everything has to be dry and usually filtered before use, and in moist climates, you might have to make it up fresh (it takes a day to dissolve). We put it in a glass slide holder (a Coplan jar used for staining slides while they stand vertically) with a cover. Put in a slide so it's about 2/3s immersed, put on cover for a few minutes (helps get a saturated environment), then drain slide and stand upright to dry. We then use diamond stylus to scratch around the margin on one side, and plunge the slide slowly at an angle into a large bowl of clean water (you clean the surface by adding a drop of the formvar, letting it solidify and drawing it off). If you are careful, the film on the top surface of the slide will lift off and float away. Check it carefully for imperfections my technician usually used a magnifying glass to do this. Then use only the good (smooth looking and gold to light gold) parts. Areas of roughness will be full of tiny holes. Place grids side by side upside down on the film. Then place a square of paper (filter papaer can work) over the film and pick it up off the surface. Dry and store in a clean place.
(7) Grids. This is very importent once you get to serious work. It is crucial to use special rigid grids. So far as I know, these are available only under the trade name "Synaptec", but I think you can get them through most suppliers. The key feature is that they are stiff enough that they do not flex when you handle them, and flexing breaks the support film. It is this advance alone that makes serial work routinely possible.
(8) One micron sections for light microscopy. Spurr is somewhat troublesome for cutting thick sections. You need to keep them as thin as possible, as you lose the cutting advantage of Spurr at much over one micron, and if sections are too thick they peel off in an odd way that may cause some damage the block surface. The other problem is toluidine blue staining. The issue here is that the stain penetrates Spurr very slowly, so longer incubation times are more heat is needed, and the latter in particular causes the sections to stretching which produces wrinkles. We found it difficult to get good flat sections for light microscopy, especially when the sections are large.