In my experience, this concentration is by far too high. Since IgG is one of the main constituents in serum (about 10 g/L), no blocking in the world can avoid a minor binding of IgG to the plate. A dilution of 1:10,000 of serum would be a good starting point.

In my experience, this concentration is by far too high. Since IgG is one of the main constituents in serum (about 10 g/L), no blocking in the world can avoid a minor binding of IgG to the plate. A dilution of 1:10,000 of serum would be a good starting point.

Yes - neat serum will give a high background.  You need to dilute way down

I agree with Weller. By any chance the antigen you are using is of bacterial/ recombinant origin? in that case you can expect high background due to anti E coli antibodies (which is common) in the serum. 

Perhaps you should try a dilution of 1:100 and 1:1000 to see if there would be any significant differences from the 1:10 you are accustomed to. Perhaps 1:10,000 could be resorted to if these fail.

Hi all, thanks for your suggestions. I tried diluting the serum down to 1:10,000, and the background indeed decreases... but so does the specific signal. The problem is the antibodies we are looking for are not likely to be present at a high enough concentration to detect, and we lack a positive control. I'll try some other blocking reagents.

Why don't try diluting  the antibodies more. 

I also suggest you do an optimization experiment using different concentrations of both the serum and antibodies so you can establish the optimum condition for your assay..

The simplest  thing is to not wash the plate with PBS-tween. Tween reduces the capacity of protein binding to polystyrene, including that in the blocking solution. Just dump out the antigen-coating solution and add blocking solution to excess volume. However, minimizing uncontrolled variables in ELISAs is essential, and using mysterious "SuperBlock" from ThermoFisher is one of them. For example, non-fat dry milk in PBS can be a very effective blocking agent, especially in Western blots, but because it containing a complex mixture of proteins, there can be unexpected artifacts such as binding of these proteins to the target antigen or binding of serum anti-casein antibodies to casein in cow's milk.  ThermoFisher keeps secret the composition of SuperBlock -- I would never use such a mysterious reagent.  A generally ideal blocking solution is bovine gelatin at 1 mg/ml in PBS-- no tween!.  The 10X stock solution may need to be heated to ~60 C to fully dissolve some forms of gelatin. Diluents for the serum and anti-Ig may also be important, and these should contain tween.  I also agree with other respondents that 1:10 serum dilution is problematic; perhaps you can use longer incubation times, even overnight, so that higher dilutions can be employed.

I agree with Robert.. I've been this strategy for 4 years and it's working very efficiently.

I would set up a series of serum dilutions to test your background. I'm familiar with 1:100 being common for human samples, so perhaps try as suggested above a log dilution of 1:10, 1:100, 1:1000, 1:10,000. I am not surprised 1:10,000 doesn't give signal though. If you don't have a positive control, you really do need to do a series of dilutions. 

Other strategies to reduce background in your ELISA is to make sure you are washing adequately with PBST. Next, I would suggest you do another titration series with your secondary antibody. You should be able to then identify a concentration of both serum and secondary in which your immune vs. non-immune serum is clearly distinguishable.

Good luck!

I had similar problems with high background. I solved it with the following protocol:

Using no-protein block from Pierce. Serum was incubated for 1hr at 4`C. 2nd antibody was a F(ab)2. Most important was the Tyramide. It is an amplification diluent. Check out ELAST from PerkinElmer. With this additional step I was able to dilute the serum up to 1-10,000 (instead of 1-100 which resulted in huge background) and still got some positive results. At the end I used 1-2000 for all my samples to not to miss weak positives. You can find more details about the protocol in my paper: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4348150/