hello,

The cross-linking leads to a drop of solubility. Lysis buffers alone are not sufficient to recover proteins and chromatin. You have to sonicate samples. It is worth playing both with cross-linking and sonication parameters in order to find a rigth balance.  You can check https://www.diagenode.com/p/universal-plant-chip-seq-kit-x24-24-rxns

It looks like your protein cross-linking might cause some protein oligomerisations and increase of its molecular size, which might cause the increase the MW of your protein linked (with/to to some other protein(s). To solve this problem you might want to use a lower concentrations of the cross-linking coupling reagent or just drop this step procedure at all from you protocol. The sonication will unlikely help you with the solubilization of the coupled proteins. If you want to purify your protein in the native condition you need to avoid any chemical reactions like cross-linking or some other chemical changes/manipulations of t's natural environments where this protein might not feel happy. 

You should quench with 750 mM Tris instead of glycine. This will give you more control over the cross-linking duration, as you are probably not completely quenching the PFA. Test different cross-linking times. Check out the paper "Utility of formaldehyde cross-linking and mass spectrometry in the study of protein-protein interactions" from 2008.