A Conversation with Susan Gasser.

Dr. Gasser: We’ve been studying the mechanisms of heterochromatin repression, of different ways to keep genes silent. Most of that is at the level of transcription through the formation of heterochromatin, but in a screen for derepression of heterochromatin we found three subunits of an RNA-controlling complex. All the other hits in this screen were chromatin modulators as we expected, so these three were a surprise. This was in C. elegans, but this complex is conserved from bacteria to yeast to humans. It’s called the LSM complex for “like Sm” proteins. We read up on the LSm complex, and found that it comes in two forms. There’s one in the cytoplasm and another in the nucleus. They share six subunits, but the seventh, either subunit 1 or 8, is specific for either cytoplasm or nucleus. The first thing we showed was that the role in silencing was through the nuclear complex. That was good, because it meant that it was probably working at the level of genes. We then looked at the proteins that are supposedly interacting with this nuclear complex. Its normal role is to bind U6 RNA, catalyze or chaperone splicing, and then help trigger the degradation of the spliced-out intron, through an RNA exonuclease called XRN2. XRN2, but not U6 RNA, was also involved in the heterochromatic silencing of our reporter. Then we asked, “Which endogenous genes are sensitive to this RNA degradation mode of silencing?”We detected several hundred genes that were derepressed—or up-regulated—in the absence of this LSM complex or of XRN2. We asked, “What’s the nature of these genes?” First, we saw that they were all very poorly expressed in wild-type worms. Second, we checked their chromatin state by monitoring enrichment of histone marks, and we found that 95% carried histone H3K27me3, the characteristic methylation deposited by Polycomb. Polycomb is known to silence genes. It’s usually thought to create facultative heterochromatin in a tissueor cell type–specific manner. It actually poises genes in an “off” state, but such that they can also be switched “on,” depending on differentiation. It’s quite intriguing that all the genes that were controlled by this LSM2–8 were Polycomb-marked facultative heterochromatin. So we were faced with the question, “How does it work?” Polycomb normally represses by transcriptional repression, by blocking transcription, but here we were looking at RNA degradation. We checked specific genes that were sensitive to this RNA-degrading complex and showed that if you mutate the LSM8 subunit, which is specific for the XRN2-binding nuclear complex, then you stabilize a low level of transcripts from such target genes. Apparently, although genes are repressed by Polycomb, there’s a low level of promiscuous transcription possible. Then we asked if the same genes are regulated by XRN2, LSM8, and Polycomb. And indeed, they overlap significantly. Finally, we asked the question, “Does RNA degradation feed back in any way to the Polycomb mark?” Does the LSM complex and RNA degradation stabilize the transcriptional repression that’s mediated by Polycomb?” And indeed it does; we see that the K27 methylation mark is reduced on target genes, but not lost overall, in the lsm-8 mutant. The LSM8 complex constitutes a backup system that ensures the integrity of the Polycomb repression pathway. This is rather a noncanonical idea, and the Polycomb field seems not very keen on having Polycomb silence by means of something other than the attenuation of Pol II [DNA polymerase II] activity. So it’s been an interesting ride to try to publish this.