Posted 29 May 2009 - 10:22 AM
Got a bit of an annoying one here. I'm trying to do Co-IP for two proteins that are known to interact - this was published way back in 2001, and it's been shown since then in a number of different ways by different groups. For various reasons, rather than just looking at endogenous protein I'm trying it by over-expressing a Myc-tagged version of one of the proteins (via transient transfection) and looking for Co-IP of its (endogenous) interacting protein.
The Myc-tagged protein expresses very well, and is pulled down just fine with anti-Myc antibody...but I'm not detecting even a whisper of the other protein. I've tried using different kits, doing it old-school (no kits, just antibodies and proteinG beads), fiddling with the buffers and so on...nothing works. There's a drop-deadline on this one so things are getting a wee bit desperate.
The conditions are not stringent - I wanted to *see* the pull-down before increasing the stringency to get rid of non-specific rubbish - and even more puzzling I'm using the exact same cell line that was used in the first two papers to demonstrate Co-IP of these two proteins. These papers showed pretty robust interaction, although to my intense annoyance they didn't give even a hint in their materials&methods as to *how* they did the IP (not even one damn buffer composition).
So, I guess I'm asking if anyone's had a similar experience of trying to replicate a well-established (and published) interaction and just had it fail totally...did you figure it out? Any ideas however bizzare or off-the-wall would be greatfully recieved.
Posted 31 May 2009 - 03:41 AM
I would check the endogenous target protein expression in the cell line. I'd see if the endogenous is detectable or otherwise. This will gives me an idea if the Ab is working, and that the endogenous protein level is detectable.
Also, I'd keep IP waste and run parallel with the IP and input. If IP waste shows presence of the target protein, the IP is either not working or, there is no interaction. But I guess with two independent labs being able to replicate the interaction does strike off the latter probability.
Check if the myc tag is stearically blocking the interaction domain of the bait. E.g. if the above papers used a C-term or N-term tagged proteins.
Posted 01 June 2009 - 01:57 PM
Maybe you could use the cells previously published as a positive control.
I think Isek makes some very good points as well.
Posted 02 June 2009 - 07:38 AM
Yup, the endogenous protein's detectable, actually highly so as the antibody we're using is great - I'm talking (for a whole cell lysate) of loading about 10 micrograms on a gel, then probing the membrane with the antibody @ 1:3000...and going for a ~10 second exposure. So, by CoIP (considering the enrichment factor) it should be readily detectable.
Also, we of course run 1% and 5% inputs from the raw lysates on the gel along with the IP samples, the endogenous protein's detected just fine in these.
Stearic blocking: Given we're talking a small tag here I think it unlikely but will consider this further...unfortunately these previous groups were looking at only endogenous proteins I believe...will check it out.
This is the really odd thing - we *are* using exactly the same cell line as previously published! This is why the non-result is so baffling.
The mystery rolls on...thanks for your input, I think I need to search the literature and see if anyone's done a tagged exogenous before and CoIP'd the endogenous protein.
Posted 04 June 2009 - 09:57 AM
the efficiency of the myc-IP is high judging by your inputs.
Is it possible that you are not IPing enough material? How much are you IPing?
Maybe you need to increase the amount to 2 mgs or so.
Posted 05 June 2009 - 09:12 AM
Or try to IP the second protein and watch for the myc-tagged one?