Rational Design of Small Molecules Targeting Oncogenic Noncoding RNAs from Sequence

Abstract

The discovery of RNA catalysis in the 1980s and the dissemination of the human genome sequence at the start of this century inspired investigations of the regulatory roles of noncoding RNAs in biology. In fact, the Encyclopedia of DNA Elements (ENCODE) project has shown that only 1-2% of the human genome encodes protein, yet 75% is transcribed into RNA. Functional studies both preceding and following the ENCODE project have shown that these noncoding RNAs have important roles in regulating gene expression, developmental timing, and other critical functions. RNA's diverse roles are often a consequence of the various folds that it adopts. The single-stranded nature of the biopolymer enables it to adopt intramolecular folds with noncanonical pairings to lower its free energy. These folds can be scaffolds to bind proteins or to form frameworks to interact with other RNAs. Not surprisingly, dysregulation of certain noncoding RNAs has been shown to be causative of disease. Given this as the background, it is easy to see why it would be useful to develop methods that target RNA and manipulate its biology in rational and predictable ways. The antisense approach has afforded strategies to target RNAs via Watson-Crick base pairing and has typically focused on targeting partially unstructured regions of RNA. Small molecule strategies to target RNA would be desirable not only because compounds could be lead optimized via medicinal chemistry but also because structured regions within an RNA of interest could be targeted to directly interfere with RNA folds that contribute to disease. Additionally, small molecules have historically been the most successful drug candidates. Until recently, the ability to design small molecules that target non-ribosomal RNAs has been elusive, creating the perception that they are "undruggable". In this Account, approaches to demystify targeting RNA with small molecules are described. Rather than bulk screening for compounds that bind to singular targets, which is the purview of the pharmaceutical industry and academic institutions with high throughput screening facilities, we focus on methods that allow for the rational design of small molecules toward biological RNAs. One enabling and foundational technology that has been developed is two-dimensional combinatorial screening (2DCS), a library-versus-library selection approach that allows the identification of the RNA motif binding preferences of small molecules from millions of combinations. A landscape map of the 2DCS-defined and annotated RNA motif-small molecule interactions is then placed into Inforna, a computational tool that allows one to mine these interactions against an RNA of interest or an entire transcriptome. Indeed, this approach has been enabled by tools to annotate RNA structure from sequence, an invaluable asset to the RNA community and this work, and has allowed for the rational identification of "druggable" RNAs in a target agnostic fashion.

Figure 1

RNA is a single-stranded biopolymer that adopts various conformations beyond those solely described by Watson–Crick base pairing. For example, RNA can form noncanonically paired internal loops, bulges, and hairpin loops. One major advantage that RNA has over other biopolymers is that information about its structure can be deduced from its sequence by using computational approaches with or without experimental constraints. Targeting these noncanonical structures can be accomplished with small molecules, which is the focus of this Account.

Figure 2

Developing methods to identify the RNA motifs that bind small molecules avidly, namely 2-dimensional combinatorial screening (2DCS). 2DCS is a library-vs-library method that probes both small molecules and RNA motif space simultaneously, and the resulting data comprise a database of annotated RNA motif–small molecule binding partners. Briefly, a small molecule library is spatially arrayed onto microarray surfaces and probed for binding to a library of RNA motifs embedded in a unimolecular hairpin cassette. Small molecules on the array capture members of the RNA motif library that they bind. Binders are excised from the microarray surface and sequenced.

Figure 3

Inforna approach to design small molecules targeting RNA from sequence. This approach is target agnostic as it computes the most optimal binders to RNA noncanonical motifs from all human miRNA hairpin precursors. Lead compounds are then tested for bioactivity and for specific modulation of the biological functions of an RNA.

Figure 4

Small molecules that target the miR-96 hairpin precursor affect downstream cellular processes. (A) Lead optimization of the initial compound afforded a dimeric small molecule that targets both miR-96’s Drosha site and an adjacent 1 × 1 nucleotide GG internal loop, affording Targaprimir-96. (B) Targaprimir-96 inhibits biogenesis of miR-96 in vivo and impedes tumor growth as measured by photon flux (the number of photons per second per unit area). (C) Targaprimir-96 inhibits miR-96 biogenesis and (D) derepresses FOXO1 in vivo.