Interphase 1: Why have a background control? Why silica enrichment after XRNAX?
A common misconception about XRNAX is that it is a procedure that only leaves behind the protein-crosslinked RNA from UV-crosslinked cells. This oversimplification can be drawn from the fact that XRNAX extracts approx. 100 times more protein/RNA from UV-crosslinked cells in comparison to non-crosslinked cells. While this clearly suggests that XRNAX extracts contain protein-crosslinked RNA it does not exclude the possibility that they also contain free protein. Importantly, they do, so that in reality XRNAX is the entry point to downstream applications which clean up protein or RNA covalently crosslinked to each other. XRNAX eliminates all components of UV-crosslinked cells other than protein-crosslinked RNA plus some free protein or RNA non-covalently associated with the protein-crosslinked RNA. This is a consequence of the initial TRIZOL extraction, which collapses protein-crosslinked RNA into the interphase. Along with covalently associated protein and RNA, non-covalently associated protein and RNA are dragged into the interphase. Evidence for this comes from several observations, first of which is that UV-crosslinking changes the interphase. Interphases from non-crosslinked cells are fluid-like a and can be easily disintegrated and even pipetted. Interphases from UV-crosslinked cells contain one sticky blob in addition to the fluid-like regular interphase contents. This blob and its fluid-like surrounding are not clearly distinct but seem more like a sponge, which soaks up the fluid it sits in. Consequently, free protein and protein-crosslinked RNA are not clearly separated in the interphase to begin with. How much free protein is retained in this way we can quantify using SILAC and a non-crosslinked control. As you can see in Figure S1D, after XRNAX the amount of non-crosslinked protein is not negligible.
An important consequence of this observation is that a meaningful control for an XRNAX experiment should be added to the sample before the TRIZOL extraction. SILAC quantification tells us that about 40 % of protein in XRNAX extracts is non-crosslinked. XRNAX on non-crosslinked cells yields only about 1 % the amount of protein produced from UV-crosslinked cells, so that treating the sample and the non-crosslinked control separately will underestimate the background by a factor of 40. Combining the UV-crosslinked sample and the non-crosslinked control before the initial TRIZOL extraction of XRNAX is only possible if SILAC is used, therefore our recommendation. Also, this line of evidence demonstrates pretty clearly why silica enrichment needs to follow XRNAX if you want to quantify. If 40 % of protein is not covalently attached to RNA, quantifying RNA-binding between conditions becomes rather meaningless. As displayed in Figure S11D, running your XRNAX extract over a silica column will remove most of the free protein, what makes meaningful quantification possible.