Protocol Online logo
Top : New Forum Archives (2009-): : Immunology

Custom antibody does not work as expect - (Dec/14/2009 )

Hi All,
I am a new guy come working with antibody. Recently I've got a great trouble and hope you guys can help :lol:

Say I have a peptide LVRPEVDVMCTAFHDNEETFLK. Because of the hydrophilicity, the sequence DNEETFLK in the C terminus are used to prepare an antibody to capture the original peptide. We pay to a biotech company who is expert in this. The company use the DNEETFLK-KLH conjugate as immunogen, and DNEETFLKC (with an extra C in C termins) as the antigen when they screen the antibody from animal serum. For what I've know, the extra C is for conjugation with the solid substrate during immunoaffinity chromatography and shouldn't be an issue I am going to mention.

Okay so I have the antibody, antigen (DNEETFLKC) and the LVRPEVDVMCTAFHDNEETFLK peptide. I did a direct ELISA with DNEETFLKC. It works perfectly (at 2ng /well level) . I did a competitive with DNEETFLKC, it works too. Now I did a direct ELISA with LVRPEVDVMCTAFHDNEETFLK, no, I don't see any signal. I coated the plate with DNEETFLKC and use LVRPEVDVMCTAFHDNEETFLK as the competitor. The longer peptide (serial dilution from 3 ug to 1 ng /well) didn't bind to the antibody at all.

At first we though there should have some trace residues causing the trouble from the synthetic LVRPEVDVMCTAFHDNEETFLK (done by other company).
I've spiked the DNEETFLKC to the LVRPEVDVMCTAFHDNEETFLK and the competitive ELISA works. So this eliminated the residue factor.

Then I suspect it is related to the orientation of the antibody-antigen pair. The epitopic sequence here is in the C terminus but not N or in the middle of LVRPEVDVMCTAFHDNEETFLK . However, the antibody recognizes the DNEETFLK from the N terminus during the antibody production.

My question is: will this be the troubling causing factor? that I will need an antibody recognizing from C terminus of DNEETFLK if I need it to recognize LVRPEVDVMCTAFHDNEETFLK?


Thanks again for your patient to read the long post. I will send an enquiry to the antibody production company too but I also appreciate any comments from you guys! ;)


Jake

-jakekei-

is it possible that the peptide is folded in a way that blocks the epitope from the antibody?

or

maybe the peptide is binding to the plate in such a way that the epitope is not accessible to the antibody?

try detecting the peptide by western blotting (i know, it is a small peptide, so use a tricine gel). sds should unravel any folding and ensure that the epitope is accessible.

-mdfenko-

mdfenko on Dec 15 2009, 01:39 PM said:

is it possible that the peptide is folded in a way that blocks the epitope from the antibody?


I think it is not likely. The other end is rather hydrophobic. When the peptide dissolves in buffer (tris 10mM) or water, they should not have stronger interaction than that between
the antigenic portion and water..

mdfenko on Dec 15 2009, 01:39 PM said:

maybe the peptide is binding to the plate in such a way that the epitope is not accessible to the antibody?

try detecting the peptide by western blotting (i know, it is a small peptide, so use a tricine gel). sds should unravel any folding and ensure that the epitope is accessible.

Thanks for this point. Actually when I found nth in the direct ELISA for the long peptide, I got this in mind. But when I coated the short peptide and use the longer one as competitor
in competitive ELISA, I saw nothing too. So I think that's already ruled out this possibility.

Western blot seems to be a good idea, I will think about it :).

-jakekei-

I was looking into having antibodies made this way about a year back because we needed to detect about 20 different proteins and were looking for a way that was less laborious then cloning the genes, purifying the proteins, and raising antibody to them. We decided against this method because all of the companies I contacted (about 5) would only guarantee that the antibodies they produced would recognize the synthetic peptide antigen to which they were prepared; none would guarantee that they would work at detecting native protein...

-HomeBrew-

HomeBrew on Dec 15 2009, 07:03 PM said:

I was looking into having antibodies made this way about a year back because we needed to detect about 20 different proteins and were looking for a way that was less laborious then cloning the genes, purifying the proteins, and raising antibody to them. We decided against this method because all of the companies I contacted (about 5) would only guarantee that the antibodies they produced would recognize the synthetic peptide antigen to which they were prepared; none would guarantee that they would work at detecting native protein...


So..typically people use the middle part of a long peptide (10-20 amino acid) rather than part of the C or N terminus as the epitope, and that usually works (to capture the original longer peptide)? ;)

I know people work in proteomic study do have library of antibody ( a single antibody per each target)..

-jakekei-

Typically, algorithms are applied to the protein sequence to predict contiguous segments likely to be antigenic (see, for example, this site). The problem is that none of these predictive methods are perfect, although your chances are better if your segment is selected based on it passing multiple tests.

The bigger problem is that none of these scale-based tests takes the three-dimensional conformation of the native protein into account very well, thus some antibodies prepared against the custom peptide will fail to recognize the native protein because although the sequence is there, it doesn't form the same epitopes as the custom peptide does. This is also why some of the antibodies thus prepared may only work in some applications (e.g. they may work in ELISA but not in westerns, or vice-versa), if they work at all.

-HomeBrew-

HomeBrew on Dec 16 2009, 05:11 AM said:

Typically, algorithms are applied to the protein sequence to predict contiguous segments likely to be antigenic (see, for example, this site). The problem is that none of these predictive methods are perfect, although your chances are better if your segment is selected based on it passing multiple tests.

The bigger problem is that none of these scale-based tests takes the three-dimensional conformation of the native protein into account very well, thus some antibodies prepared against the custom peptide will fail to recognize the native protein because although the sequence is there, it doesn't form the same epitopes as the custom peptide does. This is also why some of the antibodies thus prepared may only work in some applications (e.g. they may work in ELISA but not in westerns, or vice-versa), if they work at all.


Thanks for the very informative explanation & the website suggested. I've learnt a lot from them (and more confidence for the upcoming
graduation exam :wacko: )

I've got reply from the company who made the antibody. Hope the issues would be solved soon.

-jakekei-

jakekei on Dec 14 2009, 02:42 PM said:

Then I suspect it is related to the orientation of the antibody-antigen pair. The epitopic sequence here is in the C terminus but not N or in the middle of LVRPEVDVMCTAFHDNEETFLK . However, the antibody recognizes the DNEETFLK from the N terminus during the antibody production.

My question is: will this be the troubling causing factor? that I will need an antibody recognizing from C terminus of DNEETFLK if I need it to recognize LVRPEVDVMCTAFHDNEETFLK?

Jake


what you have written above is the most likely problem. we have made antibodies to the Nterm antigenic site of peptides and or fragments generated from the specific cleavage of a much larger intact transmembrane protein to monitor production of said peptide to various treatments. this allows us to discriminate btwn cleavage products especially when multiple peptides can be generated from the intact protein.

digest your synthetic full length peptide with Asp-N endopeptidase...it should digest at at the Asp6 and Asp14 P1' site...then do the elisa compared to undigested full length peptide. if the antibody recognizes the DNEETFLK from the Nterm as you think, you will get a positive signal on the elisa.

i assume this is a monoclonal Ab.

-eldon-