When you say signal; are you immunoblotting the IP and is your protein of interest the same mass as immunoglobulin? Try HRP conjugated to Protein G (bioRad), we use it to detect proteins around 50-60 kDa in immunoprecipitations.

When you are blocking your western blot, do not use milk or BSA. Use serum from the same animal that your secondary antibody was produced in. For example, if you have goat anti-rabbit, you will block and incubate with the antibodies in 10-5% NGS. It decreases background very nicely.

That is always a headache to us when the protein(s) is around 50kd or 25 kd. I agree with James that HRP-conjugated Protein G or Protein A may be the solution since they bind to intact IgG but not heavy or light chain. The problem for IgA will be very poor or no biding if your antibody is mouse IgG1. Protein G theoretically binds all kinds of commercial primary antobodies but is much more expensive than Protein A. Wish you luck. Beads method is the newest method but our lab tech never able to make it to work for us.

Running all of the proper negative controls is essential. You will find in addition to all of the IgG bands, for example that protein A or protein G alone will come off agarose beads when boiled in SDS sample buffer and these proteins give their own bands in a primary antibody-dependent manner. For the IgG problem it's most useful to IP with a rabbit antibody and probe with a mouse antibody, but one trick to use when you have to absolutely do a rabbit-rabbit IP-Western and look at a protein around 50kDa is to elute your protein with SDS sample buffer without reducing agent (DTT or BME), so that they heavy and light chain of the IgG remain covalently bound and migrate predominantly above 80kDa. If running these samples next to samples containing reducing agent, be sure to skip a well in between, since the reducing agent from one lane can creep over to the edge of the next lane.

In addition to running the correct controls (listed previously)... There are a couple things you can do to reduce these 'blazing' IgG bands.

(A) IF you have your own immunoprecipitating antibody source (ie you have the hybridomas or have bleeds from a rabbit) it is quite easy to crosslink your antibodies to protein G or A. (This is very easy. Pierce sells many crosslinkers that work well for this and are easily neutralized via free amine groups in tris). After you have immunoprecipitated your protein(s) of interest boil your IP as normal but without the reducing agents. Transfer the supernatant to new tube add a reducing agent, boil again, and run as per usual. This procedure works well if the light chain is the problem (ie the band around 21kD) it also works well for the heavy chain (running in around 50kD) although in this case you do not need to first boil without the reducing agent.

(B) If you do not have a lot of antibody to work with and you have the ability to change either the IP antibody or the immunoblotting primary do so such that they have been raised in different species. For example you may IP with a rabbit antibody and blot with a mouse antibody. You wold then use a secondary antibody that has minimal crossreactivity to the other species (in this instance an anti-mouse with little/no reactivity to rabbit). Such antibodies can be purchased form Jackson immunochemicals.

(C) If both antibodies are mouse monoclonal antibodies and they are not of the same isotype then you can purchase and use isotype specific secondaries (again Jackson immunochemicals is a great place to get such antibodies)

I would just like to add 2 things...use the sepharose beads that are made for this purpose as well as use HRP conjugated antibodies that will decrease the background. You will be surprised!