Just a quick note that I have updated my ChIP protocol using the Covaris instrument and preliminary data says it is working much better. I will have more on this as I get more data back, but I am really starting to believe that if you want to do ChIP seriously, you have to have a Covaris.
Ethan-omics 
Ethan
Nice job indeed , your blog is a great support for me . Thanks
I didn’t see your Covaris protocol, mind sharing the link again( excuse me if I missed out something)
It’s here now:
Click to access chip_covaris.pdf
If I update it again you can still find it under the protocols tab at the top of the page.
Yes we have a covaris here and I am trying my hands on that , from some prelim gel profiles, I found the sonication is prettu nice(except that you cannot do more than 1 sample at a time).
I am concerned how the ChIP will look like(I mean data) , no sure if the Covaris method is too harsh for DNA?
Also do you feel 5 min HCHO fixation is good enough? We normally do 10 mins(that what most protocols say).
Also why you cautioned against using pure formaldehyde rather than methanol containing formaldehyde(I suppose methanol is added to prevent HCHO polymerization to paraformaldehyde).How would methanol in formaldehyde be detrimental?
According to Covaris, which of course you have to consider the source, shearing with the Covaris is much easier on epitopes then tip sonicators or the bioruptor.
As far as fixation time and shearing time, I think you really have to dial it in to your epitope. Some proteins are bound to the DNA much tighter, e.g. nucleosomes, others will shear off really easy especially proteins that are not directly bound to DNA. Next week I will do a chip where I’ve fixed cells for 5min and 12min. I will then shear them for probably 10 min and 20 min. It is a nucleosomal epitope so I really want to bust things up.
Covaris tells me that the methanol in 37% stock methanol can cause some aggregation. Also I saw a seminar once on a totally unrelated subject where the guy was showing electron micrographs of chromatin. There was substantially more protein associated with the chromatin when fresh formaldehyde was used. Cell biologist have been using fresh formaldehyde forever because it works better and is more reproducible. It’s funny molecular biologists never took note.
It is very important to add Triton X-100 and NaCl to the chromatin after shearing. The shearing buffer is not good for immunoprecipitation. The Triton sequesters the SDS and makes the conditions less harsh and more friendly to antigen-antibody-protein G interactions.
Thanks Ethan for the elaborate reply! Appreciated
For HCHO fixation, I wonder if there is any cell: HCHO ratio to fix. I do here 100 Million cells fixation in a 50 mL falcon with final 1% HCHO for 10 mins. What you feel, is that number OK? or we need to scale up/down cells. I want to avoid overfixation as much under-fixing!
By that way am at U of Chicago.
Hi Ethan, and Subhradip,
We have also found that methanol, and ethanol increase the permeability of cells thereby rendering the fixation more potent. We carried out a simple experiment of spiking methanol-free formaldehyde with methanol, and noticed a dramatic shift in the shearing efficiency. fixing with methanol containing formaldehyde for 5 minutes is the same as fixing with methanol-free formaldehyde for 20 minutes.
Thank you
Hamid
Hamid, that is very interesting. I was thinking that freshly prepared formaldehyde would be more potent. I’ve done a ChIP where I’ve x-linked for 5 and 10 min with fresh formaldehyde and sonicated each for 10 and 25 min in the Covaris, but we are out of Qiagen minElute columns to purify my IP’s and there is no word from our distributor when our order of more will arrive. So it may be a little while before I know the results. (On a side note, this is one of those things that drive boil my blood about doing science in Greece. That and the inflated prices we have to pay. For example the Covaris microTube kite has an 80% mark up over the price from Covaris. I’m not saying anything bad about Covaris here. For every product we buy it is the same when we have to go though distributors.)
Anyway, it seems that maybe I should repeat it with some even longer fixation times to get closer to replicated the level of fixation I see with the stock 37% formaldehyde which contains methanol.
Hamid
Indeed a great observation. Thanks for the same, however my concern is that(what I wrote to Ethan last time), even if the sonication is good with Methanol free HCHO, does it mean the final ChIP will be fine (ofcourse one need to actually do these experiments and comment), but just to see what’s going on you guys minds .Is there any risk of underfixation
Ethan, sorry to hear that your experiments got stalled due to kits, did you ever tried using Chelex beads (Nature Protocols 1, 179 – 185 (2006) )?
Hamid, I’m pretty sure you are wrong about the methanol or at least under then conditions I am using. But it is also kind of difficult to imagine how 0.4% methanol final concentration would make a big difference in x-linking efficiency. I’ll make a post about this later today with some gel images. There is a huge difference between x-linking in media and x-linking in what I call Covaris Fixing Buffer. X-linking in media seems to be way more efficient making fragmentation much more difficult.
Hi Ethan,
I have carried out an experiment in which I did the following:
1. used the fresh 16% formaldehyde to make the 11% stock solution.
2. used the fresh 16% formaldehyde to make the 11% stock solution, and added sufficient methanol to make the methanol concentration the in the 11% stock 10%.
3. used the 37% formaldehyde that contains 15% methanol to make the 11% stock solution.
The cells were then fixed according to our protocol, and the nuclei isolated. The samples were then sheared with a time course of 2-12 minutes. The data clearly shows that methanol had an adverse effect on the shearing. Samples from #1 sheared perfectly, and you can see the gradual reduction in the size range and distribution. Samples treated in #2 and #3 did not show the same efficient shearing. There is a clear over fixation in #2 and #3. I am certain that with longer processing time, even #2 and 3, would shear to the desired size range, but I am certain the the IP efficiency, and recovery will be greatly reduced. That is an experiment we are in the process of carrying out for the paper.
Thank you
hamid
Hi Ethan
1)I saw your Covaris protocol, will you share how many cells you begin with per IP? how many cells you use for cross linking and in what volume of fixing buffer?
2)The sonication profile that you showed , these from ? cells.
Thanks
Subhradip
Hi Subhardip,
1) I begin with one 15 cm plate of 95% confluent HeLa cells per ml of shearing buffer. It also works nearly the same for MEFs and mouse ES cells although in these cells we are using less material per ml of shearing buffer, just because they are a little more difficult and expensive to grow.
2) The sonication profile is from HeLa cells, but it look nearly identical for MEFs and ES cells.
Hamid,
It must be something different in our protocols. I was looking at longer fixation times to get conditions for proteins that do not efficiently crosslink to the DNA with 5 minutes of fixation time. Anyway, check out the experiment I just posted.
Ethan
I have trouble in understanding why the ChIp WB that you posted showed loss of epitope with increase in sonication time? Given the fact that fixation time is same (which is main contributing factor for loss of epitope) and also given that covaris is extremely good in maintaining the temperature(so we can rule out heat mediated loss of epitope with increasing sonication time ).
Correct me if am missing something basic.
Proteins like DNA are long chains, o.k. most are folded into blobs, but they are still big. Furthermore, in the context of crosslinked chromatin, where you break a bunch of proteins crosslinked together, the break point is not necessarily going go by separating individual proteins. Bottom line, proteins get sheared just like the DNA. If you look at bigger proteins by western blot you can see what looks like degradation products as the protein gets sheared into smaller fragments. This is a big reason why fragmentation of chromatin by micrococcal nuclease can work much better then mechanical shearing. But the other side is micrococcal nuclease leaves all these giant crosslinked macromolecular structures in tact, which can cause background when trying to ChIP some proteins.
Hi Ethan,
Thanks for the helpful info on ChIP. Would you have any insight into using the bioruptor vs covaris for ChIP? How would the uniformity of shearing between those two instruments compare?