Did you cross blot your FLAG-ip samples with V5 antibody? If there is no reaction, it is possible that antibody is not working.

Karthikeyan makes a very good point about the antibody. Another point is that it could be steric hinderence, this would mean that whilst the epitope for FLAG is readily accessible to your antibody in your complex, it does not mean that the V5 epitope will be accessible to the V5 antibody. This is quite common, during one of my studies on the interaction between two proteins, I found one of the proteins in the complex can only be immunoprecipitated using an N- but not a C-Terminal antibody, but after treating the cells the N-Terminal antibody can no longer pull down the complex, but a C-terminal antibody can, indicating a change in the conformation of the complex. Do you have either C- and N- terminal tags on your candidate proteins, because if you do you could try both. you could also try pulling down with an antibody to the protein rather than the V5 tag if it is available.

Good luck

Andrew and Karthikeyan have suggested some importante possible explanations that might help you address your issue. Here I add another possibility. It could be that your flag-protein integrates only a fraction of all V5 cellular complexes. When you IP with V5 the fraction of complexes that contain your flag-protein is diluted because the V5 will pull down also the other complexes that do not contain your flag-protein. Even when the antibody is added in excess you never reach condition of immuno-depletion because antibody-epitope interaction is governed by equilibrium laws and also because those commercial antibodies were selected to have intermediate characteristics that make them suitable for several applications

Abdelhalim also makes a very good point, and I would suggest that if you take a small amount of cell protein (say 30-50micro gram of cell lysate) and sequencialy immunodeplete it with your V5 antibody (Say about 5 times) then compare it with a sham depleted lysate (use IgG), then if there is a bona fide interaction which represents a significant fraction of your target protein, you should have depleted some of your Flag reactivity in your lysate. It is a good point well made by Abdelhalim that just because two proteins interact does not mean that all of the protein is committed to that interaction. If you cannot see an interaction you may need to try an alternative strategy such as a Gal4 2-hybrid system.

I appreciate your question and all the answer because I have been struggling with the same issue for almost a year. At first I was working with Flag and V5 tagged proteins and then I decided to change the V5 to HA tag because the antibody seems to be better. Even with the change, I haven’t been able to optimize my experiment. I have tried different amounts of protein, Flag beads, Flag peptide for elution, concentration of antibody and still nothing. I did an experiment where I transfected my cells as I usually do for the Co-IP and threw some cover slips to test by immunoflurescence my transfection efficiency. I noticed that not all my cells have been double transfect. So besides what has been mentioned by Abdelhalim that your flag-protein integrates only a fraction of all V5 cellular complexes you might not have all the cells expressing both of your proteins, which makes even harder to see some binding. Due to my problem, I have been considering to do PLA, I don't know if that could work for you too. Good luck!

Thanks Andrews,

I think the tow hybrid system is a good alternative and because there is no library screen to perform the approach could provide a straightforward support to the idea that the two proteins interact. However, negative results could not be considered negatives.

Thai,

That’s a superb idea. I don’t know why PLA is not yet so popular to assess protein-protein interaction in vivo, even though last time I checked the probes were sold by Eurogentec. I myself have some experience in that but for other applications and had the opportunity to ask Landegren, in whose lab the method was developed, on some specifics. For example the two interacting proteins shouldn’t lay above a distance of 140 nm away from each others. Maybe some optimization of the probes has overcome that. This also bring into consideration that interaction between two proteins is not always direct within a complex. It could be bridged through other proteins. Hence, the steric hindrance raised by Andrew, etc… This is also important to bear in mind when performing two-hybrid in yeast. For example if the complex doesn’t have its homologue one might have negative results due to the lack of other partners.

I know, it is never so simple but I find it nice…

The only other things I can think is to re-clone your constructs in both the C- and N- with different tags (HA, V5, FLAG) and try again, but this could be a real pain. You could try and re-clone them and make the proteins in bacteria and see if they interact, there are a number of kits. Also I do not think I was very clear about my 2-hybrid system suggestion, what I was driving at was a mammalian 2 hybrid system, and again there are many kits that you can use.

Good luck!!

Thank you for the thoughtful answers. My V5 antibody beads definitely immunoprecipitate my V5 proteins, (so much so I could see a glow in the darkroom!) and both blotting antibodies work well. I especially appreciate Abdelhalim's idea that perhaps other complexes are forming that dilute the amount of my Flag-tagged protein that binds. I was at a loss because the FLAG pulldown worked so well and was reproducible. I had also considered 2 hybrid methods and will have to investigate PLA more.

Abdelhalim, thank you for your ideas. During my experiments I will keep in mind all these details.

Sorry Andrews, I have misunderstood you about the 2-hybrid system you were suggesting. Yes you are absolutely right. I know how it does work; though I have never done it myself I think it is worth a try. I always associate 2-hybrid with yeast because last time one of my former bosses asked me to perform it I said no. It is very laborious and very prone to false positives. Regarding your in vitro suggestion I think yes, it is always good to build more support to a potential interaction. He can also perform IVT.

Thai, you welcome

Hi Donna,

I am sorry I didn’t see your post. You are welcome. Sometimes RG posts are a bit anachronic. Anyways, if this is the case then you have to follow Andrew’s suggestion on sequentially IPeing with the V5. That is, you subject your unbound fraction to another IP with V5 antibody. I think, 5 times and excess V5 antibody should be ok for an immunodepletion. Then pool all and load it in one well. Report back no matter what results you get. Hope it will work