Hi Anil - I can't access the Liu paper as it is pay-walled. It may be that they just treated with Nocodazole  and isolated G2/M cells using a FACs sorter and performed Histone work and RNA-SEQ on these cells. This seems likely as in the abstract they mention that they looked at G0/G1, G1/S, and M phases of the cycle (which would require sorting I think).

If you have access to their methods it might explain in there how they obtained these cells/data. Certainly Noc treatment won't destroy the plasma membrane so the DNA (in whatever state) should remain in the cells (probably at the mitotic plate).

Sorry I can't be more help!

Gary

Hi Gary,

Thanks for sharing your thoughts. The authors performed double thymidine synchronization for G1/S transition, and released the cells into nocodazole. After 12 hrs, they collected the rounded mitotic cells by shake-off--a standard procedure that usually yields >95% pure mitotic cells. So, they did not do FACS to get the mitotic cells. They then isolated nuclei from these mitotic cells to perform GRO-seq. According to the guidelines described by John Lis in his Science paper describing GRO-seq (https://www.ncbi.nlm.nih.gov/pubmed/19056941) the starting material is purified nuclei, on which run-on transcription is conducted to generate and purify nascent transcripts. Interestingly, John Lis reviewed this manuscript for PNAS.

In my opinion, nuclei would not survive in mitotic cells obtained after 12 hrs of nocodazole. I am tempted to think that, perhaps, MCF-7 cells are not as well synchronized by nocodazole as are other common cells like HeLa. If so, the cell-isolation procedure may have gotten a big population of cycling cells, and it is on the nuclei from these cycling cells (and not from arrested mitotic cells) that the authors likely performed GRO-seq.

The problem is, they did verify the purity of their nocodazole-treated cells by Hoechst-A FACS that showed 82% G2M cells (Fig S1D). I suspect the majority of these cells were "somehow" G2 or pre-M cells that did not enter mitosis very well, and thus retained nuclei. Or am I wrong somewhere?

Something else that is puzzling is, the authors show pro-transcriptional H3K4me2 and H3K27ac levels around the TSS at a much higher level compared to G0/G1 or G1/S. I will have difficulty in believing this data unless some other labs independently show similar patterns.

I guess the paper is accessible from Pubmed, with no institutional subscription, in US, but maybe it is not available in UK.

Thanks again.

Anil