First of all: How do you define "direct ELISA"? You are coating your analyte on the plate and detect it by a primary antibody and subsequently with a labeled secondary antibody? In this case, the purity of your sample is extremely critical. How do you prepare your samples?

Hi and thanks for the response,

I'm new to the terminology, so sorry if I'm mis-using the term "direct ELISA".  I am essentially doing the same procedure as described by ABCAM here:

I am detecting the analyte/antigen with a biotinylated primary antibody, and then streptavidin -linked horseradish peroxidase is taking the role of the secondary antibody (i.e. the streptavidin should bind to the biotins on the primary antibody). 

I believe the sample was purified using a Nickel column via a His-tag which was then cleaved off.  I believe the purity was then assessed using NMR.  I am not the person who purified the protein, though, but I have been told that the purity is good.   

The sample was then lyophilized in what I believe is PBS, and it was reconstituted in the "Coating buffer" I defined above.  I aliquoted the sample into 50 uL volumes and then the aliquots were then frozen at -80 degrees celsius until they are used in the assay.  The amount of protein that is in the sample is quantified via nanodropping using a ThermoFisher protein A280 (procedure here: https://tools.thermofisher.com/content/sfs/brochures/Thermo-Scientific-NanoDrop-Products-Protein-Technical-Guide-EN.pdf).  Sorry if I wasn't being clear.

I should have mentioned this in my original post, but I have done a few controls where I have modulated the concentrations of the three proteins.  So, I had assays where I increased the vMIPII concentration and kept the primary antibody and strep-HRP low, and then an assay where I drastically increased the concentration of primary antibody but kept the analyte and strep-HRP low, etc.  I found that when I greatly increased the amount of primary antibody, there was the greatest increase in the development of color. But there was also an increase in color when I added increased levels of the other proteins.  At this point, I haven't double-checked to see if this result is reproducible, so before I use up more reagents, I want to make sure that this phenomenon of poor polyclonal antibody affinity for the purposes of ELISAs is not unheard of.

Thanks a lot for your help, and have a great day.

I think that your ELISA protocol is more or less OK. Be aware that the coating step is critical because it is non-specific. The sample needs to be a more or less purified protein. If you have too much of non-related proteins or surfactants in your sample, your assay will fail completely. All variations of your reagents will not help you. In the case you cannot purify your sample protein sufficiently, you need a format called "sandwich immunoassay". In this case, you need 2 antibodies against your protein; one of them is first coated on the plate. If you could tell me in more detail, what your sample is composed of, it would be easier to tell you, how to proceed.

A short comment on the polyclonal/monoclonal issue: It is a frequent misconception that monoclonal antibodies are better than polyclonal ones. In an analytical context, often the opposite is the case. The major disadvantage of polyclonal antibodies is that cross-reactivities may be difficult to control. But low affinity is more often a problem of monoclonals. In your application, polyclonal antibodies should be more reliable and easier to handle.

Hi and thanks so much for getting back to me!

Regarding the purity of the analyte- I understand, I unfortunately don't have that data on hand at the moment.  The protein was lyophilized over 2 years ago, so I'll have to go digging through my colleagues' lab notebooks to find what exactly happened to the protein sample that I'm using.

Is it okay if I ask you a couple of tangential questions regarding ELISA assays?

Since the coating step is non-specific, will the wells bind any inorganic and organic molecules?  Would a plate bind these molecules in the absence of coating buffer (from personal experience, it seems that in the absence of a basic solution, no protein that I am using is able to adhere to the wells unless it is in massive quantities, such as in the case of the 3% BSA in TBS).

Why would it be a frequent misconception that monoclonal antibodies are better than polyclonal antibodies?  If anything, the reason why our lab purchased polyclonal antibodies is because it seemed like there would be a higher liklihood for a polyclonal mixture to bind multiple different epitopes per analyte protein.  In the case of a monoclonal antibody, incidental effects such as antigen orientation while it is being plated & unfolding/denaturing could lead to Monoclonal AB not binding entire analyte proteins.  It just seems like the common thought would be to assume polyclonal antibodies are superior in ELISAs compared to Monoclonals.  

Although I am familiar with a sandwich immunoassay, I am curious why it is crucial that the primary and secondary antibody would have to be different for a polyclonal antibody.  While I understand that 2 antibodies would need to be used if one of them was a monoclonal antibody, I don't quite follow why I'd need two different kinds of polyclonal antibodies.  

Sorry if I am missing something obvious.

Thank you for your time.

Hi Arjan,

perhaps your lyophilized sample is simply degraded... You should not spend too much time with these old, ill-defined samples.

Coating: Usually, only larger, organic compounds, such as proteins or larger peptides will bind on a plate. Coating buffer is often not needed, only the coating efficiency might vary. If there is a coating problem with a protein, usually traces or larger concentrations of a surfactant are the problem.

Polyclonals/Monoclonals: Yes, your statements are correct from my point of view. It is definitely dependent on what you want. You obviously do not care about cross-reactivities, since you want to target different epitopes, which can occur on different proteins, sometimes on unrelated ones.

Sandwich immunoassay: It is difficult to explain this assay here in detail. One important point is that in a sandwich immunoassay you need 2 epitopes bound by 2 different antibodies. Yes, it is possible to perform a sandwich immunoassay with one immunoserum, because it contains already a set of different antibodies against different epitopes. However, in this case, you need to label one fraction of the serum e.g. with biotin. In many cases, a monoclonal (mouse) antibody is used for the capture step and a polyclonal (rabbit) antibody for the detection antibody. Subsequently, a labeled anti-rabbit secondary antibody (e.g. from goat) is incubated.