Four papers released online today detail some of the work from David Allis's group and others that's detailed in our recent article on chromatin remodeling. In two Nature papers released today, Allis, a Rockefeller chromatin researcher, along with postdocs Joanna Wysocka and Tomek Swigut and a team led by structural biologist Dinshaw Patel, from Memorial Sloan Kettering, report on BPTF, the largest subunit of the nucleosome remodeling factor (NURF). It contains a so-called PHD finger which they've now shown preferentially binds histone 3 trimethylated at lysine 4. They found the protein in pull-down assays and it appears to maintain activity at developmentally critical HOX genes.
This work is notable for adding support to the controversial histone code hypothesis. The structural work reveals an alpha helix between the PHD finger and a bromodomain. Bromodomain motifs are well known to recognize acetylated lysines, so that unstructured alpha helix might be acting...
I saw Allis give a very engaging talk at Thomas Jefferson University last Thursday, and he was hinting that the "multiple tails" option was looking pretty good. But just in case the relationship between PHD fingers and gene activity sounded straightforward enough, the two other Nature papers released at the same time (and presumably being published in the same print issue) complicate issues, revealing that the PHD domain of tumor suppressor ING2 also recognizes trimethylated H3K4. This is puzzling because ING2 is closely associated with the mSin3a-HDAC1 histone deacetylase, a complex that represses gene transcription.
Or Gozani at Stanford, Tatiana Kutateladze at University of Colorado Health Science Center and their respective collaborating groups published two papers describing the function and structure of ING2 and its ability to block gene transcription after DNA damage. This is the first protein to link H3K4 methylation to transcriptional repression, and it does it through the same recognition motif (the PHD finger) that Allis's papers describe. They write, "Together, our findings highlight the notion that the recognition of chromatin modifications by effector proteins rather than specific modification per se, determines biological function, and, as such, greatly expands the diversity of signaling at chromatin."
Allis told me he thought it was fascinating that the same mark (H3K4 trimethylated) might signal both expression and repression. Some researchers argue that if the same signal gives different, even opposite readouts , then there can't be a predictive system. But the context of other signals around H3K4 methylation may play a role in this and other situations.