Transcription through Polycomb response elements does not induce a switch from repression to activation.

Polycomb group proteins (PcG) are a group of transcriptional repressors that silence gene transcription through covalent modification of histone proteins and compaction of chromatin at target genes. In metazoans, PcG proteins are present in all cells of the body and at all stages of development, but they repress target genes only in cells where these genes should remain inactive. Work over the past decades led to the general view that PcG proteins do not initiate repression but primarily function to preserve the transcriptionally inactive state of a target gene in situations where transcriptional activators of this gene are present in a cell but the gene nevertheless needs to remain inactive. Classic examples of this type of regulation are the long-term repression of Hox genes in inappropriate body segments during Drosophila development. Repression of a target gene by PcG proteins does, however, not always mean terminal silencing of the gene in a given cell; in many instances repression needs to be reversed, even after prolonged periods of silencing. In PNAS, Erokhin et al. (1) address how PcGregulated transgenes respond to the presence of a transcriptional activator. Understanding how a PcG-repressed state is converted into a transcriptionally active state is not only important for understanding how genes are regulated during normal development but also for understanding the events that occur when PcG proteins malfunction in cancer. In their experiments, Erokhin et al. (1) analyzed the consequences of bringing the transcriptional activator Gal4 to a transgene construct in which PcG proteins, bound to a Polycomb response element (PRE), normally repress transcription from linked reporter genes. The 660-bp bxd PRE used in this study is a well-characterized PRE from the HOX gene Ubx (2–4). A simplified version of the transgene and the key conclusions from their studies are shown in Fig. 1. Erokhin et al. (1) found that continuous Gal4 expression is able to disrupt transcriptional repression mediated by the PRE, and they demonstrate that this process does not require transcription through the PRE. Moreover, the authors show that transcription through the PRE in the transgene does not cause displacement of PcG proteins. This study thus refutes a popular model that posits that transcription through a PRE would cause PcG proteins to be displaced and would be a key step in “switching” a PRE into a cis-regulatory element with enhancer-like properties (Fig. 1). To understand the relevance of the work by Fig. 1. Transcription through a PRE does not remove PcG proteins or lead to a heritable switch in gene expression. (Left) Results predicted by the transcription switch model: (A) In the absence of the Gal4 activator protein, PcG proteins at the PRE silence the expression of the white gene leading to a light eye color. (C, Upper) Transcription through the PRE displaces the PcG proteins and activates white gene expression leading to a red eye color; (Lower) repression is not relieved because the PRE is not transcribed. (E ) Expressing Gal4 for 1 h during early development switches the PRE from a repressive to an activating element, and white gene expression is maintained throughout development. (Right) Summary of the results obtained by Erokhin et al. (1): (B) white is silenced by the PRE. (D, Upper) Transcription through the PRE does not displace the PcG proteins (although the levels are reduced) but the white gene is activated; (Lower) presence of Gal4 permits expression of the white gene. (F ) Transient expression of Gal4 early in development does not lead to white gene expression later in development. *Because PcG proteins are maintained on the PRE in the case of continuous Gal4 activation, Erokhin et al. (1) did not test whether PcG proteins remain at the PRE after a pulse of Gal4. For simplicity, the hsp26-LacZ and the GFP reporter genes used as additional controls in the construct (1) are omitted from the drawing here. Author contributions: J.A.K. and J.M. wrote the paper.