Western blots are commonly run in order to determine what proteins you have in a sample, and in what quantities they exist. Their incredible sensitivity (nanograms can be detected in certain cases!) allows you to discriminate one protein from all others in a fashion that is very important for research. After performing this technique, you are left with a picture that shows only the protein that you are interested in. A quantitative comparison of different samples can be achieved by running a known amount of the protein you are interested in alongside your unknown samples. In order to run a western on your protein mixture(s) of interest, you need to do several things to prepare.
First, you need to know the quantity of total protein in your sample(s). This will help to make sure that you do not use too much, as westerns are very sensitive (mentioned above). It will also help in figuring out the amount of your protein of interest as a percentage of total protein in your starting sample. This can be done by using a homogeneous control of the protein that you wish to quantitate for comparison. I use approximately 2ug protein of each sample that I wish to perform a western on. I determine the total protein by doing a bradford assay, or a simple colorimetric determination that compares a standard BSA curve to unknown values that lie within that same curve, evaluating concentration of protein in the unknown samples.
Samples that are to be analyzed must have a homogeneous charge that relates to the size of the protein. This is the only way for the size of your protein to be determined. Normally, proteins can have an overall positive or negative charge, and therefore do not travel at the same rate or direction during separation by electrophoresis. SDS-PAGE solves this by using a detergent which gives proteins a standard charge that relates directly to amino acid chain length, following up with an electrophoretic technique used to separate sizes of molecules on the order of proteins. If you are trying to determine if a protein subunit is present, or need to dissociate proteins from one another, you must also treat your sample with beta-mercaptoethanol. This stuff smells like a mixture of rotting eggs and ass, so try not to leave it open to the air for too long (thiol groups tend to do this). So, before you run poly-acrylamide gel electrophoresis (the PAGE of SDS-PAGE), boil your protein samples in a mixture of buffer (preferably colored), SDS, and beta-mercaptoethanol for five minutes.
Run your SDS-PAGE at an appropriate percentage of acrylamide to allow for the proper polymerization to occur. I use a 14% gel to separate large glycoproteins. It is easier to handle than lower percentage gels, and does not tear as easily, however a 12% or lower gel may be required to separate higher weight proteins. Loading protein gels is a bitch, as the color of the samples is so light. Load a molecular weight marker first (they are nice and dark), and then hop over lanes from there.
After running your samples on a gel, they need to be transferred to a support membrane, commonly made out of materials such as nitrocellulose or PVDF. These will be the materials that are handled from now on, though the gel can be stained with either silver or coomassie blue to make sure that the proteins are actually affixed to the membrane, not still on your gel. This step is not as easy as it sounds, as you can easily transfer for too short an amount of time (which would mean that you have to redo your transfer), or too long, which causes the loss of your samples. This sucks a lot, because you have to start all over again (It seems hard to do, until it happens to you and you cry like the little girl you are).
These membranes are washed in PBS/tween, then blocked. Blocking means that you allow the membrane (or "blot" as it is now called) to sit in a mixture of PBS/tween and protein (normally milk) overnight. If you are performing a western for a cow protein that may easily be found in milk (aka BSA), block in another protein, or just leave in PBS/tween overnight. Blocking reduces background signal when you perform your blot, allowing for the protein signal you desire to show up more distinctly.
Now comes the actual process that is referred to as a western, however the above steps are all needed to get to this point. Take your membrane, now sufficiently blocked, and wash it in PBS/tween (thoroughly) to get rid of your blocking solution. Apply whatever you are using to stain specifically for your desired protein to your blot. This is normally an antibody, made in mouse, goat, rabbit, etc. Let us say for argument's sake that we are performing a western for BSA. You would purchase a anti-BSA antibody, let us say mouse anti-BSA. You dilute this as specified, or if the company you bought it from is full of wankers, at the dilution that works best for the money (these things can cost upwards of a thousand dollars per ml). Dilute in the same solution that you blocked in.
Now comes the visualization part. If you are unlucky enough to buy an antibody that is not conjugated to some sort of reporter, than you have to wash your blot and use another antibody that is directed to the first one. If I have to do this, I always choose a horseradish peroxidase conjugated new animal anti-animal of the first antibody (in this example HRP conjugated rabbit anti-mouse). HRP is easily visualized with chemiluminescent dyes. No radiation involved, none of that messy radioactive waste crap, none of the worry of mutations causing me to birth three eyed monsters.
Pictures of your phosphorescent dye-labeled western are easily taken with Kodak lab-bitch film, and these can actually be fun to play with. These are great for qualitative determinations, however they are not suited for quantitative determinations. For this, there is the phosphoimager. When this is broken, you are SOL, and tell your PI that you cannot give them quantity comparisons. Then, they ask you to do it again. YAY!!
Written from memory, as if you cannot tell from the bitterness.