I assume that you are sure that the fist anti-rabbit is not biotinylated.

Are both anti-rabbit obtained in the same species? If not, its  quite common and totally normal that anti-IgG antibodies cross-react with antibodies of different origins, not only with the ones against they were generated. You would have to use anti-rabbit from the same species (ideally from the same company) or  anti-rabbit antibodies (both alone and biotinylated) that explicitely say they have minimal cross-reactivity to the other species IgG. Both Sigma and Jackson provide anti-IgG antibodies depleted of undesired crossreactivities, but you have to make sure that the ones you are using fullfill this requirement. They can also offer other catalogue numbers corresponding to normal antibodies that cross-react with many IgGs. This could be the simplest explanation.

Once you check this, if you are using the right antibodies, you could try to reduce background by including skim milk at 4% instead of BSA (not only for blocking, but also to dilute the biotinylated antibody and the conjugate) or by diluting more these two molecules (perhaps they are too concentrated). But I would think first in anti-rabbit cross-reactivity

I assume that you are sure that the fist anti-rabbit is not biotinylated.

Are both anti-rabbit obtained in the same species? If not, its  quite common and totally normal that anti-IgG antibodies cross-react with antibodies of different origins, not only with the ones against they were generated. You would have to use anti-rabbit from the same species (ideally from the same company) or  anti-rabbit antibodies (both alone and biotinylated) that explicitely say they have minimal cross-reactivity to the other species IgG. Both Sigma and Jackson provide anti-IgG antibodies depleted of undesired crossreactivities, but you have to make sure that the ones you are using fullfill this requirement. They can also offer other catalogue numbers corresponding to normal antibodies that cross-react with many IgGs. This could be the simplest explanation.

Once you check this, if you are using the right antibodies, you could try to reduce background by including skim milk at 4% instead of BSA (not only for blocking, but also to dilute the biotinylated antibody and the conjugate) or by diluting more these two molecules (perhaps they are too concentrated). But I would think first in anti-rabbit cross-reactivity

Dear Aiqin, In general, BSA should work fine for blocking, I usually use 1% in PBST, 1hr at RT. Depending on the quality / grade of the BSA, it could contain some Biotin that later will react with Streptavidin conjugates and cause a false positive signal.

Having that said, when you repeat he experiment, do you get the same results? Have you already optimized your ELISA protocol i.e. done appropriate titrations of all antibodies and conjugates involved?

Dear Gertrudis,

Thank you very much for providing me those useful information.  The fist anti-rabbit is not biotinylated. And anti-rabbit IgG (alone and biotinylated) are both produced in goat and both are from Sigma. So it is supposed having minimal cross reaction. 

As for blocking, I have tried skimmed milk. It shows same results as using BSA. 

Today, I will perform this experiment again with diluted biotinylated anti-rabbit IgG  (200,000 times dilution) and conjugate (40,000 times dilute). Let see the results and I will keep you update.

Thanks again.

Aiqin

Dear Wolfgang,

I have repeat those experiment many time with adjusted conditions. However, I always got the same false positive results. I have not done the titrations. I have not used ELISA before. There are lots of things I need to learn. Could you please give me more instruction on how to perform titrations?

Thanks,

Aiqin

Titrations imply trying several dilution of  the biotinylated antibody (for example in different rows of the ELISA plate) and several dilutions of streptavidin-HRP (for example in different columns). Then you will determine which are the best dilutions to use.  Its very important that these procedure is not only performed on the anti-rabbit antibody (which you need to be negative) but also on something that should be positive (for instance wells coated with rabbit igG or anti-rabbit followed by the addition of rabbit IgG). The best dilutions will be those giving you the maximal signal/noise ratio. Are you diluting the biotinylated antibody and streptavidin-HRP in blocking soluction? I always do so, other people prefer to use PBS/Tween.  

Please dilute your biotin detection antibody 

and da-

HRP 

Could you explain the assay? It makes no sense to me.

1. Coat plate with what?

2. Add dilutions of what? 

Hello Aiqin,

Some of the basics of ELISAs revolve around blocking, which is the reason you are adding the BSA/PBS.  There are many, many binding sites on the plastic wells for appropriately charged molecules, and so I would wonder what your coating buffer pH is.  The second experiment you describe (BSA blocking first) - shows that your blocking protocol is sufficient, as the signal is obliterated by effectively coating the wells with BSA.  It also shows that the biotinylated antibody does not have BSA reactivity (some times BSA is used as a buffer reagent in antibody-production protocols unfortunately).  I would also suggest, as Gertrudis did, to dilute the biotinylated Ab (essentially your detecting antibody) in blocking solution.

I strongly think that if indeed the second experimental set of conditions produced no substrate development, then with everything else being equal, you probably need to consider that the two anti-Rb Ab's have some cross-reactivity.  One test you could do across 8 wells or so is to coat with 7 dilutions of anti-Rb and one with BSA/PBS.  Then block as usual, then detect with the biotin version, same dilution across all wells.  A differential signal pattern would strongly suggest that the signal is due to Ab reactivity.  That's the question you need to answer first in my view.

Thank you all!  Here is my update to the question.

According Gertrudis and Samar suggestions, I diluted more for both biotinylated anti-rabbit IgG (200,000 times dilution) and conjugate (40, 000 times dilute). Now, I do not have color for negative control. 

Thank you all again and good luck in your research.