Hi Mihaela. I know this must be frustrating, but this looks really interesting from a biochem point of view. I have several ideas:

1- The difference in the width of the lanes between pull down and input suggests different buffer conditions between samples. They should really be in similar conditions for comparison because all kinds of things can happen to proteins in different buffers, like aggregation, proteolysis, etc. Try eluting your beads straight in sds sample buffer (if you are eluting in glutathione).

2- Does your protein have any catalytic/enzymatic activity? Can it be chopping/ubiquitinating/aggregating the potential interactors? Can it be binding to proteins that modify the interactors?

3- I think the ~60kD band you see in every pull down in figure 2 (even a little for pull down 3) is the same protein that a) reacts non specifically with your flag antibody and b) binds non specifically to your glutathione beads. You don’t see it in input because you don’t load enough and the beads effectively enrich for that nonspecific protein. Try a pull down with lysates that don’t have any flag-tagged protein and check.

4- figure 1 is most interesting. It almost looks like a monomer and dimer of your cleaved protein that lost the epitope for the specific antibody? Maybe that interactor binds strongly and aggregates your protein, this might cause proteolysis and covalent dimerization. Do you use protease inhibitors? Try eluting in sample buffer with 50mM DTT for 10 min at RT and then add Iodoacetamide at 100mM for 10min more. That should reduce any aggregation.

Hi Bryen,

Thank a lot for the suggestions, and I'm glad you find it interesting J My answers to your comments are following:

1- I'm already eluting the proteins from glutathione beads so the conditions in the input and pull downs are different since inputs are in 1X SDS sample buffer, and the pull downs are in 2X SDS sample buffer. I could dilute the buffer before adding it to the beads to check if that's going to help.

2- My protein is a peptidase, but as far as we know, its substrates are short peptides, 3 to 10 amino acids long, it shouldn't digest another protein. In the first pulldown (Fig. 1), I was trying to pull my protein from the lysate. Also, I should mention that GST-protein that I used for pulldown in that case wasn't completely clean, I had a couple of unspecific bands, but I don't have the Ponceau staining picture for that western.

3- I have a GST control for pulldown, and I don't see any bands in it. I see inexplicable bands only in some pull downs. I will try to do a pull down from the lysate without FLAG-protein. Thanks for the suggestion.

4- In this case (Fig. 1), I'm trying to pull down my peptidase with a GST-protein whose size is around 115 kDa, but as I mentioned before that GST-protein wasn't completely clean. My peptidase is quite stable and it doesn't make dimers or aggregates. Also, I'm pretty confident that none of the 2 bands are my protein or its dimer, because I blotted it with the anti-serum to my protein and did not get any signal. I am using protease inhibitors now, but I did that pull down a while ago and I did not use the inhibitors back than. I don't know if I have any of that GST protein left since I decided to reverse the procedure and use the GST-fusion of my protein to pull down the putative interactors, but if I have any of that GST-protein left, I'll try to add DTT and Iodoacetamide like you suggested.

Thanks once again for all the suggestions. I have one question, what are the optimal protein concentrations in the lysates that you use for pulldown. Mine are usually around 1 mg/ml. Is that too high?

1- Why do you have to elute the beads in 2X? Maybe the increased SDS in the pulldowns (2X=4% SDS) solubilizes some nonspecific protein that reacts with the flag antibody? I would definitely try to normalize between inputs and pulldowns. Since a protein moves down a gel based on how much negatively charged SDS binds to it, it could make a difference in the size. But probably not as big as you see.

2- Interesting that it has some catalytic activity. I understand that it should only cleave peptides, but don't forget that when you pull down this way, you are greatly increasing the concentration of the peptidase/substrate in a small area (around the bead). Who know what could happen....

3- I see. It just seems too much of a coincidence that you see the exact same sized band in the pulldowns of all those different proteins (Fig.2). You even see it in the control (the upper of the two bands). I can't imagine what else it could be but a nonspecific protein. Definitely try the pulldown without flag protein.

4- I don't know which HA antibodies you are using, but these antibodies are usually very clean. Unless you have contaminated plasmids and expressing some other HA-tagged protein, If you are seeing such a strong signal, then it is hard to believe that it is not your HA-tagged peptidase. Does the antibody you have against your peptidase target an N- or C-terminal region? If yes, then maybe you don't see the signal with your anti-serum, because that region has been cleaved off the protein. That would explain the small decrease in size (of the lower band) and the dimerization of that fragment. Many peptidases undergo auto proteolysis for activation. And again, it may be very stable, but concentrate it greatly with a pulldown and it could start misbehaving.

5- That concentration (1mg/ml) is usually what I use. The lysis buffers matter a lot for protein interactions, so definitely keep that in mind.

Zelim ti srecu!

Hvala :)

1-I got that kind of protocol from my colleague, but maybe the difference is that she washed the beads in tubes so there's always some buffer left on the beads and when you add 2X SDS buffer it gets diluted, while I use spin columns for washing the beads and I essentially add the 2X buffer to the pure beads. I didn't think that might be the problem. Even if the higher SDS is solubilizing some unspecific protein, it is still strange that it reacts with anti-FLAG. The antibody is quite specific otherwise.

2-Peptidase is the dipeptidyl peptidase 3, it cleaves dipeptides from the N-termini of short peptides. I don't know if it could cleave a protein, but maybe it's possible when it gets concentrated.

3-I'll try the pull down without FLAG protein and see how it looks.

4-I'm using anti-HA antibody from Sigma, and I don't see unspecific bands, but I usually load low amount of protein (5-10 ug total protein). My antiserum was prepared against the whole peptidase so I would assume that it would have to detect even N- or C-terminal fragments, but as I already mentioned the GST-.protein that I used for that pull down wasn't that clean so maybe these bands are from some unspecific protein, but it is still strange that it reacts with anti-HA antibody.

5-Thanks. The lysis buffer that I use is 150mM NaCl, 20mM Tris (pH 8,0), 1mM MgCl2, 0.1% NP40, 10% glycerol. I also wash the beads with it after the binding. I add protease inhibitors for lysis and I wash the beads without inhibitors.

Is the 0.1% np-40 a typing mistake? 0.1% detergent is used to lightly permeabilize Cells for immunocytochemistry... can’t imagine you getting much protein like this. Usually non-ionic detergents like NP-40 are used at 1% (maybe 0.5% if your protein is sensitive..).

No, it is 0.1 % NP-40. I found the protocol on the internet. We lyse HEK293T cells, and when we add 400 ul of lysis buffer per 1 well in 6-well plate, we get concentration around 1 mg/ml (measured by Bradford). That seems good enough to me, and since I don't get any binding in the GST pull down, except with my control protein, I assume that raising the concentration of detergent won't help, but in the light of unspecific binding, maybe I should do that. I got these putative interactors from SILAC-MS, and I can not confirm any of the interactions by co-IP or GST pull down, and I am having a hard time excepting that maybe none of the proteins that we identified in SILAC-MS is actually binding our peptidase.