If you have access to a BluePippin you could do size selection for a range of your sample product below a specific cutoff - removing the HMW fragments. This is often how we proceed following sonication to remove products outside our desired size range.

Hi Nuria,

Normally, increasing the sonication time with a BioRuptor Plus (UCD-300) should get rid of the large fragments. But this relies on a good cell (and thus tissue) lysis. Moreover, the BioRuptor Plus has quite some strict sample constraints:

- Long sonication may be necessary: we generally use 40 - 60 cycles of 30" ON / 30" OFF.

- The hardness of the tubes used makes a very big difference. The TPX tubes sold by Diagenode work well, whereas normal Eppendorf tubes work much less.

- The BioRuptor is very sensititive to the volume in the tube. We use 250 ul in 1.5 mL tubes. Larger volumes dramatically reduce efficiency. See page 20 for volumes in different tubes: https://www.diagenode.com/files/products/shearing_technology/bioruptor/Bioruptor_manual.pdf

If the above mentioned things don't help, it may be that your tissue preparations contain solid pieces that are refractive to lysis and sonication. Before sonication, does your material look homogeneous? Or do you still have big lumps of cells?

Incubation of your tissue with collagenase (45', 37oC) and/or cell disruption with a Polytron Homogenyzer may make the downstream cell and nuclear lysis more efficient.

Finally, if this all doesn't help, it may be worth it to use a cell strainer (https://www.fishersci.com/shop/products/falcon-cell-strainers-4/p-48680) prior to cell lysis to filter out any non single-cell debris.

Good luck,

Daan

Hi Nuria,

My name is Irina Panteleeva. I am senior scientist at diagenode in charge of sample prep projects. I fully support the above recommendations from Daan.

I am wondering how did you visualzed the shaaring: agarose gel or BioAnalyzer? BA trends to over-estimate HMW fragments in chromatin samples.

You can concat me directly at [email protected] for troubleshooting.

Sonication does four things to different extents, and you want to tune each variable to the extent possible (but not all are independent parameters):

1.) It coarse shears large objects like cells/nuclei

2.) It does fine shearing of smaller objects like oligonucleosomes

3.) It shears epitope from nucleosomes

4.) It can denture nucleosomes, exposing internal epitopes that are not available in the native state for antibodies that were raised against peptide haptens (H3.3, K79me2, etc.), depending on the detergent used and the amount of energy delivered.

The degree to which each of these things happens is variable and highly dependent on cross-linking conditions, cell types, and cell-density during shearing.  Tip sonication and Bioruptor are pretty good at [1]; not so good at [2]-- the distribution range from sub-nucleosomal to oligonucleosomal fragments will be large, unless you way over do it which directly impacts [3], and [4] is variable in all methods of sonication based sharing ([3 & 4] are opaque to the experimenter, unless you have defined nucleosome calibrants).  Covaris is really good at controlling fine DNA fragmentation and can give much tighter distributions around monos [2] (still takes some fiddling with conditions) while minimizing [3], and does less [4] due to not as much heating and energy being dumped into your sample.  We also use 0.5 mm glass beads to make Covaris better at [1]. So perhaps seek out a Covaris?

New Bioruptors are nice, but they degrade in their energy delivery over time as do all bath solicitors (essentially a sanitation element glued to a metal bath). Covaris sonication is more portable.

Perhaps back off on your cross-linking conditions (although yours look like they are on the lighter side)-- perhaps you are not quenching the residual formaldehyde at the end? We use equimolar tris base to quench FA (glycine is far less efficient).

We actually do native ChIP for histone mods and variants (way better signal to noise) but still do X-ChIP for factors.

Size your library with Ampure or Serapure to reduce large fragments during library prep. Worst case scenario, you can do paired-end sequencing and computationally filter out larger fragments. That said, there are strong, and un-apreciated avidity biases to ChIP-seq so bigger fragments are often strongly favored in the IP, unless the epitome density in chromatin is sparse. (see: Grzybowski, Chen, Ruthenburg, Mol Cell 2015).