IMMUNOPEROXIDASE FOR ELECTRON-MICROSCOPY

[Modified for tissue culture cells, by Vega-Salas et al., J.C.B.104: 1249-1259, 1987; originally described for tissues by Brown and Farquhar, Cell 36:295-307, 1984]

1) Grow your cells on a substrate suitable for sectioning. Ideally, it should also be transluscent. Some good substrates are:

Translucent filters (e.g., Transwell, Costar, or Falcon).

Reconstituted Type I collagen gels (e.g. from rat tail).

Matrigel gels.

Filters seem to embed better in Spurr (with a PO step). If you can embed in Epon, you can also grow your cells on tissue culture plastic Petri dishes, embed WITHOUT a PO step and, when the Epon has hardened for 1-2 days, detach the block from the plastic (the cells will go in the Epon). The drawback of this is that the monolayer will be on one side of your block. Therefore, you must re-embed in fresh Epon.

Whatever the substrate you choose, make sure it will embed and section properly before you start your experiment.

Remember to plate cells for your NEGATIVE CONTROLS!!!!

2) when the cells are ready, wash them in any isotonic saline solution without amino groups and with Ca-Mg (E.g., PBS, 0.1 mM CaCl₂, from nw on: “PBS) fix them as hard as you can. Ideally, you should use 3% PFA, 0.1% glutaraldehyde in PBS. If this is too much for your antigenicity, try 2% PFA alone.

It is highly advisable to test the fixation and posterior permeabilization steps by immunofluorescence before proceeding with the EM. That will also ensure that your antibodies work.

Fixations in methanol or methanol/acetone may be suitable for IF, but are hopeless for EM.

Time of fixation may vary with your cells, but for a monolayer is usually – 30 min.

3) Wash in PBS and permeabilize membranes. One of the best permeabilization procedures seems to be saponin in PBS (try 0.005% to 0.05%). It is mild and preserves membrane structures. It may extract PI-anchored proteins. 0.1% TX-100 is good but harder on membrane ultrastructure. In general, membrane proteins do not extract in TX-100 after aldehyde fixation.

It has been shown though, that both Tx-100 or alcohol permeabilizations can induce relocalization of proteins ini the cell, even after fixation.. Proper controls include different fixations and different permeabilization methods, judged, at the very least, by IF procedures.

4) Kill the endogenous peroxidase. If you want to skip this step you should run controls without antibodies. In general, this step does not affect antigenicity. However, it is conceivable that it may do so in specific cases.

Treat your cells with 2% hydrogen peroxide in PBS for at least 5 min. If you have a lot of endogenous peroxidase, as in the case of some tissues, you may want to go for 5%, 15 min.

5) Quench free aldehyde groups by your favorite method. I use 50mM ammonium chloride in PBS for 5 min. Aminoacid solutions work well. If you used glutaraldehyde and you are experiencing problems, you may also want to try the borohyde. In my experience it does not make any difference, but many people have reported substantial improvements in the signal/background ratio.

6) Quench inespecific background as well as you can. Our option is 1% BSA, 50μg/ml pre-immune IgG (of the same species as your second antibody) for 20 min. The addition of 0.1% TX-100 to the buffers helps; however, if you want membrane preservation, the same reserves described in (3) are valid. So, in general we avoid it.

7) Add your first antibody diluted as much as possible. Again, preliminary IF will be helpful to determine optimal dilutions of the first antibody. Most antibodies will work in 20 min. Occasionally, low affinity ones may need longer times.

8) Wash (e.g. 3X PBS, 5 min each).

FROM THIS POINT FORWARD, ALL THE SOLUTIONS, INCLUDING PBS, ETC., MUST BE CAREFULLY CHECKED FOR pH 7.5 AND ADJUSTED IF NECESSARY. REALIZE THAT MANY BUFFERS USED TO STORE pH METER ELECTRODES CONTAN AZIDE, A POTENT INHIBITOR OF PEROXIDASE, SO WATCH YOUR pH ELECTRODE EXTENSIVELY. ALSO, PEROXIDASE MUST BE PROTECTED FROM LIGHT.

9) Add a second antibody coupled to peroxidase. Affinity purified antibodies are a must. Again, 20-30 min is usually enough. Fab fragments coupled to peroxidase are the best for tissue penetration!

10) Wash 4X in PBS, 5 min each.

11) Incubate in 1% BSA in PBS for 20 min.

12) Rinse rapidly in PBS once (less than 1 min), and FIX in 1% glutaraldehyde in PBS for 30 min. The reason for this step is to create a tight network of a cross-linked BSA within the cells, to contain the diffusion of peroxidase products in the next step.

13) Wash 3X in PBS.

14) Wash 3X in 50mM Tris-Cl, 7.5% sucrose, pH 7.5.

15) Add the diaminobenzidine. You can use Sigma pellets. Otherwise: 0.2% DAB, 0.01% hydrogen peroxide in buffer (14). Watch your cells every 5 min under invested microscope WITHOUT phase contrast. A pattern in brown identical to your IF images should develop in 5 – 30 min. Occasionally, very scarce antigens may take up to 2 h. In this case you should replenish the DAB solution with a fresh one every 30 min. If you cannot see an image clearly resembling your IF, something has gone wrong. Check your reagents and don’t even bother to continue.

16) Wash 3X in your standard EM fixation buffer: (e.g., 0.1 M sodium phosphate, or 0.2 M sodium cacodylate; 7.5% sucrose helps preservation of structure, but I don’t know why).

17) Post-fix in 1% osmium tetroxide, 1% potassium ferrocyanide (add the ferrocyanide from powder immediately before use), in the same buffer you used in (15), for 45 min. This step is very important, because DAB product is not opaque to electrons by itself.

18) Wash 2X in distilled water.

19) Incubate for 1 h in saturated uranyl acetate, wash 2X H₂O.

20) Dehydrate in alcohols (50, 70, 95 and 100%), ~5 min each. Try to avoid the PO step. If you have to use PO, do it as shortly as possible. Embed.

Although the samples can be stored in alcohol 100% overnight (at this time you are usually homebound), it is advisable to start the embedding at once, leaving the samples in the pre-embedding step (e.g. 50% Spurr, 50% ethanol).

21) Lead post-staining should be avoided at least in some grids to have a good contrast from peroxidase product.

Needless to say, comparison with appropriate negative controls will tell you if your dark images are specific or not.