Protein factors can regulate gene expression by binding to specifically modified DNA-associated proteins. Small molecules that selectively interfere with such interaction may be of therapeutic value. See Article p.1067 & Letter p.1119
Protein factors are crucial for controlling gene expression. One group of such factors affects gene activity by 'reading' epigenetic marks — reversible modifications such as the addition of phosphate, acetyl or methyl groups — on proteins after their translation from RNA. The factors' target proteins are histones, which associate with DNA to form chromatin. The reading ability of these protein factors is a result of specific, well-folded subdomains, sometimes called readers, which can distinguish between the post-translationally modified state of their binding partner and its unmodified state. Two papers1,2 in this issue describe highly potent and selective inhibitor molecules that compete with acetylated histones for binding to a set of such readers. These data have therapeutic implications.
Post-translational modifications can often influence transient protein–protein interactions by creating or disrupting binding surfaces on the molecules. Among such modifications, acetylation on lysine amino-acid residues has been centre stage: originally discovered3 more than 40 years ago as a regulator of chromatin structure, this modification has now been detected in thousands of other proteins4. Acetyl-lysine modifications facilitate the interaction of the protein with proteins that contain bromodomains — evolutionarily conserved subdomains that can specifically bind, or read, the acetylated form of the lysine during regulatory processes5. Such interactions are thought to regulate transcription and to be involved in various diseases, including cancer.
A range of acetyltransferase enzymes (writers) add acetyl groups to lysine residues, and two families of deacetylase enzymes (erasers) remove these groups6. Two recently approved anticancer drugs7 — SAHA and depsipeptide — work by blocking deacetylases, and have galvanized the pharmaceutical industry's interest in targeting chromatin modifications. In fact, several start-up biotech companies have attempted to target erasers and writers of lysine acetylation. In general, however, even highly specific inhibitors of acetyltransferases and deacetylases that mediate post-translational modification can have undesired side effects, because blocking these enzymes can affect many different protein substrates and biochemical pathways.
As for targeting protein–protein interactions, with several notable exceptions the use of small-molecule drugs has been considered extremely difficult8 because their binding regions frequently consist of wide, shallow surfaces. For example, despite decades of work, pharmacologically practical compounds that disrupt the binding of phosphorylated proteins to their SH2-domain-containing protein partners have remained elusive. There have also been a couple of attempts to use small molecules to inhibit the interactions between proteins containing bromodomains and those carrying acetyl-lysines, but focus on this line of research has generally been limited9.
Using very different approaches, Filippakopoulos et al. (page 1067) and Nicodeme et al. (page 1119) now converge on a closely related set of chemical scaffolds — the triazole-diazepine-fused ring compounds JQ1 and I-BET — that inhibit the acetyl-lysine-reading ability of a specific class of bromodomain. Both sets of compounds bind tightly to bromodomains in proteins of the BET family by exploiting the unusual pockets characteristic of this protein family (Fig. 1).
The bromodomains of BET proteins show a strong preference for housing doubly modified acetyl-lysine histone tails in their wide and highly structured hydrophobic pockets10. Because of their shape and electrical properties, these pockets are also well suited for binding small molecules. Indeed, the present papers' structural data1,2 confirm that JQ1 and I-BET fit snugly into the acetyl-lysine pockets in a stereospecific fashion. Thermodynamic measurements further establish that both of the bromodomain–inhibitor interactions are of high affinity (with dissociation constants below 100 nM) and, compared with their interaction with other non-BET types of bromodomain, show great selectivity (at least 100-fold).
The two teams also pursue distinct biomedical applications for JQ-1 and I-BET. Filippakopoulos et al.1 examine whether JQ1 can antagonize the growth of a rare but aggressive form of cancer called midline carcinoma. This cancer is defined by a gene fusion that results in the unnatural linkage of BRD4 — a BET protein containing two bromodomains — with another protein called NUT. The BRD4–NUT fusion protein mediates increased acetylation of certain chromatin domains that are normally transcriptionally inactive11, and so it was predicted that inhibitors of the BRD4 bromodomains would shut down tumour growth mediated by this mechanism. Filippakopoulos and colleagues confirm this hypothesis, showing that JQ1 could blunt the growth of midline carcinoma cells in culture, as well as in mice into which the tumour cells were introduced.
Nicodeme et al.2 investigate whether I-BET modulates genes mediating immunological and inflammatory responses. They find that it inhibits the expression of a subset of genes normally induced in response to toxic injury, with histone acetylation being reduced in the chromatin regions around these genes. In a practical application of these findings, the authors demonstrate that treating mice with I-BET protects against the excessive inflammatory response to septic shock. Such results point to the clinical potential of BET-bromodomain inhibitors in immuno-modulation therapies.
The two papers1,2 provide credibility for the idea of extending the pharmacology of targeting chromatin modifications beyond enzymatic activities and into the challenging arena of protein–protein interactions. One appeal of this strategy is that it avoids the promiscuity of enzyme inhibitors.
The studies further raise the prospect of identifying inhibitors of other readers, such as those that bind proteins containing methyl-lysine modifications. Nonetheless, antagonizing a reader of post-translational modifications might also prompt unknown and unwanted alterations in biological pathways — a complication that necessitates extensive follow-up studies before such agents can move into the clinic. Moreover, the possible uniqueness of BET-bromodomain structures makes it difficult to predict whether small-molecule inhibitors would be similarly effective in antagonizing other bromodomain forms and reader modules. Nevertheless, the new tools described by these studies will undoubtedly prove attractive to biologists interested in the dynamics of chromatin and gene expression in physiology and disease.
Filippakopoulos, P. et al. Nature 468, 1067–1073 (2010).
Nicodeme, E. et al. Nature 468, 1119–1123 (2010).
Gershey, E. L., Vidali, G. & Allfrey, V. G. J. Biol. Chem. 243, 5018–5022 (1968).
Choudhary, C. et al. Science 325, 834–840 (2009).
Taverna, S. D., Li, H., Ruthenburg, A. J., Allis, C. D. & Patel, D. J. Nature Struct. Mol. Biol. 14, 1025–1040 (2007).
Cole, P. A. Nature Chem. Biol. 4, 590–597 (2008).
Lemoine, M. & Younes, A. Discov. Med. 10, 462–470 (2010).
Wells, J. A. & McClendon, C. L. Nature 450, 1001–1009 (2007).
Sachchidanand et al. Chem. Biol. 13, 81–90 (2006).
Morinière, J. et al. Nature 461, 664–668 (2009).
Reynoird, N. et al. EMBO J. 29, 2943–2952 (2010).
About this article
Scientific Reports (2020)
PLOS Biology (2015)
Molecular & Cellular Proteomics (2014)
A PWWP Domain-Containing Protein Targets the NuA3 Acetyltransferase Complex via Histone H3 Lysine 36 trimethylation to Coordinate Transcriptional Elongation at Coding Regions
Molecular & Cellular Proteomics (2014)