I've been working on trying to get a ChIP assay to work on a protein that isn't known to be a transcription factor, but that our lab suspects acts as one in a certain disease state. My project is to induce nuclear localization of my protein of interest and then try to ChIP it and ultimately do sequencing to get an idea of where this protein is binding promoters in the genome. I've been getting low yields of DNA from my ChIPs and have been assuming failure of the protocol, but I've read many papers suggesting that ChIP DNA yield is rarely high, especially for transcription factors. The most recent ChIP assay I did gave me a concentration of 0.645 ng/uL in 50uL , so about 32ng yield, assessed by Qubit. I'm wondering whether this information alone tells me that my protocol hasn't worked, or whether I might want to consider sending the DNA out for library preparation (my university has a genomics core that will do this). I would naturally like to do qpcr, but don't have any targets to search for enrichment. Any advice would be appreciated; our lab is molecular biology, and only now incorporating -omics.
Hi Ryan, when I performed ChIP-seq I obtained DNA yields close to yours. Then libraries were ok. I'm working in a mass spec group and I'd like to suggest to you to further investigate if your protein is enriched upon DNA degradation (by treating your sample with Benzonase you should observe an increase of the signal arising from your protein by WB). Then for sure you should validate your antibody before doing ChIP.
Hi Ryan.
Getting 32ng ChIP DNA is really a good yield to ahead for ChIP sequencing. Since you don't have any known targets to look for enrichment and success of your ChIP; there was couple of ways to validate it. One is to determine the concentration of DNA in your "No antibody" control which is used while performing the experiment. If you get lesser yield, it might give you some hint towards your ChIP success. Another way which I have done and has worked with me is to do one ChIP with a protein whose targets are known and perform your required ChIP in parallel. You can check for your control ChIP enrichment and take the point that your ChIP is working if the control ChIP works. It might be wastage of some reagents but it guarantees that your ChIP is working.
Hope it helps!
If you want to make sure that you ChIP worked technically try using positive control IPs . histone H3 or H3k4me3 usually work very well. Since there is loads of H3J4me3 preset at most TSSs you could design primers a giant a promoter of a housekeeping gene and if you have a good signal you will know that at least the protocol is working well (Ct values should be at least 18-23). If you can't ChIP H3 or H3k4me3 there is very little hope for anything that is less abundant. if the H3K4me3 ChIP fails I would then double check the chromatin fragmentation to make sure you are not over sonicating/over digesting your DNA. As far at the ChIP for your favourite TF goes I would recommend you check its expression in the cells you use and try to use te antibody to see if you can IP the TF out of a concentrated cell lysate. There is one more factor that can affect the DNA yield and that is X-linking. I use 1% formaldehyde for 9-10 min depending on cells I am using And these conditions give me good results even for the lowly expressed TFs. I hope this helps. Good luck !
Hi Ryan,
I agree with the above advice - although getting 32ng DNA pulled down compared to say 10ng for an IgG doesn't always guarantee specificity in the 32ng pulled-down DNA. One thing we do is xlink cells and make chromatin as usual. But instead of pulling-down for ChIP, we do an IP-western with the soluble chromatin. Use a different species antibody for the blot probing step. If the IP-western works, at the very least, you can say that your ChIP antibody has the ability to interact specifically with an epitope on your antigen (whichever chromatin protein you are looking at) in the harsh context of PFA xlinked chromatin. Thus, any DNA pulled down should give you more confidence of its validity. All the best, Donncha.
Hi, Donncha and Vikas. Thank you for your replies! I didn't mention in my initial description, but I have done 'ChIP's' where at the last step I elute the beads in Laemmli buffer and blot with a mouse antibody (my IP Ab is rabbit). I get very nice signal for my protein, and no signal for the IgG IP's. Would this be more indicative of success? I'm planning another experiment and my PI basically just said to try the library preparation, but if proper IP is a positive sign, then I'll be more confident to send the samples to the genomics core.
Hi Ryan, Well you beat me to it, in reference to the western. That sounds promising. I'd go for it, with a repeated IP useful for reproducibility. Good luck.
Hi Ryan,
I have the same problem as you have. We don't know the targets of our protein of interest. We have done just as Carlo Yague and Donncha Dunican said. We checked by WB the IPed material and by Qubit comparing input, negative control (IgG antibody) and ChIP material. We have a tagged version of the protein so we also used the no-tag control. We made a % of IPed material vs input and considered more than 0.05% of input material was OK for our protein (it is not an abundant transcription factor).
Your experiment looks promising! I will encouraged you to send your ChIP to sequence and once you have the targets check the sites by qPCR.
Good luck!
Zaida.
Hi Ryan. I guess with all these positive signs, you can go ahead with your library preparation and you should be able get your hits.
Best wishes,
Vikas
Considering all the discussions already in the chat I also feel that your experiment looks definitely promising. I would proceed preparing libraries and move on getting the sequencing done. If you don't have any other previous information about where your protein could bind, there would be no other experiment that could give you a sure response. Obviously once you have the targets you will have the opportunity to validate the results by checking the sites by qPCR.
Good Luck!
Carmelo
If there are no known targets of your protein and you want to check if your protocol is working at all, you could also take along antibody for protein with known targets and see if results are reproduced. Just remember that different anitbodies can still behave differently.
Anyway, good luck with your sequencing!
I believe you need to check the IP to start with. So you will know if the protein binds to any DNA. Any protein predictions to help? maybe bioinformatics can be a tool as well....
As an update: I have used a previous publication from our lab that led my work into ChIP to identify a plausible positive control locus for ChIP-qPCR. I have done the experiment a few times, and it seems as though I'm seeing enrichment of the locus of interest, but I'm not 100% sure what I should be expecting in terms of the degree of enrichment. If anyone could take the time to look over the data from a representative experiment, it would be greatly appreciated.
Hi Ryan,
sorry for the delay answer. I use a different method for analysis (absolute quantification) so I had to look your results in detail. The ChIP sample should be at least 0.01% of the input sample, if the protein of interest is not abundant, or at least 0.1% of the input for high expressed TF or common epigentic signatures. The no-antibody sample should be less than 0.01% of the input sample. in your case is 0.12 and 0.008 respectively, so congratulations!!
I just have a question, why did you analyze two different input samples? I use the same input sample for the ChIP and the no-antibody samples, because both samples came from the same isolated chromatin.
Good luck with your sequencing and happy holidays.
Zaida.
hello Ryan Coleman , can yo tell me that how did you managed to know the binding site of your hypothetical protein because you even dont know whether it is a TF or not?
My second question is how you performed ChIP-PCR because there are a large no of papers illustrating about ChIP-Seq and ChIP-CHIP.
- Badges
- Science method