Do you think Histone deacetylase is less abundant in your sample? It may be below the detection level in the input. I suggest instead of concentrating (is that right cut-off membrane you used?; may be it passed through the membrane due to its higher order/lower order structure) with centricon, try making TCA precipitation which gives your whole protein content as a pellet.

Do you think Histone deacetylase is less abundant in your sample? It may be below the detection level in the input. I suggest instead of concentrating (is that right cut-off membrane you used?; may be it passed through the membrane due to its higher order/lower order structure) with centricon, try making TCA precipitation which gives your whole protein content as a pellet.

Hi Christian,

I think it will help if you provide further detail such as the no of cells used in the Co-IP, the % of input you've loaded and whether or not you did over-express your transcription factor, etc…Anyway, if you think it is a dilution problem, then yes I agree with Ananda you can use TCA precipitation.

Well I thought exactly the same. However, since after centricon tubes I usually load all the input, I don't know if TCA precipitation will change a lot. The CO of the centricon tubes is well under the MW of the histone deacetylase..anyway, i was wondering if it could be possible that a fraction of the proteins could remain attached to the membrane during the spin of centricon tubes. When I did gel elettro-elution to recuperate proteins from sds pages, in the protocol it was suggested to reverse the voltage for ten seconds to detach proteins from the celophan membrane put around the gel to avoid the loss of extracted proteins. In this case TCA precipitation should do the trick. I will try. Thanks a lot

I really confused with "electo elution.... stuffs". I agree as you said, there might be proteins sticking on the membranes. I strongly suggest TCA because you need proteins only for western blots not for any downstream process (where centrion/similar methods are highty recommended), TCA precipitation is relatively quick,easy to do and at the end it gives you the total protein as a pellet (be careful not to lose the pellet during wash). Good luck.

Hi Christiano,

are you using the supernatant from the IP for input? If so, why don't you simply keep some of your lysate from before the addition of antibody? You might simply have depleted all HDAC from the lysate with your Co-IP. Concentrating proteins for input purposes is not the best approach in my eyes, since you never know how much is sticking to the membrane.

I don't know if I understood your approach correctly, but if this IP is about the differentiation of HDAC bound and HDAC unbound to a certain TF you could run into a lot of problems, since the amount of antibody used, buffer conditions, number of washes and so on can strongly influence the outcome.

Using an antibody directed against the HDAC and probing for certain TFs might be an alternative.

Best,

Kai

Thanks Kai, but if I use the lysate not immunoprecipitated how I check if there's a change in the HDAC quantity after immunoprecipitation with different antibodies and how I check controls?? I tried to immunoprecipitate directly the HDAC, but it does not work...don't know if the antibody is not good for CoIP or if it's due to the high number of different TFs that interact with this molecule, so that the level of the TFs I am interested in is really low. Thanks for your answer.

Maybe I don't understand you correctly but do you ever see the deacetylase in WB without enrichment with IP? If you have loaded the same percentage (of volume) of your initial sample and your IP eluate and don't see anything in initial but in the eluate than something is wrong. Are you sure that the band in WB acctually corresponds to the correct protein? You could run an SDS-PAGE, stain for protein and do in-gel digestion of this band to verify its identity with mass spectrometry.

Hi,

You say "When I analyze, by western blotting, the fraction of proteins bound to the beads works well. When I analyze the fraction not bound to the beads (the input) I cannot detect the histone deacetylase in the controls, even if I use the same antibody used to analyze the bound fraction."

1. This is not surprising at all. As you would guess, the concentration in the IP is so much more than in the input. ALL of protein in the lysate is now in the beads and in say 2o uL of sample buffer, vs say 200 uL of lysate. So 10x enrichment....

But you then did the correct experiment:

"I tried to immunoprecipitate directly the HDAC, but it does not work...don't know if the antibody is not good for CoIP"

2. And you gave the correct answer !!!. IF YOU ARE USING THE SAME ANITBODY FROM WESTERN FOR IP, the antibody in the IP is binding to the undenatured protein, while after SDS-PAGE the antibody used is recognizing the denatured protein. Big difference. Antibodies for Western blotting do not necessarily work in IP.

SO you need to find a HDAC antibody specifically for IP (other papers perhaps??). You can still use the other Ab for Western detection. You should not go further until you can detect the HDAC by itself by IP. There is low concentration of HDAC in your lysate.

3. OR you can lyse your cells in 10x LESS volume of lysis buffer (or same volume used for loading IP on gel). It will be viscous, but you only need 10-20uL. And then do a total lysate since you have now the same concentration as in IP.

4. Centicon-ing crude lysates and losing protein is and age-old problem !! :-)

Good luck !!

Hi Christian,

What you are calling input is your flowthrough fraction. You are not really interested in the unbound fraction, except to make sure the co-ip is working at all steps. You may also experience some protein loss by non-specific binding to the tubes. Ideally, you would compare the %bound HDAC to total HDAC in the lysate and see if this changes with higher TF levels. There are fewer variables to worry about than comparing bound to flow-through. To do this, perform IP on the TF first, then western blot for HDAC. You would have to do this for the TF that is increased and another control TF. Alternatively, IP for HDAC, then probe for TF of interest and a control TF and comapre to total TF from input for each. If the increased TF is competing with the control TF, you will see less control TF after Co-IP.

Regarding antibodies: this is usually the most difficult part for IP. You'll need two great antibodies, one for IP and another for western. They will not likely be the same antibody as discussed earlier by Prem. Before you begin IP, make sure you have worked out the antibody conditions for western detection. If this is not working well, you cannot trust your IP.

If the IP conditions are difficult to work out from the mouse cells, you might consider engineering a fusion protein of the HDAC with an affinity tag and performing pull-down assays with competing TFs.

Dear Jason, the first part of your answer is exactly what I did. I already compared the samples IP with two different antibodies (the antibody for the TF of interest the putative competitor) and the results are great. I do not agree that checking the "flow-through" has no sense. All papers showing this experiment had this control. This means that it matters...regarding to the second part of your answer, well I work on tissues, so I guess I should electroporate the fusion protein with a gfp expressing vector in the tissue (it's the only way to have it at high levels) and than do a cell sorting. Guess it would be complicated and I would not like to study this system on cell lines...however, thanks a lot for your answer

Just FYI, nowadays companies sell specific antibodies that works for IP as well as WB purpose, we have to just check them in the data sheet before we buy it.

Hi Christian,

Thank you for entertaining my suggestions. And sorry for getting away from the original topic of detection of the HDAC in the flowthrough fraction. It is possible that it is completely bound to the column. Have you seen it in this fraction before?

Also, could you attach one of the papers you reference? I can't wrap my head around using the flowthrough fraction as anything other than an internal control to make sure the IP is working. You would still need to compare your bound proteins to total protein in the input.

I think electroporation of tissue and cell sorting is a much more complicated method than over-expressing your proteins in cell lines. Also, you don't necessarily want to overexpress the HDAC in the tissue because there would be enough HDAC to bind all of the TFs. I think what you are observing in the tissue is a limited pool of HDAC and the TFs are competing for binding. Overexpressing HDAC in the tissue would mask the competition. With cell line expression of a fusion protein and pulldown assays, you can titrate that amount of TFs in the system and show competition, but the trade-off is that the results do not translate directly to what is occuring in the cell.

The attached reference is also looking at HDAC binding competition between TFs and may be useful. They just look at input and elute for their co-IPs. They also use ChIP to detect the displacement of factors.

Good luck!

Think I got a lot of new ideas from your answers. Thanks pals. I will try and let u know. The winner is...u will know in the next episode

Hi Christian,

I completely agree with Jason. It is not about winners but interaction with your peers makes you foresee the reactions and concerns that reviewers will tell you had your work to be sent for publication. I think you should imperatively use input. I am not sure about the rest as you didn’t provide more details about your experiment, understandably though. You don’t need to respond to this. Are you considering a TF and a mutant variant which you think has higher affinity for HDAC complexes? Do they still bind the same DNA sequence and is it with the same affinity? Anyway, assaying for HDAC in input is not a big deal. I will always try to optimize the extraction procedure for both the TF and its variant, etc…At least you should show that they extract equally well...

I totally agree on what u said about peers Abdelhalim and the story of the winner was just a bad joke. If u read the previous answers it was me insisting that the input is a fundamental control. However, to be more specific, I study a mutant in which one of the genes differentially expressed (in this case strongly up-regulated) binds quite specifically an HDAC, subtracting it to another transcription factor. For the moment I am going to try TCA precipitation to concentrate my input, rather than centricons. To assess if the other TF still binds the upstream promoter of its downstream targets I am actually doing some ChIP experiments on wt and mutant mice to verify if there is any variation in the binding of these sites. I assure u I really appreciate all the observations made by all of you and I will follow for sure some of them. Hope it will work. Best