13 replies to this topic

[leptopelis]

Hello!
I'm trying to establish IP in our lab. My problem is that after IP my IgG band is much stronger then the antigen band. If IP would be perfect, then should't the antigen band be about as strong as the IgG band? (I use the same antibody for IP as for the subsequent Western blot).
I ran on my gel the original lysate (that has not been immunoprecipitated) and the supernatant of the IP reaction too, and in the IP supernatant there is as much antigen as in the unprecipitated lysate. And the solute eluted from the beads (that is the precipitated sample) contains as much (or even a little less) antigen as the unprecipitated sample in spite that the immunoprecipitated sample must be much more concentrated then the unprecipitated one.
I tried to reduce the ammount of the antibody, but in that case both, the IgG and the antigen band were lowered in the IP sample. Why does not bind my antibody more antigen? (The antibody I use is very good.)  
Any answer would be appreciated.

[little mouse]

About concentration by IP, I cannot answer.

but if you don't want to see the Ig band in your western, you should use a HRP-conjugated protein G instead of the HRP-conjugated anti Ig. protein G will mainly recognize the native primary antibody you use to detect the antigen, but not the denatured Ig that are loaded on the gel.

[leptopelis]

My problem is not the presence of the IgG band, but the small size of the antigen band compared to the IgG band.
Nevertheless, later the presence of the IgG band might be a factor to consider, and so thanks for the idea. So far I have thought of crosslinking as a method to eliminate the IgG band, but this other method might work as well.
So the question is: why doesn't my antibody bind more antigen?

[little mouse]

from my experience, I never had a concentration of my antigen by IP.( It was not the goal. )
When I want to purify an antigen, I dilute more my sample, and I load the sample on a column where the antigen is covalently linked. the antigen is eluted by decreasing the pH.
I don't know if others have a concentration of their antigen after IP?

[leptopelis]

Well, I just thought that I should get about as strong antigen band in my IP as the IgG band (or a little lower as not all of the antibodies bind antigen of course). But my IgG band is much stronger than the antigen band. And the supernatant of the IP reaction (that is the solution remaining after IP) has as strong antigen band as the control (not immunoprecipitated) sample. I don't know if it is normal or not.

[mikew]

Important parameter:

What is the species of your primary?
Is it the same as your secondary?
If so, it is likely anything you see is an artifact. Primary and secondary
antibodies really should be different species. This is especially true for Rabbit and Goat antibodies.

What size is your protein?
Why do you think the amount of your IPed protein would be equal intensity to your IgG band?
This is utter nonsense. Your IgG band will always be stronger. You antibody may not be very efficient.

[leptopelis]

Yes, all my primary antibodies are of rabbit origin. I thought that one molecule of antibody binds one molecule of antigen. Though not all the antibody binds antigen, I thought that their intensity must be around equal. But thinking again it is very reasonable that only a portion of the antibody binds antigen. I simply didn't know what to expect. So does protein A have a higher affinity to the antibody as the antibody has to the antigen? I had no information about it at all.
About artefacts: I hope, it is not.   I have a band at around 50 kDa (it must be the IgG, as it is missing from the control sample and more importantly it is missing from the non precipitated sample). The other band is a bit higher and it seems to be O.K., as the antigen is protein kinase B (Akt), whose mol weight is aroun 56 kDa.
In the non-immunoprecipitated sample (and in the supernatant of the IP reaction) two strong bands are present (the higher one a bit stronger but the lower one a bit more compact) while in the IP samples only the lower, more compact band can be seen (and the IgG of course). It is important that first I tried to catch estrogen receptor alpha, but failed. In that case I saw the IgG band only. So I hope that my other band in the last IP recations is the Akt.
I used 400 mcl lysate for the IP reactions and at the end eluted the beads in a volume of 80 mcl (including the beads). Non precipitated samples were the original lysates diluted with 2x sample buffer. And after Western blot I see that the IP reaction contains as much (or a little lower) ammount of Akt as the non-precipitated samples. It is not what I expected.

[leptopelis]

Could I improve my antigen yeald if I first incubated the protein A-agarose beads with antibody and then added these beads to my lysate? My idea is that in this way I could raise the local concentration of the antibody on the surface of the beads.

[little mouse]

leptopelis, on Feb 19 2009, 09:58 PM, said:

So does protein A have a higher affinity to the antibody as the antibody has to the antigen? I had no information about it at all.

protein-antigen interaction has the higher affinity

[little mouse]

leptopelis, on Feb 20 2009, 09:04 AM, said:

Could I improve my antigen yeald if I first incubated the protein A-agarose beads with antibody and then added these beads to my lysate? My idea is that in this way I could raise the local concentration of the antibody on the surface of the beads.

I never tried this way.
normally you have an excess of antibody compared to antigen, and an excess of beads compared to  antibodies, to be sure to IP as much of antigen as possible (that's why you will always have a higher band for antibodies than for antigens.
if you incubate first the beads and antibody, one could say that you could have a sterical "encombrement" that doesn't allow the antigen to bind to the antibody; or not so well.
I don't know. Have a try

[leptopelis]

Thanks for all the answers.
This sterical inhibition is something to consider. Next week I'm going to try IP with more beads (I used 20 mcl of it (50% packe volume). I'll go on with 40 mcl. And anyway I will try the "antibody first with the bads" method too.
But unfortunately I can't affors to use antibidies of different orgin to IP than to the subsequent Western blot.

[rkay447]

Increasing the beads will do nothing but increase your background or non-specific binding.  I am a strong advocate for prebinding the antibody to the bead before adding lysate.  I recommend the following: take 10-15 ul of protein A agarose beads.  Wash twice and add antibody with lysis buffer or pbs.  Let bind for at least one hour at room temp with the tubes agitating end-over-end.  It's important to keep the beads in slurry and not at the end of the tube.  Wash the beads twice with buffer and block your beads in fresh made and filtered 5%BSA/PBS for at least another hour but I prefer four hours to overnight.  Meanwhile, make your lysate and pre-clear by incubating with control rabbit IgG bound to agarose beads.  This helps remove non-specific binding in the IP.  After at least one hour but again, I prefer at least four, carefully remove the supernatent from the preclear beads (you don't want to add any of the non-specific binding back to the IP).  Bradford the lysate to check protein concentation.  Wash your BSA-blocked beads at least 2-3 times.  It's critical you get out any free BSA.  Add lysate, incubate for 4hours to overnight.  The time needed here completely depends on the antibody.  Longer times might increase the antigen IP but might also increase background.  This you much optimize to your own conditions.  Wash the beads well, boil in 20-30ul sample buffer and run on the gel.  For detection in the western, because you are using both rabbit antibodies in the IP and the western, you are  lighting up  the heavy and light chains of the IP antibody which makes the western (especially for a 56kDa protein) nearly impossible.  Rather than using anti-rabbit HRP as your secondary, you should try protein A-HRP.  Otherwise I can't stress the importance of having a mouse antibody for the western blot.  If money is an issue, find a group that published with the mouse and request an aliquot.  Call the companies and beg for a sample aliquot.  There are many ways to get your hands on antibodies without paying for them.  Also, if your advisor has published a paper using a santa cruz antibody (and I think abcam has starting doing this as well) you are entitled to a FREE antibody!

[mdfenko]

leptopelis, on Feb 19 2009, 03:58 PM, said:

Yes, all my primary antibodies are of rabbit origin. I thought that one molecule of antibody binds one molecule of antigen. Though not all the antibody binds antigen, I thought that their intensity must be around equal. But thinking again it is very reasonable that only a portion of the antibody binds antigen.

since you are using antibodies from rabbit i assume that they are polyclonal. were they made against a small portion of the protein or the whole molecule?

with polyclonals, especially if made against whole protein, you will see more than one molecule of antibody binding to one molecule of antigen.

you can see how this might affect the intensity of staining.

[mikew]

Hi again,

I think hormone receptors can be very difficult to IP.
You might want to think about transfecting in a myc-tagged estrogen receptor or FLAG-tagged and IPing
with the Flag (or myc) antibody.
  Here is a recent paper where they did some co-IP with the Estrogen receptor.
Note that they use different species of antibody for IP and Western
However, if you are Co-IPing for the Estrogen Receptor and Western blotting for the Estrogen receptor the
intensity fo the band in your Western should be strong enough to out power any background.

Good luck.

Regulation of estrogenic effects by beclin 1 in breast cancer cells.
John S, Nayvelt I, Hsu HC, Yang P, Liu W, Das GM, Thomas T, Thomas TJ.
Cancer Res. 2008 Oct 1;68(19):7855-63.