FMRP Directly Interacts With R-Loop and Shows Complex Interplay With the Dhx9 Helicase

Abstract

Fragile X Syndrome (FXS) occurs when mutations in the FMR1 gene cause the absence or dysfunction of FMRP, known mainly as a translation repressor. We recently showed that FXS cells suffer genome-wide DNA double-strand breaks near R-loops under replication stress. The expression of FMRP, and not an FMRP-I304N mutant of the K-homology 2 RNA-binding domain, suppresses the R-loop-induced DNA breakage. These observations led us to hypothesize that FMRP safeguards the genome through promotion of R-loop detection and/or resolution. Here, we demonstrate that FMRP directly binds R-loops through multivalent interactions between the carboxy-terminal intrinsically disordered region and the R-loop sub-structures. We also show that the amino-terminal folded domain of FMRP directly binds DHX9, an R-loop resolvase, in a KH2-dependent manner. The FMRP-DHX9 interaction is recapitulated by co-immunoprecipitation in human cells. Our findings are consistent with a model in which FMRP recruits DHX9 to R-loop forming sites by bridging their interaction through its amino- and carboxy-termini, respectively.

A Tale of Loops and Tails: The Role of Intrinsically Disordered Protein Regions in R-Loop Recognition and Phase Separation

“…We systematically tested the ability of FMRP full-length, N-Fold and C-IDR to bind nucleic acid structures including ssDNA, dsDNA, RNA, DNA:RNA hybrid, and R-loops with and without RNA overhang (Figure 3B). Indeed, we observed direct binding between FMRP full-length and R-loops and, of all the tested protein-nucleic acid pairs, the C-IDR and the R-loop substrate without overhang produced the highest affinity (K D 4.73 ± 3.83 nM) (Chakraborty et al, 2021). However, the interaction was significantly weakened when a 5' RNA overhang was present in the R-loop (K D 148.3 ± 10.03 nM), suggesting that the C-IDR may interact with the triple junction where the trailing RNA emerges.…”

“…Finally, we subsequently showed that FMRP coimmunoprecipitates with DHX9 in vivo and directly binds the methylated Arginine residues in the RGG region of DHX9 in vitro, using the two Agenet domains within the N-Fold (Chakraborty et al, 2021). DHX9 is a multifunctional ATP-dependent nucleic acid helicase that unwinds various DNA and RNA substrate structures, including R-loops and G-quadruplexes (Chakraborty and Grosse, 2011).…”

“…In fact, there are numerous instances of IDRs of proteins playing critical roles in the biological functions of many proteins due to their inherent ability to form flexible linkers between folded domains, for being the predominant sites for post translational modifications (PTMs) and for serving as sites for direct protein and nucleic acid binding as well as for being the dominant drivers of LLPS for many known phase separating proteins (Van Der Lee et al, 2014;Uversky, 2017). In this manuscript, we investigate the amino acid composition and properties of the proteins in the R-loop interactome and explore the hypothesis that, in some R-loop binding proteins, the IDRs provide the dominant site of interaction with the R-loop structure as we recently demonstrated for FMRP (Chakraborty et al, 2021).…”

“…While these studies missed some key R-loop proteins, they recovered most proteins traditionally known to be involved with R-loop regulation, such as topoisomerases [e.g., showing the relative strength of binding between FMRP constructs (top row) and various nucleic acid substrate controls (left column), such as R-loops with (out) RNA overhang, DNA bubble, dsDNA, DNA:RNA hybrid, ssDNA, and ssRNA. Binding affinity for each FMRP construct and nucleic acid substrates was calculated as dissociation constants (K D s) averaged from at least two Electrophoretic Mobility Shift Assay (EMSA) experiments (Chakraborty et al, 2021). K D values are scaled according to a log scale: (++++) means 1-10 nM; (+++) means 11-100 nM; (++) means 0.1-1 μM; (+) means 1-10 μM; and (−) means no observed interaction under the binding conditions used for the EMSA assay (i.e., 1 nM of γ-P (Muramatsu et al, 2000) R-loop, RNA-DNA hybrid, dsDNA, bubble DNA, ssDNA, or RNA substrate mixed with protein at various concentrations in a buffer composed of 25 mM Tris-HCl (pH 7.5), 100 mM KCl, 5 μg/ml BSA, and 5 mM EDTA.…”

“…While studying the impact on genome stability due to the absence of FMRP in Fragile X patient-derived (FX) cells, we discovered that FX cells undergo R-loop-mediated genome-wide DNA double-strand breaks (DSBs) under aphidicolin-induced DNA replication stress (Chakraborty et al, 2020). We subsequently demonstrated that FMRP directly interacts with R-loops, predominantly via its C-IDR, with the three N-terminal folded RNA binding KH domains providing additional weak contacts through binding to the various R-loop substructures (Chakraborty et al, 2021). This surprising and exciting finding prompted us to investigate the role of IDRs in the other R-loop interacting proteins, especially in proteins that lack the canonical R-loop processing enzyme domains such as helicases or nucleases.…”