Handle fragments bigger than 5 Kb is tricky. I would advice to do an additional partial  digestion  with Sau3A. Fix all parameter and control time of digestion to render fragment from 3-5 Kb. Since cuttings are at random you will have overlapping fragments.  After elution, use a procedure to assemble them. Take a look to the link below for the assembling procedure. Good luck!

http://www.ncbi.nlm.nih.gov/pubmed/21601685

Hi Alexandro,

Yes I have been trying to digest the genome to various sizes to reduce the size for pulldown. I am not looking for 100% efficiency, but better than 10% would be nice.

What do you think about this paper? They claim they can do up to 9.4kb fragments, but there is no data shown. http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0017353

My real question is what prohibits either streptavidin/biotin or Protein-G-antibiotin/biotin from efficient binding to large DNA...steric hindrance?

Hi Kenneth and Alexandro,

I did a lot of experiments with TdT-labeled DNA fragments, and from my experience, it's the labeling that makes the difference, because we could immunoprecipitate very large DNA fragments (genomic DNA upto 100kb or plasmid DNA) with either anti-biotin antibodies or streptavidin coupled to Dynabeads. There are two essential elements. The enzyme: we have tried many commercially available TdT enzymes, and Roche's is the best by far as it is a recombinant TdT that was selected because of its very high processivity. The nucleotide mix: dATP is the preferred nucleotide of the TdT, so by creating a mix of dATP and Biotin-dATP, you have the perfect mix to create a long biotynilated tailing as the "cold" dATP allows the TdT to add more Biotin-dATP. It's certain that it will be difficult to IP large DNA fragments with only one biotin as "the hook", but several is another story. However, Alexandro makes a good point, reassembling the strands after elution is a great challenge depending on your upstream applications!

It may work, but the efficiency is very low. Linear 10 Kb and beyond occupy a lot of space. It depend on average size of the dynabeads, by you may have with that size strong steric interference. But if the needed efficiency is not to high, just try.  I took a look to the paper and what you meant is refereed as "Data not shown". Honestly, I would like to see it in a gel. If you take a look in the Southern blot, there is almost nothing upper 6-7 Kb, indicating a drastic drop in efficiency when approaching to 10 kb. If you have time, you can try. Good luck.   

It's not a drop of efficiency, it's because the genomic DNA was cut by enzymes! Otherwise, we were immunoprecipitating the whole genome!

Hi Frederic and Alexandro,

I appreciate your answers. I honestly tried to immunoprecipitate the whole genome using undigested gDNA a few times, and my pulldown appeared very weak. I am using a long PCR assay to test pulldown of a specific target and actually find my input amplifying better than my pulldown. So my hypothesis was that the genome probably needed to be digested first. I used a restriction enzyme strategy much like Frederic's paper, but using enzymes to generate a 6, 12, and 18kb fragment that I tried to pulldown. I found the same result using my long PCR compared to input. I agree that maybe the TdT labeling requires longer tails in order to improve the capture efficiency.

I will try this. My next experiment I will just take HINDIII and undigested lambda DNA and see if I can pulldown after using a few TdT labeling conditions.

I will let you know how it goes!

Hi Frederic and Alexandro,

 So here I took lambda genome and tailed it directly with TdT and also tailed HINDIII digests of lambda genome. I used the same described conditions as the paper, but with 200 U of TdT (even this seemed like over-excess!)I then captured with goat anti-biotin using the exact same procedure as the paper including all the wash buffers. In parallel I also used streptavidin-Dynabeads and washed the same number of times as the IP, but with a different buffer.

I ran a quick gel and included 30% and 50% of input into the capture. The capture is very inefficient, but I only used 75ng of input labeled lambda genome (I wanted lower input on purpose). One can see that the streptavidin-Dynabeads looked like they worked, while IP did not. It also looks like digestion may help a little. 

I am wondering if there is something wrong with my elution from the beads. I heated to 90 C as the paper did...I am concerned that denaturation of the immunocomplex could cause issues with the release of the fragments?...or possibly impede the migration of captured fragments on the gel...

There is also a concern about the type of DNA dye you used. You may have a lot of single strand DNA that will not show up on a gel if you are using a double-strand specific dye! By the way, I'm pretty sure you didn't label you DNA enough as the fragments are quite sharp. When extensively labeled, it doesn't migrate as evenly because of the addition of single-strand DNA.

The capture is very poor. I'm afraid that the labeling is not that good. One thing, if you have some cobalt salt in the lab, you can add it to the buffer from 2-5 mM. This may enhance the activity of TdT. By the other hand, for analytic purpose, I suggest to concentrate the elution using one Speed-Vac or similar and put all in the gel. Because if the input is too low, when you elute your sample, the concentration drop under 5 ng/ul, you are in the limits. If you can use SYBR Gold dye, it is much better, it binds to both single and double strand DNA. Additionally, I remain skeptic about the capacity of this system to capture efficiently DNA over 8 Kb. Maybe with some modifications may work. But lambda/HindIII has a lot of fragments that you should capture. Let see if you get it. Good luck.  

Thanks for your comments. I may order the Roche TdT. I may also try to design an assay to test for optimization of tailing. I should have run unlabeled lambda side by side to see if there is any difference in smearing of the bands. I do typically use Sybr in my gels, but I wasn't expecting the capture to be this bad. I dried down the entire yield by SpeedVac and loaded all of it on this gel. Albeit the capture is very poor, why did the IP not capture anything? Maybe best to use Streptavidin-Dynabeads.

Hi,

I would just like to make some comments around the Dynabeads. The Dynabeads kilobaseBINDER Kit (Prod. No. 601.01) is designed for the binding of large (greater than 2 kb) DNA or RNA fragments. The kit contains Dynabeads M-280 Streptavidin plus a unique and Binding Solution. With this Binding Solution, it is possible to immobilise 70 pmoles of a 4 kb DNA fragment and 80 pmoles of a 10 kb fragment per mg Dynabeads. This enables various applications to be conducted, including pure template isolation for gene expression, protein-DNA interaction studies, mRNA processing studies etc. 

kind regards

ketil

[email protected]

Hi Ketil,

  Yes I found this the other day. I will be ordering this to try it. Thanks so much for bringing this to my attention!

its all my pleasure. good luck.

Hi Kenneth,,

Did this method work out? If yes, do you have a protocol/publication for this, or anyone else? I also want to digest genomic DNA, but with sonication. I think that the average DNA length will be 0.05-1 kb.

Hi Arthur,

  I never got anything to work well with IP for large fragments, so  I went back to basic capture with Streptavidin Dynabeads with Kilobasebinder, which seems a little better. With an average size of 1kb I would think it would be easier to IP. Please let me know if you try. I would start with following the TdT end label methods in the paper that was referenced earlier in this discussion, although I am uncertain of your goals.