ChIP against transcription factors - (Dec/13/2004 )
Hi everybody! Anyone out there doing ChIPs with abs against transcription factors? I've done hundreds of ChIPs using abs against acetylated, methylated, ubiquitinated, you-name-it'ed, histones and they work great, but I'm having troubles getting the technique to work for RNApol II CTD (Abcam ab), PPARg or CBP (both Santa Cruz). These are abs which are suppossed to work, at least in the papers published by other people, but they never do in my hands. I've used both the Upstate protocol and the Abcam protocol; I usually work with 2-3 million cells/IP, and my sonicated DNA has sizes below 2 kb. Histone abs always perform great, in all conditions I use. I'll appreciate your suggestions!
For all ChIPs I think the impt. thing is to have a positive and neg. control.
I've used CBP from Santa Cruz and POlII CTD, and phos. ser.5 PolII CTD.
I have great results with them (paper is submitted so I can't give you a ref).
Check to see if someone have used thes antibodies in the cells you are using and get those primers. As a general rule, 5 microliters of antibody from upstate,abcam or convance and 10 microliters for santa cruz antibodies.
For example, if you add hormone (estrogen, ret.acid etc) there are published genes you can look at for Pol II, CBP etc. THese results work in many cell lines.
That's about all I can offer.
I am also doing ChIP with antibodies to methylated, acetylated histones and have some questions for you.
I have found that loss of DNA during washs after antibody precipitation is a big concern and eventually I could not get an equal signal even for duplicates of samples. What measures do you take to minimize DNA loss and how do you control for that. I know I can use control gene such as beta-actin for acetylated histones, but for transcriptionally silenced genes such as by histone methylation, what gene can be used as control.
I think input DNA can only be used to control for equal amount of starting DNA. Do you agree?
Thank you in advance.
Hi again! Thanks for your suggestions, mikew. I think I might not be using enough antibody. I use 2 ul of the Abcam and Upstate histone abs and they work, but maybe for TF is better to use higher volumes, as in a short stretch of DNA there are a lot of histones but maybe only 1-2 molecules of RNAPolII or CBP. I'll try that. Thanx! Also, I agree that controls are very important. I use beta-actin as a positive control, as it is highly acetylated and methylated, I would expect to see binding of RNAPolII to it but I don´t see it. As negative controls I use tissue-specific genes that are not expressed in the tissue or cell line I'm working with at the moment (insulin, albumin, myoD...).
Hi, green. As for your questions, I've also done duplicate precipitations and got different strengths of signal. I've tried to use the Abcam protocol which skips the LiCl and TE washes, and I've got stronger signals, but noisier. I believe that the loss of DNA is related to the size of the sonicated fragments. When for some reason I got bad sonications (>5kb) my PCR signals are lower than usual, but when I manage to get better than usual sonications (<1kb) I got high signals, very reproducible and clean.
I agree that input DNA only indicates the amount of STARTING material. What I usually do to account for precipitation efficiency (which may vary from tube to tube, even when using the same ab and the same sample) is analyze my IP'ed samples by duplex PCR. I co-amplify my gene of interest with a positive control (usually beta-actin) which is not suppossed to change. I calculate the amount of problem/control in the sample as a percentage of the problem/control ratio in the input DNA. I usually try to get all my primers to work in duplex with primers for beta-actin so that they more or less amplify at the same efficiency (I use several dilutions of the input DNA to make sure of that).
Hope it helps.
thank you Nice-cell (not bad). Your reply helps a lot.
If you are studying a gene which is inactivated for example by histone methylation, then it will be hard to find a gene that is similarily silenced as control.
Hi, green! Mmm, I'm not really familiar with the silencing histone modifications. The only time I did H3-MethK9 IPs I was working with primary hepatocytes from mouse and as a positive control I used some gene from the chr.X, I seem to remember it was MAGE-A2. In those experiments I always used female mice, 'cause one of their chr.X is inactivated by histone methylation, of course. I remember I got a faint signal for my control but not for any of my problem genes, so I never used that ab again. Now I'm working with cell lines, and I'm not sure which genes would I use as a control for silencing methylation marks. Maybe I will try again this trick of choosing a gene from the chr.X, or an imprinted gene such as igf2. Or maybe a highly tissue-specific gene which is not expressed in my cell line (insulin? albumin?). Humm, I'm not sure. Does anybody out there have any suggestions?
was just wondering if the ChIP against transcription factors is working?..... i have the same problem.
I am using the upstate protocol, and i use about 5microgrm of antibody. I get great results for H3 but hardly any cycle difference for POLII. I am using the Bcells.
Just a thought, but is it possible that, while your crosslinking regime is sufficient for histones, which are tightly bound to the DNA, it's possibly not enough for TFs? Have you tried increasing your crosslinking time? We have done ChIP for TBP and CBP and while we don't get as spectacular a result as for modified histones (optimal PCR cycles from these ChIPs are much higher) we do still get results we would expect for our positive control primers. We use a formaldehyde concentration of 1.42% and a crosslinking time of 15min at room temp.