This study has equally interesting findings as it's back story!
Oct 30th, 2006 -- 9:06pm
I was returning late from hockey practice and this meant my PCR had run more cycles than I planned (30 instead of preferred 25). I quickly ran a gel and saw only one bad -- the one band that would change my scientific career.
It was the first time a specific RNA was observed to bind to the epigenetic juggernaut -- Poly comb repressive subunit 2 (PRC2). That sounds complicated but in simple terms PRC2 is the key factor that is required to turn genes off so cells can change their fate (e.g. from a stem cell to a liver cell).
The long-standing conundrum has been that PRC2 is present in all the cells in our body, yet needs to turn off very specific & different genes in each cell setting. So how does it know where to go in which cell ? RNA may now provide a new clue.
We weren't the only ones with evidence pointing to a role for RNA. Bernstein and colleagues also found that polycomb needed RNA in general for it's localization using independent approaches. Thus, in 2006 the new clue emerged -- RNA may influence how PRC2 can find it's proper locations in different cells.
For over a decade no less than 502 studies have been published on PRC2-RNA interactions. In fact, every possible model of how RNA could interact with and influence PRC2 had been proposed. In some cases RNA guided PRC2 to it's location and in others RNA prevented PRC2 from binding -- polar opposite effects.
Worse many studies proposed RNA had nothing to do with PRC2. So which is it? It's so complicated.
Enter a more simple approach to the problem. We first had to tackle the key question: Does RNA "bridge" PRC2 to DNA at all? The standard way to see where PRC2 is bound on DNA is called Chromatin-Immunoprecipitation or ChIP. This works by freezing the cells in time with a strong cross-linking reagent -- or like a freeze frame in a movie with everyone frozen in their current place.
We reasoned if RNA was important at all --- than once PRC2 was frozen in place and we then removed all the RNA (pulling the rug out from under PRC2) it would fall off the DNA and we would no longer retrieve the DNA and see these binding sites removed. If PRC2 stays put then RNA could not be "bridging" PRC2 in its current location. I hate to mention this but I had been suggesting this experiment for over a decade. Two very brave postdocs took on this challenge: Taeyoung Hwang and Yicheng Long from the Rinn and Cech labs respectively (neighboring in space).
It was only months after we moved from Harvard to Boulder that Taeyoung and Yicheng were already far along in understanding PRC2-RNA interactions. As a shot in the dark I suggested adding RNAse (which eats up all the cells RNA) and see if it pulls the rug out from under PRC2 being on DNA. They did it and it worked! When the RNA was removed so was PRC2. But what we really needed was a control -- could the RNAse just do something artificial to the assay -- intuitively no, but we needed to do more.
So we chose direct DNA binding proteins: the classic Tata Binding Protein that makes numerous direct interactions with DNA and could not feasibly have an "RNA bridge". Same for Polymerase II that makes DNA into RNA via direct interaction with DNA that could not be "bridged by RNA". They jumped right on it and incredibly the RNAse had no affect on TBP or POL II -- if anything the replicates may have been cleaner.
-- We could now conclude that RNA is required for PRC2 to maintain it's position on DNA --
The Cech lab had been working for years on finding the exact spot RNA binds to PRC2. Yicheng had found the elusive sites in a previous study. He found that if he mutated only two amino-acids (protein building blocks) on PRC2 it could no longer bind RNA (well not nearly as well). The Cech lab had taken the bold step of using CRISPR-CAS9 genome editing approaches to make this mutant version in human induced pluripotent cells (iPSC) -- or pretty much similar to human stem cells.
I have always said "genetics is to biology as math is to physics" it's like a proof. If the lack of this mutant to bind RNA causes a cell phenotype we can conclude that the RNA is also functional. Thus, Yicheng's approach is heroic to take the ultimate step in testing RNA-PRC2 interactions are physiologically relevant.
To this end Yicheng engineered the mutant line where PRC2 was completely normal in all activities (enzymatic, complete complex assembly etc) -- except it doesn't bind RNA very well. The stem cells were fine and acting normal, that is actually expected as it is know that removing PRC2 from stem cells doesn't affect the pluripotent state -- only when the stem cells differentiate into other cell fates. So Yicheng took the hIPSCs and differentiated them into cardiomyocytes (probably because it so cool to seem them beating like a heart in the dish). However, what was observed is that RNA defective PRC2 lines could not make heart cells let alone have them beat.
Yicheng took it one step further rationalizing that if he reintroduced the normal PRC2 that can bind RNA -- would it fix this lack of ability to make heart cells? The answer was yes.
Thus, we could now conclude that PRC2-RNA interactions are physiologically relevant.
Moving forward these two simple yet clear conclusions will hopefully simplify and focus this fundamental question of how epigenetic complexes get to where they need to go. The whole thing got a lot simpler IMHO.
On a personal note this was a very special collaboration. Our laboratory is hosted in multiple labs with the idea it would draw natural collaboration -- or "innovation through integration". It worked, Taeyoung and Yicheng met as postdocs with out borders encouraged to collaborate across labs -- I think they may even be good friends now -- well they at least are starting to have similar interests :) Moreover, seeing them working with Karen Goodrich and Anne Gooding is one of those special things that makes the hardship of science a little lighter.
Of note Yicheng is on the job market to start his own lab (Taeyoung soon to follow) -- catch them while you can !