negative control region issues - (Apr/09/2010 )

Hey guys,

Just trying to optimize my chip using a Pol II Ab and primers to GAPDH (postive control) and various other loci for negative controls. I have gotten good results in the past using GAPDH and GLUT4 (not expressed in liver) but since I prepped a new batch of chromatin, I am seeing very similar % input at both positive and negative loci. IgG is not a problem and is always non-detectable so I do have a good signal.

I have noticed then when I add more chromatin to the IP, my Ct's actually INCREASE instead of decreasing as you might expect (by as much as 2 Ct); even the input Ct's increase slightly but not as much as my Pol II samples. My % input therefore decreases and still ends up being similar between postive (GAPDH) and negative (GLUT4, Rhodopsin) regions.

My question is do you guys think I am using WAY too much chromatin for each IP and might this be interferring with the IP? I have good sonication, probably an average of about 500 bp.

It seems to me that increasing your amount of DNA in your IP could indeed increase your background (negative control region) over your signal (positive control), as you may have saturated your positive control signal and are just bumping up your background. For what it's worth, I typically use 2 ug of DNA (as measured using a Nanodrop on samples that end up being run on a gel to check for shearing). As I tell the undergrads I work with when critiquing their writing....less is more!

Not sure how to explain the increase in Ct (did you run the different chromatin amounts on the same plate for qPCR? If not then I think directly comparing Ct's is not useful).

MM

@Mighty Mouse:

What are you ChIP'ing for, out of interest? I'm using a H3ac antibody, and use around 4 ug of chromatin for each ChIP. I did notice an increase in background if I used more (7 ug), but I haven't tried to use less than 4 ug.

It's my understanding that increased amounts of input chromatin in the ChIP will also cause inhibition by the increased amounts of SDS in the reaction?

Does the amount of input affect the overall DNA yield from the ChIP (i.e. how many ng are available for sequencing)? I.e. if I want to reduce background by decreasing input chromatin, will this also decrease the amount of DNA I get out of the ChIP, and necessitate doing more ChIPs to get enough for sequencing?

jamessmith01 on May 26 2010, 08:26 AM said:

If you want to increase your yield without decreasing your signal to background ratio you should increase the amount of antibody you use proportionally with the increase in chromatin input. If you are using agarose beads it's unlikely that you'd need to increase the amount that you use. Since the capacity is typically much higher than you need to bind all your antibody, you would just be increasing your background when increasing the amount of beads. If you're using magnetic beads, though, you should check the capacity and make sure you can still bind all the antibody.

Hi James,

So I'm ChIP'ing for both TFs and histone modifications. I find ChIP'ing for histone modifications to be much more robust than TFs. So far I've ChIP'ed for H3Ac and CREB and am working on pCREB and thyroid receptor and will probably tackle H4Ac soon. If you decreased your input by 1/2 you would likely decrease your IP signal as well, so if you're on the edge of where the data starts to get noisy on qPCR you might not want to decrease your input. This is of course assuming that you have not saturated your antibody-protein binding at 4ug, if you have then you could probably decrease your DNA no problem. I guess you can only really determine that empirically.

MM

2ug DNA???!!
The senior student in my lab suggests me starts with 200ug per sample!!
By using 2ug, how many cts of input/ChIP product do you usually get?
I ChIPed with 50ug and got 25~28ct for Input
I cannot imagine that what's the product amount by using 2ug Input.
I'm doing other transcription factors recently and I got 32~35 for ChIP product only.
Even not sure if that's real signal or what........(and also similar IgG)
Is that I use too many DNA????
I'm so shocked right now...... /__\

Chairfly on Tue Aug 17 17:56:45 2010 said:

I think it all depends on how you are quantifying the DNA content of your chromatin. If you're using A260 the the value you get are not realistic. There are so many things in crosslinked chromatin sample that will impinge on that wavelength that you can't measure DNA accurately.

A good rule of thumb for doing your first experiments is to use an amount of chromatin which is equivalent to from somewhere between 0.5 to 2x10^6 cells.

Chairfly on Tue Aug 17 17:56:45 2010 said:

2ug is what others have used doing similar work to what I do; I use dissected mouse brain as my substrate as opposed to cell culture; not sure if that makes any difference. When I run my input samples I actually run 1% input (so 0.02 ug) and get Ct values between 26-29 for my input samples and if I get enrichment of my target, Cts around 29-32. Usually if I get good enrichment my data comes out pretty clean overall. I suppose I could use more chromatin, but the brain region I dissect isn't the largest so I can get a lot more ChIP's out of one set of dissections by using only 2ug of DNA.

KDPE, to your point, I don't use cross-linked chromatin for my A260 quantification. I remove 5 uL of my sample following sonication, then I proteinase K treat, reverse crosslinks and purify using spin columns (consequently I also use this chromatin to check shearing). My 260/280 and 260/230 values for my DNA purification typically look very good (between 1.5-2), so I think I'm getting an accurate assessment of my DNA concentration; if you can think of some reason that I wouldn't get a good assessment using this method I would certainly like to know...

MM

Mighty Mouse on Wed Aug 18 00:01:39 2010 said:

Sorry, my comment wasn't directed at your method. I've recently spoken to a number of new ChIPpers who have tried quantitating their sheared chromatin directly.

KPDE on Wed Aug 18 01:02:39 2010 said:

No worries...have they come to any consensus? Seems to work pretty well for me. Sometimes I'll get to the lower threshold of what the Nanodrop can determine, but for the most part I believe I'm in a fairly robust range (~10-20 ng/uL).