Integrative Omics Indicate FMRP Sequesters mRNA From Translation and Deadenylation In Human Neuronal Cells

Tatsuaki Kurosaki; Shuhei Mitsutomi; Alexander Hewko; Nobuyoshi Akimitsu; Lynne E Maquat

doi:10.1016/j.molcel.2022.10.018

. Author manuscript; available in PMC: 2023 Dec 1.

Published in final edited form as: Mol Cell. 2022 Nov 9;82(23):4564–4581.e11. doi: 10.1016/j.molcel.2022.10.018

Integrative Omics Indicate FMRP Sequesters mRNA From Translation and Deadenylation In Human Neuronal Cells

Tatsuaki Kurosaki ^1,^2,^*, Shuhei Mitsutomi ^1,^2,³, Alexander Hewko ^1,², Nobuyoshi Akimitsu ³, Lynne E Maquat ^1,^2,^4,^*

PMCID: PMC9753132 NIHMSID: NIHMS1849220 PMID: 36356584

SUMMARY

How Fragile X Syndrome Protein (FMRP) binds mRNAs and regulates mRNA metabolism remains unclear. Our previous work using human neuronal cells focused on mRNAs targeted for nonsense-mediated mRNA decay (NMD), which we showed are generally bound by FMRP and destabilized upon FMRP loss. Here, we identify >400 high-confidence FMRP-bound mRNAs, only ~35% of which are NMD targets. Integrative transcriptomics together with SILAC–LC-MS/MS reveal that FMRP loss generally results in mRNA destabilization and more protein produced per FMRP target. We use our established RIP-seq technology to show that FMRP footprints are independent of protein-coding potential, target GC-rich and structured sequences, and are densest in 5'UTRs. Regardless of where within an mRNA FMRP binds, we find that FMRP protects mRNAs from deadenylation and directly binds the cytoplasmic poly(A)-binding protein. Our results reveal how FMRP sequesters polyadenylated mRNAs into stabilized and translationally repressed complexes, whose regulation is critical for neurogenesis and synaptic plasticity.

Keywords: fragile X syndrome protein, human neuroblastoma cells, FMRP RIP-seq footprinting, SILAC, LC-MS/MS, mRNA translation, mRNA decay, poly(A)-binding protein, deadenylation, translationally silenced and stabilized FMRP–mRNA complexes

Graphical Abstract

graphic file with name nihms-1849220-f0001.jpg

eTOC Blurb

Deficiency in fragile X syndrome protein, FMRP, is the leading single-gene cause of intellectual disability and autism. Kurosaki et al. integrate transcriptomic and proteomic approaches to define FMRP-binding sites within neuronal mRNAs and show how concomitant binding of FMRP to mRNA bodies and poly(A) tail-bound PABPC1 inhibits their translation and deadenylation.

INTRODUCTION

Fragile X syndrome (FXS) is the most common single-gene cause of inherited intellectual disability and autism spectrum disorder (ASD) (Richter and Zhao, 2021; Hagerman et al., 2017). The disease-causing mutations consist of CGG-trinucleotide expansions within the 5'UTR of the fragile X messenger ribonucleoprotein 1 gene, FMR1. The pathogenesis of FXS is attributable to the loss of FMR1 expression, which encodes the fragile X syndrome protein (FMRP). FMRP functions in neurogenesis, synaptic plasticity, cognition, and reproduction.

At the molecular level, FMRP is an RNA-binding protein best known as a translational repressor (Richter and Zhao, 2021; Hagerman et al., 2017). Consistent with translational repressor function, we recently reported for human SH-SY5Y neuroblastoma cells in which FMRP was knocked down that the decay rate of mRNAs targeted for nonsense-mediated mRNA decay (NMD), which depends on translation, is hyperactivated (Kurosaki et al., 2021a). One mechanism by which FMRP is recruited to and/or stabilized on NMD targets is mediated by interacting with the central NMD factor UPF1. Consequently, when FMRP is deficient, UPF1 can no longer promote and/or stabilize FMRP binding to NMD targets, which are thus relieved from FMRP-mediated translational repression, augmenting their degradation (Kurosaki et al., 2021a).

Missing from this work is identifying transcriptome-wide those human SH-SY5Y-cell mRNAs that are bound by FMRP. This point is crucial given that half-life studies demonstrated that FMRP depletion destabilizes a large number (n = 7,604) of SH-SY5Y-cell mRNAs, including most (n = 984/1277) of the defined NMD targets (Kurosaki et al., 2021a). Also missing is a clear understanding of how the loss of FMRP influences mRNA translation transcriptome-wide for mRNAs that are or are not normally bound by FMRP.

Considering confusion over what constitutes an FMRP-binding site (Anderson et al., 2016; Ascano et al., 2012; Brown et al., 2001; Darnell et al., 2001, 2005; Li et al., 2020; Vasilyev et al., 2015), it also becomes important to define FMRP-bound sequences. FMRP-binding sites have been identified throughout the coding sequence (CDS) of polysome-associated target mRNAs with no noted nucleotide-sequence specificity (Darnell et al., 2011) or with a preference for optimal codons (Shu et al., 2020). FMRP-binding sites have also been mapped to G-quadruplex sequences within the CDS (Darnell et al., 2001; Goering et al., 2020; Schaeffer et al., 2001) or to short specific sequences in the CDS and 3'UTR of target mRNAs (Ascano et al., 2012; Ray et al., 2013; Tran et al., 2019).

Here we use anti-FMRP RIP-seq footprinting (Kurosaki et al., 2014; 2018b) to gain insight into what defines an FMRP-binding site on SH-SY5Y-cell mRNAs. To detect FMRP footprints that include those assisted by UPF1 i.e., those that typify NMD targets and, in theory, may or may not involve FMRP binding directly to RNA (Kurosaki et al., 2021a), our method does not involve UV crosslinking. Thus, it avoids artificial crosslinking bias (Singh et al., 2014; Wheeler et al., 2018). Importantly, It also precludes assaying binding that occurs after cell lysis. We report that FMRP footprints map to structured GC-rich and G-quadruplex sequences within mRNA 5'UTRs, CDSs, and 3'UTRs, suggesting that FMRP need not recognize a translating ribosome or codon optimality as previously reported (Chen et al., 2014; Shu et al., 2020) but instead binds to sequences and/or structures per se. Moreover, normalizing FMRP binding to mRNA length reveals that FMRP footprints are most dense in 5'UTRs, most likely due to their GC-richness.

FMRP function in neuronal cells remains unclear. Although FMRP has been widely described as a translational repressor (Ascano et al., 2012; Darnell et al., 2011; Napoli et al., 2008; Shah et al., 2020), evidence exists that it activates the translation of some transcripts (Liao et al., 2008; Tang et al., 2015; Greenblatt and Spradling, 2018; Liu et al., 2018; Dionne and Corbin, 2021). To date, global assessments of how FMRP alters mRNA translation have largely relied on Ribo-seq, which is plagued by the inability to differentiate elongating from stalled ribosomes (Zhao et al., 2019), either with or without controlling for mRNA abundance or whether the mRNA binds FMRP.

Here, we performed stable isotope labeling of proteins by amino acids in cell culture (SILAC) using wild-type (WT) and FMR1-KO cells followed by protein quantitation using liquid chromatography-tandem mass spectrometry (LC-MS/MS). We simultaneously performed RNA-seq to normalize the level of individual proteins to the level of mRNA from which each protein derives. Results indicate that, in FMR1-KO cells relative to WT cells, the majority of mRNAs analyzed are (i) reduced in abundance, in agreement with Shu et al. (2020), but not based on the identity of optimal codons, in disagreement with Shu et al. (2020), which we show instead reflects GC-richness and structures, and (ii) generate more protein per mRNA whether or not they bind FMRP. That noted, in the absence of FMRP, most mRNAs that normally bind FMRP undergo the largest increase in both translation and destabilization. We focus on the 75% of SH-SY5Y mRNAs that are normally FMRP-bound and undergo enhanced translation and decay upon FMRP loss. Remarkably, FMRP-binding to solely the 5'UTR, CDS, or 3'UTR via structured GC-rich sequences in reporter mRNAs results in mRNA stabilization and decreased translation.

Unexpectedly, anti-FMRP IPs performed in the presence of RNase I using SH-SY5Y-cell lysates followed by LC-MS/MS identified the major cytoplasmic poly(A)-binding protein PABPC1, which we demonstrate binds FMRP directly. Quantitations of poly(A) abundance confirmed that efficient FMRP-binding to mRNAs requires a poly(A) tail and protects mRNAs from deadenylation. We discuss how FMRP in neuronal cells compartmentalizes mRNAs into stabilized and translationally repressed complexes until signaling activates their localized translation and decay.

RESULTS

FMRP-bound transcripts function in neurodevelopmental and signaling pathways in human neuronal cells

Studies of FMRP binding to transcripts either directly, via crosslinking technologies, and/or indirectly, via RNA IP (RIP) using a variety of cells or tissues – non-neuronal human cells (Ascano et al., 2012; Van Nostrand et al., 2016), mouse brain tissue (Brown et al., 2001; Darnell et al., 2011; Tabet et al., 2016; Maurin et al., 2018; Sawicka et al., 2019; Shu et al., 2020), adult human frontal cortex (Tran et al., 2019), in vivo-differentiated human pluripotent stem cell-derived neurons (Li et al., 2020), or in vitro-differentiated human forebrain organoids (Kang et al., 2021) – have reported that FMRP primarily binds within mRNA coding sequences (CDSs). However, defining those FMRP-bound mRNAs in human neuronal cells that contribute to neuronal development and synaptic signaling is in its infancy (Tran et al., 2019; Li et al., 2020; Kang et al., 2021).

We previously identified FMRP-bound NMD targets in human SH-SY5Y neuroblastoma cells using an antibody specific for FMRP (Figures S1A and SIB) and RIP (Kurosaki et al., 2021a). Given that UV preferentially generates crosslinks between pyrimidines (C or U) in single-stranded RNA and the side chains of aromatic amino acids (F, Y, or W) (Singh et al., 2014; Wheeler et al., 2018), we utilized RIP-seq instead of CLIP-seq for four reasons: crosslinking experiments may underrepresent FMRP binding (i) to G-rich RNA sequences, such as G-quadruplexes, due to the relatively weak UV-photoreactivity of proteins to purines (Singh et al., 2014; Wheeler et al., 2018); (ii) to structured RNA sequences, such as G-quadruplexes, by the FMRP RGG box (Darnell 2001; Vasilyev et al., 2015; Maurin et al., 2018; Goering et al., 2020), which lacks aromatic amino acids (Figure S1C); (iii) via the FMRP GXXG motif in its two KH domains (Musco et al., 1997; Ray et al., 2013; Myrick et al., 2015; Hu et al., 2015), which also lack aromatic amino acids (Figure S1C); and (iv) that is promoted and/or stabilized by other RNA-binding proteins, such as UPF1 (Kurosaki et al., 2021a), if binding does not involve direct contacts between FMRP and mRNA. As important controls, we performed lysate mixing experiments in which results analyzing cellular mRNAs confirmed that our RIP methodology detected FMRP binding prior to and not after cell lysis (Figures S1D-S1F). The specificity of our RIP methodology was further demonstrated by (i) the presence of MYC-UPF1 mRNA, which is an FMRP target (Kurosaki et al., 2021a), but not β-actin mRNA, which is not an FMRP target, after anti-FMRP RIP, and (ii) the absence of both mRNAs after mouse (m)IgG RIP (Figures S1D-S1F).

For anti-FMRP RIP-seq footprinting (Figure 1A), we generated ~20 million reads from each of four cDNA libraries – two control libraries, one deriving from Input, i.e., prior to IP, and the other deriving from mIgG IP samples; and two test libraries deriving from biological replicates of IPs that utilized the FMRP-specific antibody. After computationally removing sequences encoding rRNAs and repetitive elements, ~2 million reads in common to input samples and both anti-FMRP IP samples were mapped to a unique genomic sequence (All, i.e., 14,212 SH-SY5Y mRNAs in Figure S1G). Among these mRNAs, anti-FMRP RIP-seq enriched for transcripts that encode proteins associated with FXS-disease phenotype, including proteins from genes that are previously defined FMRP targets (Brown et al., 2001; Darnell et al., 2011; Ascano et al., 2012; Suhl et al., 2014; Figure 1B) and reported to be misregulated in autism (Figure 1C).

We next identified 436 high-confidence FMRP-bound mRNAs (hereafter denoted FMRP targets) and 723 high-confidence mRNAs that are not bound by FMRP (hereafter denoted Not FMRP targets) based on a stringent (log₂-fold change >1 and P < 0.05 or log₂-fold change <−1 and P < 0.05, respectively) enrichment of anti-FMRP RIP-seq footprint reads relative to both Input RNA-seq and control mIgG RIP-seq footprint reads (Figure S1G; Tables S1 and S2). These FMRP targets constituted ~3% (436/14,212) of SH-SY5Y mRNAs (Figure S1F), are significantly enriched in two previously reported FMRP CLIP-seq datasets (Van Nostrand et al., 2016; Li et al., 2020) (Figure S1H), and include previously defined FMRP-bound KIF4, MAPK8IP2, and CDC42BPG mRNAs, whose neurite localizations are impaired in FMR1-KO Cath.-a-differentiated cells (Goering et al., 2020). GO term enrichment for biological process (Figure 1D) and network analyses (Figure 1E) revealed that their encoded proteins are enriched for functions in many aspects of neuronal development and maturation.

RIP-seq-derived FMRP footprints (Figures 1F and S1I) were further verified using RT-qPCR analyses of specific mRNAs (Figures 1G and 1H), confirming FMRP enrichment on mRNAs misregulated in FXS and autism (Figures 1B and 1C). As examples, FMRP-enriched targets include mRNAs encoding GPRIN1, a G protein-regulated inducer of neurite outgrowth 1; mTOR, a protein kinase that controls protein synthesis, cell growth and proliferation; TSC2, a GTPase-activating protein that negatively regulates the mTOR complex 1; ANKRD11, which regulates nerve-cell proliferation and brain development; IRF2BPL, which functions in central nervous system development and neuronal-cell maintenance; and SHANK3, a constituent of the synaptic scaffold (Figures 1F and 1H). Consistent with our previous findings (Kurosaki et al., 2021a), FMRP was also enriched on neuronal NMD targets encoding, e.g., SHANK3, dual specificity phosphatase 3 (DUSP3), ribosomal protein S6 kinase B1 (RPS6KB1), and synaptic Ras GTPase activating protein 1 (SYNGAP1), but not GAPDH, whose mRNA is not an NMD target and does not bind FMRP (Figure 1H).

FMRP binding occurs throughout mRNA bodies, primarily at structured and GC-rich sequences without sensing protein-coding potential, and correlates with increased mRNA stabilization

In search of parameters within mRNAs that correlate with FMRP binding, we calculated the log₂ ratio of the number of read counts in anti-FMRP RIP-seq footprints for individual mRNAs relative to the number of read counts for that mRNA in Input samples. We call this value the FMRP-binding index (FBI). Notably, the FBI showed a significant positive correlation with the relative occupancy (%) of anti-FMRP RIP-seq peaks as defined using the CLIPper peak-calling algorithm (Figure S2A). We also calculated the FBI using previously reported FMRP CLIP-seq data deriving from human pluripotent stem cell (hPSC)-derived neurons (Li et al., 2020) and K562 human erythroleukemia cells (Van Nostrand et al., 2016). For all data sets, FBIs correlated positively with mRNA GC content, secondary structuredness, G-quadruplex (G4) density, and mRNA length (Figures 2A, S2B-S2J). Of these parameters, the correlation coefficient for FMRP enrichment was highest for mRNA GC-content (Figures 2A, S2B-S2D) and secondary structuredness (Figures 2A, S2B, S2C and S2E), including G-quadruplexes (Figures 2A, S2B, S2C and S2G), and weakest for mRNA length (Figures 2A, S2B, S2C and S2I). mRNA half-life changes determined for SH-SY5Y cells upon FMRP-knockdown (KD) relative to control-KD using pulse-chase TRIC-seq methodology (Kurosaki et al., 2021a) revealed that FBIs correlated with mRNA instability upon FMRP-KD (Figures 2A, S2B, S2C and S2J), i.e., while the half-lives of Not FMRP targets had approximately an equal chance of being lengthened (~52%) or shortened (~48%), the half-lives of most (~94% of) FMRP targets were significantly (P < 2.2 × 10⁻¹⁶) shortened upon FMRP-KD (Figures 2B-2D and Table S3). Taken together, these data indicate that FMRP generally binds to structured and GC-rich sequences, and its loss results in mRNA destabilization to a degree that reflects the FBI.

Figure 2. — (A) Bar plots showing Spearman’s correlation coefficients of the FMRP-binding index (FBI) (x-axis) and the specified SH-SY5Y-cell mRNA feature (y-axis). RSCU, relative synonymous codon usage; G4 density, G-quadruplex density; cAI, codon adaptation index; half-life change with FMRP siRNA (siFMRP), the latter derived from data in Kurosaki et al. (2021a).

(B) Heatmap of the mRNA half-life changes for 391 FMRP targets in the presence of two different siFMRPs, each normalized to control siRNA (siCtl).

(C) Stacked histogram representations of the percent (%) of mRNAs whose half-life increased (blue) or decreased (red) with siFMRP relative to siCtl treatment.

(D) Cumulative fraction of log₂ fold-change in measurable mRNA half-lives for All mRNAs, FMRP targets, or Not FMRP targets upon siFMRP relative to siCtl treatment, n, number of mRNAs; P-values were calculated using the two-sided Wilcoxon rank-sum test.

(E) For Input and after α-FMRP IP samples, stacked histogram representations showing the % of reads that mapped to each mRNA region as read number (upper two histograms), or read density i.e., read number after normalization to the length of each region (lower two histograms).

(F) Scatter plots of FBI (y-axis) vs. mRNA GC content (x-axis) for 5'UTRs, CDSs, and 3'UTRs. n, number of mRNAs; r, Spearman’s rank correlation coefficient; P, P-value.

(G) Histogram representations showing Spearman’s correlation coefficients between FBI (y-axis) and the 256 tetra-nucleotide densities in SH-SY5Y-cell mRNAs, which are arranged based on correlation coefficient strength (x-axis). High GC content (red); low GC content (blue).

FMRP has been reported to bind primarily to mRNA CDSs (Ascano et al., 2012; Darnell et al., 2011; Shu et al., 2020) in a way that correlates with codon adaptation index (cAI), which is a measure of codon optimality (Shu et al., 2020). We also observed these correlations when codon optimality was measured using either relative synonymous codon usage (RSCU) or cAI (Figures 2A, S2B, S2C, S2F and S2H), and when FMRP footprints were not normalized to the length of the mRNA region, i.e., 76% of FMRP footprints map to CDSs, with 7% mapping to 5′UTRs and 17% mapping to 3′UTRs (Figure 2E, Read count). However, SH-SY5Y CDSs generally constitute the largest percentage (62%) of mRNA sequences when compared to either 5′UTR (4%) or 3′UTR (34%) sequences (Figure 2E, Read count). This is consistent with data deriving from NCBI RefSeq gene datasets, which revealed that median lengths of human mRNA 5′UTRs, CDSs, and 3′UTRs are, respectively, 203 nts (8%), 1278 nts (53%), and 938 nts (39%) (Pivesan et al., 2016). After normalizing to length, we found that most (45%) of FMRP footprints resided within 5′UTRs, with 33% and 22% residing within, respectively, CDSs and 3′UTRs (Figure 2E, Read density), consistent with FBIs correlating with GC-richness (Figures 2A, 2F and S2B-S2D). FMRP density was also highest in 5′UTRs using RIP-seq peak densities and CLIPper peak-calling analysis for each mRNA region (Figures S2K and S2L). Our findings are consistent with meta-analyses of human GENCODE mRNA annotations, indicating that the hierarchy in GC-richness is 5′UTRs > CDSs > 3′UTRs (Figure S2M). Moreover, our mining results from FMRP CLIP studies (Van Nostrand et al., 2016; Li et al., 2020) likewise revealed that, per length, FMRP binds primarily to 5′UTRs (Figure S2N), which is also consistent with more recent data reported by Van Nostrand et al. (2020).

Consensus FMRP-binding motifs within mRNAs from human HEK293 cells, mouse brain, and hPSC-derived neurons were previously defined as ACUK, GACR and/or WGGA (Ascano et ak, 2012; Suhl et al., 2014; Anderson et al., 2016), where W is A or U, K is G or U, and R is A and G. To further evaluate FMRP binding at these and a total of 256 tetra-nucleotides, we used our anti-FMRP RIP-seq footprints to calculate nucleotide correlation coefficients based on four-nucleotide windows. We found that FMRP binding preferentially occurred at GC-rich sequences such as GGCC, GCCC and GGGC (Figure 2G and Table S4), but not at AU-rich sequences, including ACUK (Figure 2G).

It has also been reported that FMRP preferentially binds to GAC, UAU, UAC, UGG, and GGA codons in HEK293 cells and mouse brain (Anderson et al., 2016) and optimal codons such as CUG in cortical neurons dissected from embryonic mice (Shu et al., 2020). Thus, we used our SH-SY5Y-cell data to calculate the nucleotide correlation coefficient for FMRP binding to nucleotide triplets constituting in-frame codons in mRNA CDSs. As a control, we did likewise for mRNA 3′UTRs, continuing in the same frame. We found for both CDSs and 3′UTRs that FMRP binding occurred largely at GC-rich but not AU-rich trinucleotides (Figure S2O and S2P) regardless of reading frame (Figure S2Q and S2R), in keeping with the importance of GC-richness rather than protein-coding potential or ribosome binding to (or in the vicinity of) the trinucleotide (Figures 2A, S2B and S2C).

Globally defining changes in protein abundance upon loss of FMRP

Metabolic labeling of cellular proteins for a relatively short (1-hr) period demonstrated that, compared to WT mouse brain, Fmr1-KO mouse brain manifests a global increase in protein synthesis (Qin et al., 2005; Dölen et al., 2007). This finding, together with data showing that FMRP copurifies with translationally active polysomes (Brown et al., 2001; Darnell et al., 2011; Khandjian et al., 1996; Mazroui et al., 2003), was taken to indicate that FMRP is a translational repressor. However, quantitative proteomics comparing individual proteins in mouse-derived embryonic fibroblasts labeled for 14 days (Matic et al., 2014), cortical neurons labeled for 18 days (Liao et al., 2008), dorsal hippocampal slices labeled for 5 hr (Bowling et al., 2019), or cortical synaptosomes labeled > 45 days (Tang et al., 2015) revealed that not all proteins are increased in abundance in the absence of FMRP.

Comparable studies of human cells, either FMR1-KO or FXS patient-derived cells, have never been performed. Thus, we subjected WT SH-SY5Y cells and three independently generated FMR1-KO SH-SY5Y cell lines (Kurosaki et al., 2021a) to SILAC followed by LC-MS/MS (Figure 3A). Labeling was performed for 14 days to obtain steady-state measurements of individual proteins that could later be normalized to the mRNA from which each derives. WT cells (in triplicate) and each of the three KO cell lines were cultured for 14 days in SILAC light medium and, in parallel, in SILAC heavy medium. Subsequently, WT cells, cultured in either SILAC light or heavy medium, were mixed with an equal number of KO cells (one of each of the three) cultured in the opposite medium.

Western blotting indicated that FMRP expression in WT cells was unaffected by either medium (Figure 3B). Quantitative LC-MS/MS using each of the six cell mixtures detected 3,358 proteins, of which 64% (n = 2,160) were upregulated, 33% (n = 1,098) were downregulated, and 3% (n = 100) remained unchanged in FMR1-KO relative to WT cells (Figure 3C, gray dots and Table S5). For proteins that derive from FMRP targets, 53% were upregulated, 43% were downregulated, and 3% remained unchanged in FMR1-KO relative to WT cells (Figure 3C, red dots). For proteins that derive from Not FMRP targets, 60% were upregulated, 37% were downregulated, and 3% remained unchanged in FMR1-KO relative to WT cells (Figure 3C, blue dots). Our finding that whether protein abundance was upregulated, downregulated or unchanged was comparable for proteins deriving from mRNAs that do and that do not bind FMRP indicates that FMRP has appreciable effects on mRNAs that are not direct FMRP targets (Figure 3D; see below, where protein abundance was normalized to mRNA abundance).

GO term enrichment analyses revealed that, in FMR1-KO cells relative to WT cells, upregulated proteins deriving from “All mRNAs” are involved in a number of synaptic signaling pathways, vitamin B6 metabolism, glycolysis, and Parkinson’s disease (Figures 3E, S3A and S3B), whereas downregulated proteins are enriched for functions in pyrimidine and cholesterol biosynthesis (Figure 3E). Such metabolic alterations typify FXS patient samples and/or Fmr1-KO mice (Altimiras et al., 2021; Berry-Kravis et al., 2015; Bricout et al., 2008; Çaku et al., 2017; Fulks et al., 2010; Hall et al., 2011; Leboucher et al., 2019; Lisik et al., 2016; Lumaban and Nelson, 2015).

Western blotting was used to confirm LC-MS/MS results for those proteins whose abundance was most changed upon FMR1-KO, demonstrating that FMRP loss increased the abundance of chromodomain helicase DNA binding protein 5 (CHD5), Unc-13 homolog D (UNC13D), and rabphilin 3A (RPH3A), and decreased the abundance of N-myc downstream regulated 1 (NDRG1), transcriptional mediator complex subunit 1 (MED1), and post-glycosylphosphatidylinositol attachment to proteins inositol deacylase 1 (PGAP1) (Figure 3F), none of which is derived from an FMRP target (Table S1).

Globally defining the amount of protein made per mRNA upon loss of FMRP

The molecular mechanism of FMRP in translation remains unclear. For example, some mRNAs that were shown to bind FMRP in HEK293 cells produce abnormally high levels of protein in FXS patient-derived prefrontal cortex, hippocampus, and cerebellum (Ascano et al., 2012). While these and other data support FMRP function as a translational repressor, the degrees to which the observed increased protein abundance upon FMRP loss is attributable to increased mRNA translation, increased mRNA abundance, or both is unknown. This lack of information can be remedied by normalizing the level of each protein to the level of mRNA from which each derives. Normalization is important because FMRP loss alters mRNA abundance in both humans and mice (Liu et al., 2018; Shu et al., 2020; Kurosaki et al., 2021a; Kurosaki et al., 2021b; Figures 2A-D). Moreover, FMRP has been implicated as a translational activator of some mRNAs in mouse cortical neurons and quiescent Drosophila oocytes (Liao et al., 2008; Tang et al., 2015; Greenblatt and Spradling, 2018; Liu et al., 2018). Notably, previous measurements of translational efficiency by quantitating ribosome footprints on individual mRNAs (Liu et al., 2018) cannot distinguish translationally active ribosomes from stalled, i.e., inactive, ribosomes (Zhao et al., 2019): a more accurate measure of translational efficiency would result from quantitations of proteins per se and the subsequent normalization of each protein to the level of its mRNA template.

To this end, we purified RNAs from each of the 12 samples generated for SILAC coupled to LC-MS/MS analysis and performed RNA-seq to quantitate mRNA abundance (Figure S4A and Table S6). In FMR1-KO cells relative to WT cells, there were more downregulated mRNAs (~72% for FMRP targets; ~56% for Not FMRP targets) and fewer upregulated mRNAs (~28% for FMRP targets; ~43% for Not FMRP targets) (Figures 4A and 4B), consistent with loss of FMRP largely destabilizing mRNAs (Figure 2A-2D). Here again, these data indicate the existence of indirect effects of FMRP loss on the abundance of Not FMRP targets as well as, undoubtedly, FMRP targets. As expected, given the role of FMRP as a translational repressor, normalizing protein abundance to mRNA abundance demonstrated that FMR1-KO upregulated the translational efficiency (the level of protein per mRNA) of most (~75% of) mRNAs, including most (~74% of) FMRP targets (Figure 4A). These include previously defined FMRP targets (Suhl et al., 2014) (Figure S4B) and autism-associated transcripts (Figures S4C), some of which were confirmed using protein/mRNA ratios measured by western blotting (data not shown) and RT-qPCR (Figure S4D). We also observed increased translational efficiency for a large fraction (~62%) of Not FMRP targets (Figure 4A and 4B), suggesting that FMRP elimination generally increases translational efficiency not only directly via mRNA binding but also indirectly.

Figure 4. — (A) Stacked histogram representations. (Left) As a proxy for translational efficiency, the percent (%) of mRNAs whose protein product, when normalized to the level of mRNA, is increased, unchanged, or decreased in abundance in *FMR1*-KO relative to WT cells. (Middle) As in Left, but for mRNA abundance. (Right) As in Left, but for protein abundance.

(B) Derived from A, the cumulative fraction of the log₂ fold-change in protein/mRNA abundance, mRNA abundance, or protein abundance for *FMR1-KO* relative to WT cells. P-values were calculated using the two-sided Wilcoxon rank-sum test.

(C) Scatter plot of log₂ fold-change in protein abundance, as determined using SILAC–LC-MS/MS data for *FMRI*-KO cells relative to WT cells (y-axis), vs. log₂ fold-change in mRNA abundance, as determined using RNA-seq data for *FMR1*-KO cells relative to WT cells (x-axis). r, Spearman’s rank correlation coefficient; P, P-value.

Consistent with data deriving from FMR1-KO mouse neural stem cells (Liu et al., 2018) and human SH-SY5Y cells (Kurosaki et al., 2021a) demonstrating that some mRNAs can be translationally buffered-up in the absence of FMRP, an abnormally high level of protein was evident for mRNAs whose level was abnormally low (~31% for all mRNAs, ~31% for FMRP targets, and ~25% for Not FMRP targets) (Figure 4C). Translational buffering-up occurs when, in the absence of FMRP, the relief from translation repression is sufficient to overcompensate for the reduction in mRNA abundance.

It is worth noting that evidence for FMRP possibly functioning as a translational activator derives from our finding that the ratio of protein produced per mRNA for a small percentage of mRNAs (26% of FMRP targets, and 38% of Not FMRP targets) is decreased in FMR1-KO cells (Figures 4A-C). The only significant (P = 4.6 × 10⁻³) feature that distinguishes this class of FMRP-bound mRNAs from FMRP-bound mRNAs that manifest an increase in protein produced per mRNA upon FMRP loss is a lower FBI (Figure S4E). Considering that multiple FMRP-binding sites exist in all FMRP targets examined, and that FMRP loss can affect not only mRNAs that bind FMRP but also mRNAs that do not detectably bind FMRP, any evidence indicating that FMRP may be a translational activator could, in fact, reflect aspects of mRNA metabolism that do not reflect FMRP binding per se. We conclude that FMRP binding to mRNA in SH-SY5Y cells plays a primary role in protecting that mRNA from degradation and repressing translation (see Discussion).

FMRP binding to an mRNA 5'UTR, CDS, or 3'UTR inhibits mRNA degradation and translation

To determine if there are functional differences in FMRP binding to different mRNA regions, we inserted the 81-nt GC-rich FMRP-binding sequence (Fbs) from the CDS of GPRIN1 mRNA (Figure 5A) into the 5'UTR, CDS, or 3'UTR of an EGFP reporter mRNA (Figure 5B). Insertion was either the wild-type Fbs sequence (FbsWT) or a mutated Fbs (FbsMUT) sequence, the latter of which reduced GC content and was predicted to destroy G-quadruplex formation (Figure 5A). Transient expression of each reporter mRNA together with the mRFP reference mRNA in SH-SY5Y WT cells followed by anti-FMRP IPs (Figure 5C) demonstrated that EGFP mRNA harboring FbsWT co-immunoprecipitated with FMRP 2-4-fold more efficiently than did EGFP mRNA or EGFP mRNA harboring FbsMUT (Figure 5D). As controls, neither mRFP mRNA nor endogenous β-actin mRNA co-immunoprecipitated with FMRP (Figure S5A). RT-qPCR analysis revealed that EGFP mRNA abundance (Figure 5E) was increased upon insertion of FbsWT, but not FbsMUT, into either the 5'UTR, CDS or 3'UTR (Figure 5E). This is consistent with FMRP binding increasing mRNA stability (Figures 2A-D) regardless of where within the mRNA body it binds.

Figure 5. — (A) Diagram of the source and sequence of the major 81-nucleotide FMRP-binding site from *GPRIN1* mRNA, i.e., FbsWT, and its mutated variant, FbsMUT, which fails to bind FMRP. Yellow highlight, mutated nucleotides in FbsMUT; Light-blue highlight, U-to-A substitution introduced only into EGFP 5'UTR Fbs mRNAs eliminating an AUG; Underline, predicted G-quadruplex using QGRS Mapper (Kikin et al., 2006).

(B) Diagrams of EGFP reporter mRNAs into which FbsWT or FbsMUT was inserted.

(C) Western blots of SH-SY5Y WT cells transiently expressing an EGFP reporter construct and the reference mRFP construct before (−) or after IP using anti-FMRP or, as a control, mIgG. Results represent three independently performed experiments.

(D) Histogram representations of RT-qPCR analyses of RNAs isolated from cells analyzed in C. Results are shown as means with S.D., where n = 3. (*)P <0.05 and (**)P <0.01 pertain to comparisons between EGFP FbsWT mRNA and EGFP FbsMUT mRNA (two-sided unpaired t-test).

(E) Histogram representations of RT-qPCR analyses of EGFP mRNAs normalized to mRFP mRNAs. Results are shown as means with S.D., where n = 4. (**)P <0.01 and (***)P <0.001 pertain to comparisons between EGFP FbsWT mRNA and EGFP FbsMUT mRNA (two-sided unpaired t-test).

(F) Diagram of the splint RT-qPCR assay used to measure the poly(A) tail content for individual mRNAs. App, 5' adenylation; ddC, 3' dideoxycytidine.

(G) Histogram representations of splint RT-qPCR analyses of RNA purified from samples analyzed in E. Results are shown as means with S.D., where n = 4. (**)P <0.01 pertain to comparisons between EGFP FbsWT mRNA and EGFP FbsMUT mRNA (two-sided unpaired t-test).

(H) Histogram representations of EGFP fluorescence normalized to mRFP fluorescence, each of which was normalized to the level of the encoding mRNA. Results are shown as means with S.D., where n = 4. (*)P <0.05 and (***)P <0.001 pertain to comparisons between EGFP FbsWT mRNA and EGFP FbsMUT mRNA (two-sided unpaired t-test).

To determine how FMRP binding increased mRNA stability, we used poly(A) tail-specific splint RT-qPCR (Kurosaki et al., 2018a) to quantitate the abundance of poly(A) at the 3'-end of each EGFP mRNA (Figure 5F). Poly(A) tail abundance was upregulated upon insertion of FbsWT, but not FbsMUT, into either the 5'UTR, CDS, or 3'UTR of EGF mRNA (Figure 5G). Assays of EGFP fluorescence normalized to mRFP fluorescence, each of which was then normalized to the level of the corresponding mRNA, demonstrated that each insertion of FbsWT into EGFP mRNA reduced amount of protein produced per mRNA relative to the amount of protein produced per either EGFP mRNA or each insertion of FbsMUT into EGFP mRNA (Figure 5H). We conclude that FMRP binding to an mRNA 5'UTR, CDS or 3'UTR protects that mRNA from decay, at least in part by maintaining its poly(A) tail, and also represses its translation.

FMRP binding to poly(A)-bound PABPC1 protects against mRNA deadenylation

In search of molecular mechanism, we used SH-SY5Y-cell lysates to perform FMRP IP and, in parallel as a negative control, mIgG IP, both in the presence of RNase I, followed by LC-MS/MS. We found that the major cytoplasmic poly(A)-binding protein PABPC1 and its paralog PABPC4 were in the top three high-confidence proteins to immunoprecipitate with FMRP (Figures 6A and 6B; Table S7). Additional IPs of SH-SY5Y-cell lysates confirmed that PABPC1 co-immunoprecipitated with FMRP in a largely RNase I-resistant manner (Figure 6C). A direct FMRP–PABPC1 interaction was tested in IPs of a mixture of human FMRP, human PABPC1 and EGFP, each separately produced in and purified from E. coli (Figure S6A). Our finding that PABPC1 co-immunoprecipitated with FMRP, FMRP co-immunoprecipitated with PABPC1, and EGFP failed to co-immunoprecipitate with either protein (Figure 6D) verified a direct interaction between FMRP and PABPC1.

Figure 6. — (A) Silver-stained polyacrylamide gel (left) and western blot (WB; right) of SH-SY5Y-cell lysates bound in the presence of RNase I to, and subsequently eluted from, Protein G Dynabeads alone (beads-bound), bound to mIgG, or bound to anti-FMRP (α-FMRP) and used for LC-MS/MS.

(B) Using samples in A, histogram representations of the LC-MS/MS probability score, i.e., SEQUEST score, for SH-SY5Y-cell proteins of interest that were immunoprecipitated using anti-FMRP in an RNase I-resistant manner.

(C) Western blots before (−) and after IP using anti-FMRP or mIgG and SH-SY5Y-cell lysates in the absence (−) or presence (+) of RNase I with the specified antibody. Results represent three independently performed experiments.

(D) Western blots before (−) and after IP in the presence of RNase I of a mixture of human FMRP, human PABPC1 and, as a negative control EGFP, each produced in and purified from *E. coli*, with the specified antibody. Results represent three independently performed experiments.

(E) Using the protocol described in Figure 1A and WT SH-SY5Y cells, the relative abundance of poly(A) lengths (nt) after anti-FMRP IP or mIgG IP present in raw RIP-seq footprint data and normalized to before IP.

(F) Diagram of the poly(A) tail-specific primer extension of total-cell RNA. rSAP, recombinant shrimp alkaline phosphatase; App, 5' adenylation; ddC, 3' dideoxycytidine.

(G) Left, poly(A)-specific primer extension analyses of bulk poly(A)-tail lengths in the specified *FMR1*-KO cell line relative to each of three WT SH-SY5Y-cell lines. Normalizations were to the level of 18S rRNA, determined using semi-quantitative RT-PCR. Right, Densitometric profiles of lanes in the left panel.

(H) Histogram representations of splint RT-qPCR analyses of RNA from samples analyzed in G. Results are shown as means with S.D., where n = 4. (*)P <0.05, (**)P <0.01 and (***)P <0.001 pertain to comparisons of each *FMR1*-KO cell line with WT cells (two-sided unpaired t-test).

Additional high-confidence proteins found to co-immunoprecipitate with FMRP in the presence of RNase I included FXR2 and FXR1, which are known to constitute RNA granules that are regulated by FMRP (Lai et al., 2020), and RBM14, DDX17 and ATXN2L that, together with PABPC1 and PABPC4, are known to constitute generic RNA granules (Youn et al., 2018). TUBA1B, which is a constituent of microtubules that transports FMRP granules to neuronal-cell sites at which the constituent FMRP-bound mRNA can be translationally activated (Otero et al., 2002; Antar et al., 2005; El Fatimy et al., 2016), was also co-immunoprecipitated (Figure 6B). One interpretation of our data, when compiled, is that FMRP granules sequester the associated FMRP-bound mRNAs, which are protected from translation and deadenylation, in a configuration that is largely resistant to RNase I treatment (see below). Consistent with this possibility, immunofluorescence microscopy demonstrated that FMRP co-localized in the cytoplasm of SH-SY5Y cells with the established axonal granule marker Ras GTPase-activating protein-binding protein 1 (G3BP1) as well as PABPC1, regardless of whether the cells were or were not differentiated to neuron-like cells (Figure S6B). Previously, FMRP was reported to co-localize with G3BP1 in rat cortical neurons (Sahoo et al., 2018), human fetal ovaries (Rosario et al., 2016), and cultured HeLa, HEK293 and human dermal fibroblasts (Taha et al., 2021).

Given our finding of a direct interaction between FMRP and PABPC1, going back to the raw RIP-seq footprinting data, i.e., to sequences that were initially discounted due to their redundancy (Figure 1A), we indeed found poly(A) stretches that were protected from RNase I and sufficiently long (Figure 6E) to encompass the ~ 12-29-nucleotide footprint that characterizes human-cell PABPC1 binding to RNA (Yi et al., 2018). Offering an indication that FMRP binding to PABPC1 promotes FMRP binding to the transcribed body of an mRNA that contains one or more FMRP binding sites, we find no evidence that FMRP binds poly(A)-minus mRNAs: Poly(A)-minus mRNAs (Yang et al., 2011) were present in our anti-FMRP RIP-seq data below background levels (Figure 1 and Figure S6C). mRNA-specific RT-qPCR confirmed that all tested cell cycle-regulated histone mRNAs and zinc-finger mRNAs, each of which is poly(A)-minus, were not enriched in anti-FMRP IPs (Figure S6D).

Corroborating these findings, variants of H4 reporter mRNAs (Figures S6E) previously reported (Maquat and Li et al., 2001) demonstrated that inserting the GPRIN1 Fbs into FLAG-H4 mRNA that harbored the H4 3'UTR and was thus poly(A)-minus, unlike FLAG-H4 mRNA that harbored the β-globin 3'UTR and was thus poly(A)-plus, failed to significantly co-immunoprecipitate with FMRP (Figures S6F and S6G). Concomitantly, inserting the Fbs into poly(A)-plus mRNA increased mRNA abundance and decreased protein production per mRNA (Figures S6H and S6J). These results are consistent with the notion that PABPC1 bound to a poly(A) tail is required for FMRP binding to mRNA that contains an Fbs, which in turn allows FMRP to protect that mRNA from translation and a decrease in abundance.

Consistent with FMRP protecting mRNAs from deadenylation, a poly(A) primer-extension assay (Figure 6F) demonstrated fewer and shortened poly(A) tails in FMR1-KO relative to WT SH-SY5Y cells (Figure 6G). Indeed, poly(A) tail-specific splint RT-qPCR (Figure 5F) revealed that reduced poly(A) was evident on FMRP targets GPRIN1, IRF2BPL, and ANKRD11 mRNAs relative to GAPDH mRNA (Figure 6H) in a mechanism that was reversed by restoring FMRP levels to near-normal (Figures S6K and S6L).

Our findings support the idea that FMRP binding protects an mRNA from deadenylation through direct interaction with PABPC1 to form an RNase I-insensitive mRNA granule in which mRNA translation is repressed.

DISCUSSION

Here we utilize RIP-seq to identify transcriptome-wide human neuronal-cell mRNAs that bind FMRP. We find that FMRP binds many mRNAs that are critical for neuronal-cell function (Figures 1, 2 and S1). Our finding that FMRP binding to transcribed regions of mRNAs at GC-rich and structured sites is not restricted to CDSs, being most dense within 5'UTRs (Figures 2E, S2K and S2L), does not substantiate the idea (Brown et al., 2001; Darnell et al., 2011, Ascano et al., 2012; Chen et al., 2014; Shu et al., 2020) that FMRP recruitment to cellular RNAs depends on protein-coding potential or translating ribosomes. However, either may influence FMRP recruitment considering data indicating that FMRP associates with translationally active ribosomes (Khandjian et al., 1996; Corbin et al., 1997; Feng et al., 1997; Ishizuka et al., 2002; Mazroui et al., 2003; Stefani et al., 2004). While these data did not differentiate between direct and indirect interactions with translationally active ribosomes, FMRP has been shown to directly bind ribosomes in vitro (Chen et al., 2014).

Several important conclusions derive from our studies. First, loss of FMRP causes global mRNA destabilization to a degree that is largest for FMRP targets (Figure 2), including NMD targets (Kurosaki et al., 2021a). Second, loss of FMRP results in global translational activation to a degree that is largest for FMRP targets (Figures 3, 4), including NMD targets (Kurosaki et al., 2021a). Third, FMRP binding to 5'UTRs, CDSs and/or 3'UTRs cannot only translationally repress but also protects the mRNA from deadenylation (Figure 5) because FMRP binds directly to PABPC1 to prevent deadenylation, followed by decay of the transcript body, and to stabilize FMRP binding to the transcript body (Figures 6 and 7). That loss of FMRP can affect the translation and abundance of mRNAs that do not detectably bind FMRP most likely reflects indirect effects mediated by FMRP-bound mRNAs and their encoded proteins.

Figure 7. — FMRP binds to the transcribed body of a polyadenylated mRNA and to PABPC1 at the mRNA poly(A) tail, resulting in FMRP-mediated mRNP sequestration that prevents translation and deadenylation. Other granule-containing proteins are not shown.

Connecting our mechanistic findings to neuronal-cell metabolism, G-quadruplex structures are reported to be enriched in dendritic mRNAs and function as a neurite localization signal (Subramanian et al., 2011). In fact, FMRP has been shown to promote RNA localization to neuronal projections through interactions between its RGG box and G-quadruplex RNA sequences (Ramos et al., 2003; Goering et al., 2020). Thus, polyadenylated mRNAs that we find are bound by FMRP in a translationally silenced and stabilized complex – LC-MS/MS data indicate with other RNA granule constituents and a component of the granule transport machinery (Figure 6) – may represent those phase-separated and phosphorylation-modulated condensates that, once properly localized to neuronal projections, respond to signaling pathways (Figure 7). These pathways result in FMRP dephosphorylation by, e.g., PP2A in response to group 1 mGluR activation, after which FMRP is thought to be released from translationally inactive mRNPs and, as a consequence, there is a burst in translation before the mRNA undergoes degradation by deadenylation (Ceman et al., 2003; Nalavadi et al., 2012; Narayanan et al., 2007; Kim et al., 2019). Results using condensates formed in vitro are consistent with the view that phosphorylated FMRP binding to an mRNA protects that mRNA from translation (Kim et al., 2019; Tsang et al., 2019). Nevertheless, additional details of the mechanism by which translationally inactive FMRP-bound mRNPs become translationally activated remain unclear. While this work was under review, studies of the FMRP-related protein FXR1 revealed that, like FMRP, FXR1 bound to stored mRNAs undergoes condensate formation, i.e. liquid-liquid phase separation, to eventually merge condensates with the translational machinery; this results in the FXR1-dependent translational activation of stored FMR1-bound mRNAs so as to drive spermiogenesis (Kang et al., 2022). Such a conversion from a translationally inactive state to a translationally active state may also be exemplified by the FMRP-mediated enhanced translation of stored mRNAs in Drosophila oocytes (Greenblatt and Spradling, 2018). It is clear that further data are required to understand the transitioning of synaptic FMRP-containing mRNP granules to a translationally activated and ultimately destabilized state, allowing for a burst in protein synthesis.

The formation of extensive intramolecular higher-order structures places the 5'-cap and 3'-poly(A) tail of an mRNA in physical proximity, thereby facilitating their synergistic regulation mRNA translation and stability (Ermolenko and Mathews, 2021). Thus, our model (Figure 7) of physical connections between FMRP that is bound simultaneously to the transcribed body of an mRNA and its 3'-poly(A) tail via PABPC1 would likely add to this compaction, resulting in sequestration not only of the poly(A) tail but also of the 5'-cap as well.

Limitations of the study

Data indicating that loss of FMRP may repress the translation of ~26% of FMRP targets (Figure 4A) could reflect indirect effects or, alternatively, e.g. for the small fraction of FMRP targets that are translationally downregulated and stabilized (Figure 4C), that FMRP binding enhances translation. Alternatively, or additionally, the loss of FMRP binding could be replaced by the binding of other proteins, one or more of which might repress translation. In such a scenario, it could be concluded that FMRP is a translational activator, but it would actually be a translational repressor whose function is obscured by secondary effects. Further studies are required to understand how, for some FMRP-bound mRNAs, the loss of FMRP binding results in decreased protein produced per mRNA, with or without mRNA stabilization.

STAR★METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to Lynne E. Maquat (lynne_maquat@urmc.rochester.edu).

Material availability

All plasmids and cell lines utilized in this study are available upon request directed to Lynne E. Maquat (lynne_maquat@urmc.rochester.edu).

Data and code availability

Sequencing datasets (FASTQ files), including TRIC-seq and anti-FMRP RIP-seq (Kurosaki et al., 2021a), have been deposited in the DNA Data Bank of Japan (DDBJ) Sequence Read Archive: DRA005644. Datasets from RNA-seq used for SILAC–LC-MS/MS analysis are available through the Gene Expression Omnibus (GEO): GSE197221. Proteomics data obtained using SILAC–LC-MS/MS or α-FMRP IP have been deposited in the ProteomeXchange Consortium: PXD032341 or PXD032339, respectively. Unprocessed and uncompressed imaging data are available in Mendeley Data: https://dx.doi.org/10.17632/5k773sxkcw.1.
This paper does not report original code.
Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Human SH-SY5Y neuroblastoma cells (ATCC) or human embryonic kidney (HEK)293T cells (ATCC) were grown in Dulbecco’s Modified Eagle’s Medium/Nutrient Mixture F-12 (DMEM/F-12; Thermo Fisher Scientific Gibco), or DMEM (Thermo Fisher Scientific Gibco) supplemented with 10% fetal bovine serum (FBS; VWR), respectively. When specified, cells (1-5 x 10⁷) were transiently transfected with 20-50 pmol of siRNA using the TransIT-X2 Dynamic Delivery System (Mirus Bio) or Lipofectamine RNAiMAX (Thermo Fisher Scientific Invitrogen). Also, when specified, cells were transiently transfected with 0.1-1 μg of plasmid DNA using the TransIT-X2 Dynamic Delivery System. Where noted, SH-SY5Y cells were differentiated to neurons by culturing in DMEM/F-12 medium with 10% FBS and 10 μM retinoic acid (RA) for three days, followed by culturing in DMEM/F12 medium with 50 ng/ml brain-derived neurotrophic factor (BDNF) for three days (Encinas et al., 2000).

For SILAC analyses, wild-type (WT) SH-SY5Y cells or each of three independently generated FMR1-KO SH-SY5Y cell lines (Kurosaki et al., 2021a) were cultured (~8×10⁶ cells/150-mm dish) for two weeks in DMEM: F-12 SILAC medium (Thermo Fisher Scientific) supplemented with 1× GlutaMAX (Thermo Fisher Scientific), 2.44 mg/ml sodium bicarbonate (Thermo Fisher Scientific), 10% dialyzed fetal bovine serum (Thermo Fisher Scientific), and either L-lysine-2HCl and L-arginine-HCl for light-medium or ¹³C₆ ¹⁵N₂ L-lysine-2HCl and ¹³C₆ ¹⁵N₄ L-arginine-HCl for heavy-medium according to manufacturer instructions.

METHOD DETAILS

Plasmid constructions

The FMRP expression plasmids (pGEX-FMRP and pFLAG-CMV2-FMRP), UPF1 expression plasmid (pCMV-MYC-UPF1), PABPC1 expression plasmid (pGEX-PABPC1), and the monomeric red fluorescent protein expression plasmid (pmRFP) were previously described (Park et al., 2013; Elbarbary et al., 2013; Kurosaki et al., 2014; Kurosaki et al., 2021a) as was pcDEF-HA-PABPC1 (Kumar et al., 2011). All primer pairs used in plasmid constructions are listed in the Key Resources Table.

KEY RESOURCE TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Mouse anti-FMRP	MilliporeSigma	MAB12160; RRID: AB_2283007
Rabbit anti-FMRP	Abcam	Ab17722; RRID: AB_2278530
Rabbit anti-GAPDH	Cell Signaling Technology	2118; RRID: AB_561053
Rabbit anti-Calnexin	Enzo Life Sciences	ADI-SPA-865; RRID:AB_10618434
Rabbit anti-CHD5	Cell Signaling Technology	44829; RRID: AB_2799274
Rabbit anti-UNC13D	Proteintech	16905-1-AP; RRID: AB_2212713
Mouse anti-RPH3A	Proteintech	11396-1-AP; RRID: AB_2181145
Rabbit anti-NDRG1	Cell Signaling Technology	9485; RRID: RRID:AB_2721143
Rabbit anti-TRAP220/MED1	Novus Biologicals	NB100-2574; RRID: AB_609013
Rabbit anti-PGAP1	Proteintech	55392-1-AP; RRID: AB_11232412
Mouse anti-GFP	Santa Cruz Biotechnology	sc-9996; RRID: AB_627695
Rabbit anti-eIF4A3	Bethyl Laboratories	A302-981A; RRID: AB_10748369
Goat anti-UPF2	Santa Cruz Biotechnology	sc-20227; RRID: AB_2272706
Mouse anti-FXR1	MilliporeSigma	05-1529; RRID:AB_1977197
Rabbit anti-FXR2	Cell Signaling Technology	7098; RRID:AB_10891808
Rabbit anti-PABPC1	Abcam	ab21060; RRID:AB_777008
Mouse anti-PABPC1	Santa Cruz Biotechnology	sc-32318; RRID:AB_628097
Rabbit anti-G3BP1	Cell Signaling Technology	61559; RRID:AB_2909406
Goat anti-mouse IgG poly-HRP	Thermo Fisher Scientific	32230; RRID:AB_1965958
Goat anti-rabbit IgG, peroxidase conjugated	Thermo Fisher Scientific	31462; RRID:AB_228338
Goat anti-Mouse IgG (H+L) Cross-Adsorbed Ready Probes Secondary Antibody, Alexa Fluor 488	Thermo Fisher Scientific	R37120; RRID:AB_2556548
Goat anti-Rabbit IgG (H+L) Cross-Adsorbed Ready Probes Secondary Antibody, Alexa Fluor 594	Thermo Fisher Scientific	R37117; RRID:AB_2556545
Chemicals, Peptides, and Recombinant Proteins
TransIT-X2 Dynamic Delivery System	Mirus Bio	MIR 6003
Halt Protease and Phosphatase Inhibitor Cocktail, EDTA-free (100X)	Thermo Fisher Scientific	78443
TRIzol Reagent	Thermo Fisher Scientific	15596018
RNAiso Plus	TaKaRa	9108
IgG from mouse serum	Sigma-Aldrich	15381
Ambion RNase I (cloned, 100U/μl)	Thermo Fisher Scientific, Invitrogen	AM2294
RNaseOUT Recombinant Ribonuclease Inhibitor	Thermo Fisher Scientific Invitrogen	10777019
Shrimp Alkaline Phosphatase	New England Biolabs	M0371
T4 RNA Ligase 2, truncated KQ	New England Biolabs	M0373S
SuperScript III Reverse Transcriptase	Thermo Fisher Scientific, Invitrogen	18080085
KOD DNA Polymerase	EMD Millipore	71085
Q5 High-Fidelity DNA Polymerase	New England Biolabs	M0491S
Urea	EMD Millipore	9510
Acrylamide/bis-acrylamide 19:1 (40%)	Fisher Scientific	BP1406-1
Acrylamide/bis-acrylamide 29:1 (40%)	Fisher Scientific	BP1408-1
Ammonium persulfate, crystal	Mallinckrodt Chemicals	3460-04
UltraPure Temed	Thermo Fisher Scientific, Invitrogen	15524010
Sodium dodecyl sulfate	Sigma-Aldrich	75746-1KG
TWEEN 20	Sigma-Aldrich	P7949-500ML
TRITON X-100	Sigma-Aldrich	T9284-500ML
Immobilon-P^SQ Transfer PVDF Membranes	Millipore	ISEQ00010
Paraformaldehyde solution, 4% in PBS	Affymetrix	19943 1LT
Protein Assay Dye Reagent Concentrate	Bio-Rad	5000006
SYBR-Gold Nucleic Acid Gel Stain	Thermo Fisher Scientific	S11494
Ethidium bromide	Teknova	E3050
Retinoic acid	Cayman Chemical	11017
BDNF	ProSpec	CYT-207
DAPI	Sigma-Aldrich	D9542
ProLong Gold Antifade Reagent	Thermo Fisher Scientific	P36930
NdeI	New England Biolabs	R0111S
EcoRI-HF	New England Biolabs	R3101S
BsrGI-HF	New England Biolabs	R3575S
SacI-HF	New England Biolabs	R3156S
Dulbecco’s Modified Eagle’s Medium (DMEM)	Thermo Fisher Scientific/Gibco	11965092
DMEM/Nutrient Mixture F-12	Thermo Fisher Scientific/Gibco	11320033
DMEM:F-12 for SILAC	Thermo Fisher Scientific	88370
SILAC Protein Quantitation Kit (Trypsin), DMEM	Thermo Fisher Scientific	A33972
GlutaMAX Supplement	Thermo Fisher Scientific/Gibco	35050061
Soduim Bicarbonate 7.5% solution	Thermo Fisher Scientific/Gibco	25080094
Avantor Seradigm Premium Grade Fetal Bovine Serum (FBS)	VWR	97068-085
Bovine Serum Albumin (BSA)	Rockland Immunochemicals	BSA-50
Blotting-grade Blocker Nonfat Dry Milk	Bio-Rad	1706404
Dynabeads Protein G	Thermo Fisher Scientific	10004D
Oligo-d(T)₂₅ Magnetic Beads	New England Biolabs	S1419S
Coaster Spin-X Centrifuge Tube Filter	Corning	8162
AMPure XP	Beckman Coulter	A63882
DynaMarker Prestain Marker for Small RNA	BioDynamics Laboratory	DM253S
SYBR Select Master Mix	Thermo Fisher Scientific	4472920
Critical Commercial Assays
SilverQuest Staining Kit	Thermo Fisher Scientific Invitrogen	LC6070
TrueSeq cDNA Library Prep Kit	Illumina	RS-122-2001
Deposited Data
FMRP RIP-seq footprinting	Kurosaki et al., 2021a	DRA005644
TRIC-seq	Kurosaki et al., 2021a	DRA005644
SILAC RNA-seq	This study	GSE197221
FMRP eCLIP-seq	Van Nostrand et al., 2016	GSE91670; ENCSR331VNX
FLAG-FMRP CLIP-seq	Li et al., 2020	GSE128860
SILAC LC-MS/MS	This study	PXD032341
FMRP-IP LC-MS/MS	This study	PXD032339
Unprocessed imaging data	This study	http://dx.doi.org/doi:10.17632/5k773sxkcw.1
Experimental Models: Cell Lines
HEK293T	ATCC	CRL-11268
SH-SY5Y	ATCC	CRL-2266
SH-SY5Y (FMR1-KO#1)	Kurosaki et al., 2021a	N.A.
SH-SY5Y (FMR1-KO#2)	Kurosaki et al., 2021a	N.A.
SH-SY5Y (FMR1-KO#3)	Kurosaki et al., 2021a	N.A.
Oligonucleotides
Random Hexamers (50 μM)	Thermo Fisher Scientific	N8080127
Silencer Negative Control #1 siRNA	Thermo Fisher Scientific/Ambion	AM4636
FMRP#1 siRNA: 5'-AAAGCTATGUGACUGAUGA-3'	Dharmacon/GE Healthcare	N.A.
FMRP#2 siRNA: 5'-CAGCUUGCCUCGAGAUUUC-3'	Dharmacon/GE Healthcare	N.A.
FXR1 siRNA: 5'-CCAUACAGCUUACUUGAUA-3'	Dharmacon/GE Healthcare	N.A.
FXR2 siRNA: 5'-GAGAGAAGCCUGCUCCAAU-3'	Dharmacon/GE Healthcare	N.A.
UPF2 siRNA: 5'-GGCUUUUGUCCCAGCCAUC-3'	Dharmacon/GE Healthcare	N.A.
pET-EGFP-S: 5'-GAACATATGGTGAGCAAGGGCGAGGA G-3'	Integrated DNA Technologies	N.A.
pET-EGFP-AS: 5'-CGCGAATTCTTACTTGTACAGCTCGTC-3'	Integrated DNA Technologies	N.A.
GPRIN1-5'UTR-FbsWT-S: 5'-AATTCCCGCCGCTGTAGCGCCCCCGGAGCCGGCTGAGCCCGTGCGAGACGTGAGCTGGGACGAGAAGGGCAAGACGTGGGAGGTATACA-3'	Integrated DNA Technologies	N.A.
GPRIN1-5'UTR-FbsWT-AS: 5'-AATTTGTATACCTCCCACGTCTTGCCCTTCTCGTCCCAGCTCACGTCTCGCACGGGCTCAGCCGGCTCCGGGGGCGCTACAGCGGCGGG-3'	Integrated DNA Technologies	N.A.
GPRIN1-5'UTR-FbsMUT-S: 5'-AATTCCCACCGCTATAGCGCCCCCAAAGCCAACTAAGCCCGTGCGAGACGTGAGCAAAGACGAGAAAAACAAGACGAAAGAAATATACA-3'	Integrated DNA Technologies	N.A.
GPRIN1-5'UTR-FbsMUT-AS: 5'-AATTTGTATATTTCTTTCGTCTTGTTTTTCTCGTCTTTGCTCACGTCTCGCACGGGCTTAGTTGGCTTTGGGGGCGCTATAGCGGTGGG-3'	Integrated DNA Technologies	N.A.
GPRIN1-FbsWT-S: 5'-AATTCCCGCCGCTGTAGCGCCCCCGGAGCCGGCTGAGCCCGTGCGAGACGTGAGCTGGGACGAGAAGGGCATGACGTGGGAGGTATACA-3'	Integrated DNA Technologies	N.A.
GPRIN1-FbsWT-AS: 5'-AATTTGTATACCTCCCACGTCATGCCCTTCTCGTCCCAGCTCACGTCTCGCACGGGCTCAGCCGGCTCCGGGGGCGCTACAGCGGCGGG-3'	Integrated DNA Technologies	N.A.
GPRIN1-FbsMUT-S: 5'-AATTCCCACCGCTATAGCGCCCCCAAAGCCAACTAAGCCCGTGCGAGACGTGAGCAAAGACGAGAAAAACATGACGAAAGAAATATACA-3'	Integrated DNA Technologies	N.A.
GPRIN1-5'UTR-FbsMUT-AS: 5'-AATTTGTATATTTCTTTCGTCATGTTTTTCTCGTCTTTGCTCACGTCTCGCACGGGCTTAGTTGGCTTTGGGGGCGCTATAGCGGTGGG-3'	Integrated DNA Technologies	N.A.
GPRIN1-3′UTR-insert-S: 5'-GTACAAGTGATACTCAGATCTCGAGCT-3'	Integrated DNA Technologies	N.A.
GPRIN1-3′UTR-insert-AS: 5'-CGAGATCTGAGTATCACTT-3'	Integrated DNA Technologies	N.A.
FLAG-H4-S: 5'-CCATGGACTACAAAGACGATGACGACAAGCTTT-3'	Integrated DNA Technologies	N.A.
FLAG-H4-AS: 5'-AAAGCTTGTCGTCATCGTCTTTGTAGTCCATGG-3'	Integrated DNA Technologies	N.A.
Fbs-H4-S: 5'-AGCTTGCCGCTGTAGCGCCCCCGGAGCCGGCTGAGCCCGTGCGAGACGTGAGCTGGGACGAGAAGGGCATGACGTGGGAGGTATACC-3'	Integrated DNA Technologies	N.A.
Fbs-H4-AS: 5'-AGCTGGTATACCTCCCACGTCATGCCCTTCTCGTCCCAGCTCACGTCTCGCACGGGCTCAGCCGGCTCCGGGGGCGCTACAGCGGCA-3'	Integrated DNA Technologies	N.A.
DNA adaptor: 5'-App-CTGTCTCTTATACACATCTCCGAGCCCACGAGAC-ddC-3'	Integrated DNA Technologies	N.A.
DNA adaptor RT primer: 5'-TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG-3'	Integrated DNA Technologies	N.A.
Splint RT primer: 5'-TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGTTTT-3'	Integrated DNA Technologies	N.A.
GPRIN1 mRNA F: 5'-AAAGCAGGCCGATTCCACTTC-3'	Integrated DNA Technologies	N.A.
GPRIN1 mRNA R: 5'-TCCTTCCTCGGTGACACTGTA-3'	Integrated DNA Technologies	N.A.
mTOR mRNA F: 5'-TCCGAGAGATGAGTCAAGAGG-3'	Integrated DNA Technologies	N.A.
mTOR mRNA R: 5'-CACCTTCCACTCCTATGAGGC-3'	Integrated DNA Technologies	N.A.
TSC2 mRNA F: 5'-CCAAACCAACAAGCAAAGATTCA-3'	Integrated DNA Technologies	N.A.
TSC2 mRNA R: 5'-AGGTCTTCGTTGGAAGGGTAA-3'	Integrated DNA Technologies	N.A.
ANKRD11 mRNA F: 5'-AGATGACGACACGCCTTTG-3'	Integrated DNA Technologies	N.A.
ANKRD11 mRNA R: 5'-CCTAACAGGAGGTTCACCATCG-3'	Integrated DNA Technologies	N.A.
IRF2BPL mRNA F: 5'-GGACTTCTCGGAACCCGTATG-3'	Integrated DNA Technologies	N.A.
IRF2BPL mRNA R: 5'-GCGCTGTCTCGATCACGAAT-3'	Integrated DNA Technologies	N.A.
SHANK3 mRNA F: 5'-TGGGGATCACCGACGAGAAT-3'	Integrated DNA Technologies	N.A.
SHANK3 mRNA R: 5'-GCACAGCTCTCCTGGTTGTAG-3'	Integrated DNA Technologies	N.A.
DUSP3 mRNA F: 5'-TTGGCTCAAAAGAATGGCCG-3'	Integrated DNA Technologies	N.A.
DUSP3 mRNA R: 5'-CGCATCATGAGGTAGGCGAT-3'	Integrated DNA Technologies	N.A.
RPS6KB1 mRNA F: 5'-GCTGGCCTAGAGCCTGTG-3'	Integrated DNA Technologies	N.A.
RPS6KB1 mRNA R: 5'-TTTCGCACCTGGAACACCTT-3'	Integrated DNA Technologies	N.A.
SYNGAP1 mRNA F: 5'-CGAGTCCAGTCGCAACAAACT-3'	Integrated DNA Technologies	N.A.
SYNGAP1 mRNA R: 5'-GATGGAGCTTTTTAGCCGTCG-3'	Integrated DNA Technologies	N.A.
GAPDH mRNA F: 5'-GAAGGTGAAGGTCGGAGTCA-3'	Integrated DNA Technologies	N.A.
GAPDH mRNA R/pre-mRNA R: 5'-GTTGAGGTCAATGAAGGGGTC-3'	Integrated DNA Technologies	N.A.
β-actin mRNA F: 5'-AATCGTGCGTGACATTAAG-3'	Integrated DNA Technologies	N.A.
β-actin mRNA R: 5'-ATGATGGAGTTGAAGGTAGT-3'	Integrated DNA Technologies	N.A.
18S rRNA F: 5'-GGGAAACCAAAGTCTTTGGG-3'	Integrated DNA Technologies	N.A.
18S rRNA R: 5'-GGAATTAACCAGACAAATCGC-3'	Integrated DNA Technologies	N.A.
EGFP mRNA F: 5'-ACGTAAACGGCCACAAGTTC-3'	Integrated DNA Technologies	N.A.
EGFP mRNA R: 5'-AAGTCGTGCTGCTTCATGTG-3'	Integrated DNA Technologies	N.A.
mRFP mRNA F: 5'-CCCCGTAATGCAGAAGAAGA-3'	Integrated DNA Technologies	N.A.
mRFP mRNA R: 5'-CTTGGCCATGTAGGTGGTCT-3'	Integrated DNA Technologies	N.A.
UPF1 mRNA F: 5'-ACCTATTACACGAAGGACCTCC-3'	Integrated DNA Technologies	N.A.
UPF1 mRNA R: 5'-ACGTCCGTTGCAGAACCAC-3'	Integrated DNA Technologies	N.A.
MYC-UPF1 mRNA F: 5'-GCAGAAGCTGATCTCAGAGG-3'	Integrated DNA Technologies	N.A.
MYC-UPF1 mRNA R: 5'-GAGGTCCTTCGTGTAATAGGTG-3'	Integrated DNA Technologies	N.A.
HIST1H2AG mRNA F: 5'-GCTAAGGCCAAGACTCGCTC-3'	Integrated DNA Technologies	N.A.
HIST1H2AG mRNA R: 5'-GACCCGCTCGGCATAGTTG-3'	Integrated DNA Technologies	N.A.
HIST1H3B mRNA F: 5'-ATGGCTCGTACTAAACAGACAGC-3'	Integrated DNA Technologies	N.A.
HIST1H3B mRNA R: 5'-TTCCGAATCAGCAACTCGGTC-3'	Integrated DNA Technologies	N.A.
HIST1H4C mRNA F: 5'-GCAAAGGCGGAAAAGGCTTG-3'	Integrated DNA Technologies	N.A.
HIST1H4C mRNA R: 5'-TAGCCGGTTTTGTAATGCCCT-3'	Integrated DNA Technologies	N.A.
HIST2H2BE mRNA F: 5'-ATGCCTGAACCGGCAAAATC-3'	Integrated DNA Technologies	N.A.
HIST2H2BE mRNA R: 5'-TGGATCTCGCGGGATGTGAT-3'	Integrated DNA Technologies	N.A.
ZNF124 mRNA F: 5'-AATGAACTCGGTTGCCTTTG-3'	Integrated DNA Technologies	N.A.
ZNF124 mRNA R: 5'-ATGCTCTGGTCTTCCCCTTT-3'	Integrated DNA Technologies	N.A.
ZNF460 mRNA F: 5'-GCCTTGTACGTGGAGGTGAT-3'	Integrated DNA Technologies	N.A.
ZNF460 mRNA R: 5'-GGTCTGTCCCTGGTCTCAAA-3'	Integrated DNA Technologies	N.A.
Recombinant DNA
pET28A	MilliporeSigma	69864
pEGFP-N1	Clontech	6085-1
pEGFP-C3	Clontech	6082-1
pmRFP	Elbarbary et al., 2013	N.A.
pGEX-FMRP	Kurosaki et al., 2021a	N.A.
pFLAG-CMV2-FMRP	Kurosaki et al., 2021a	N.A.
pCMV-MYC-UPF1	Kurosaki et al., 2014	N.A.
pGEX-PABPC1	Park et al., 2013	N.A.
pET-EGFP	This study	N.A.
pEGFP-5′UTR-FbsWT	This study	N.A.
pEGFP-5′UTR-FbsMUT	This study	N.A.
pEGFP-CDS-FbsWT	This study	N.A.
pEGFP-CDS-FbsMUT	This study	N.A.
pEGFP-3′UTR-FbsWT	This study	N.A.
pEGFP-3′UTR-FbsMUT	This study	N.A.
pmCMV-FFAG-H4-H4 3′UTR	This study	N.A.
pmCMV-FFAG-H4-Gl 3'UTR	This study	N.A.
pmCMV-FFAG-Fbs-H4- H4 3'UTR	This study	N.A.
pmCMV-FFAG-Fbs-H4-Gl 3'UTR	This study	N.A.
Software and Algorithms
Cutadapt 1.12	Martin et al., 2011	http://code.google.com/p/cutadapt/
STAR_2.5.2b	Dobin et al., 2013	https://code.google.com/archive/p/rna-star/
featureCounts	Liao et al., 2014	http://subread.sourceforge.net
edgeR (v.3.14.0)	Robinson et al., 2010	https://bioconductor.org/packages/release/bioc/html/edgeR.html
G4Hunter	Bedrat et al., 2016	https://github.com/AnimaTardeb/G4Hunter
RNALfold 2.5.0	Hofacker et al., 2004	https://www.tbi.univie.ac.at/RNA/RNALfold.1.html
ViennaRNA Package 2.0	Lorenz, 2011	https://www.tbi.univie.ac.at/RNA/
stringr	Wickham, 2010	https://stringr.tidyverse.org/
DescTools	Signorell et al., 2021	https://andrisignorell.github.io/DescTools/
CLIPper ver. 0.1.4	Lovci et al., 2013	https://www.encodeproject.org/software/clipper/
BEDTools ver. 2.29.1	Quinlan and Hall, 2010	https://bedtools.readthedocs.io/en/latest/
PANTHER Classification System	Mi et al., 2013	http://www.pantherdb.org/
Cytoscape ver. 3.6.1.	Shannon et al., 2003	https://cytoscape.org/
QGRS Mapper	Kikin et al., 2006	http://bioinformatics.ramapo.edu/QGRS/index.php
SEQUEST	Eng et al, 1994	https://www.thermofisher.com/order/catalog/product/OPTON-30945?SID=srch-srp-OPTON-30945
Image Studio Lite ver. 4.0	LI-COR Bioscience	https://www.licor.com/bio/image-studio-lite/

Open in a new tab

To construct pET-EGFP, EGFP sequences were PCR-amplified using pEGFP-N1 (Clontech) and the primer pair pET-EGFP-S and pET-EGFP-AS. The resulting PCR product was digested using NdeI and EcoRI, and inserted into NdeI and EcoRI sites of pET28a (MilliporeSigma).

To generate pEGFP-5′UTR-FbsWT or pEGFP-5′UTR-FbsMUT, the primer pair GPRIN1-5'UTR-FbsWT-S and GPRIN1-5'UTR-FbsWT-AS or GPRIN1-5'UTR-FbsMUT-S and GPRIN1-5'UTR-FbsMUT-AS were annealed to one another after phosphorylation with T4 polynucleotide kinase. Annealed products were inserted into the EcoRI site of pEGFP-N1 (Clontech).

To generate pEGFP-CDS-FbsWT or pEGFP-CDS-FbsMUT, the primer pair GPRIN1-FbsWT-S and GPRIN1-FbsWT-AS or GPRIN1-FbsMUT-S and GPRIN1-FbsMUT-AS were annealed to one another after phosphorylation with T4 polynucleotide kinase. Annealed products were inserted into the EcoRI site of pEGFP-C3 (Clontech).

To generate pEGFP-3′UTR-FbsWT or pEGFP-3′UTR-FbsMUT, the primer pair GPRIN1-3′UTR-insert S and GPRIN1-3′UTR-insert AS were annealed to one another after phosphorylation with T4 polynucleotide kinase. Annealed products were inserted into the BsrGI and SacI sites of pEGFP-CDS-FbsWT and pEGFP-CDS-FbsMUT.

To generate pmCMV-FLAG-H4-H4 3'UTR or pmCMV-FLAG-H4-Gl 3'UTR, FLAG-H4-S and FLAG-H4-AS were first annealed and phosphorylated using T4 polynucleotide kinase, and then inserted into the StuI site of pmCMV-H4 or pmCM-H4-Gl (pmCMV-H4-Gl 3'UTR) (Maquat and Li et al., 2001).

To generate pmCMV-FLAG-Fbs-H4-H4 3'UTR or pmCMV-FLAG-Fbs-H4-Gl 3'UTR, Fbs-H4-S and Fbs-H4-AS were first annealed and phosphorylated using T4 polynucleotide kinase, and then inserted into the HindIII site of pmCMV-FLAG-H4-HA 3'UTR or pmCMV-FLAG-H4-Gl 3'UTR.

All plasmids were sequenced to confirm their successful construction.

Cell lysis, and protein and RNA preparations

SH-SY5Y-cell or HEK293T-cell lysates were generated, proteins were quantitated, and RNA was purified using TRIzol reagent (Thermo Fisher Scientific) and quantitated as previously described (Kurosaki et al., 2014; Kurosaki et al., 2021a).

siRNAs

All siRNA sequences are described in the Key Resources Table.

Western blotting

After electrophoresing cell lysates in 8-2% polyacrylamide, proteins were transferred to a nitrocellulose (Bio-Rad) or polyvinylidene difluoride (Millipore) membrane and probed as described (Kurosaki et al., 2021a) using antibodies listed in the Key Resources Table.

Immunoprecipitations (IPs)

Samples were generated before and after IP in the presence or absence of RNase I (Thermo Fisher Scientific) as described (Kurosaki et al., 2014; Kurosaki et al., 2021a) using 5-10 μg of mouse monoclonal anti-FMRP (Millipore) or control mouse IgG (Sigma) per 0.5 ml of cell lysate at 1-2 mg protein/ml. The relative amount of cell lysate used in western blotting before IP compared to after IP was 1:10.

RT-qPCR

Reverse transcription (RT)-coupled to real-time (q)PCR (RT-qPCR) was undertaken to quantitate pre-mRNAs and mRNAs as previously described (Kurosaki et al., 2014; Kurosaki et al., 2021a). In brief, total-cell RNA was treated with RQ01 DNase I (Promega), random hexamers (Thermo Fisher Scientific) were used to prime RT by Superscript III (Thermo Fisher Scientific), and PCR was then undertaken with transcript-specific primers as described in the Key Resources Table.

cDNA library constructions for RIP-seq and RNA-seq

RIP-seq footprinting libraries were constructed essentially as detailed (Kurosaki et al., 2018b; Kurosaki et al., 2021a). For RNA-seq library constructions, total-cell RNA concentrations were determined using a NanopDrop 1000 spectrophotometer (NanoDrop, Wilmington, DE). RNA quality was assessed using the Agilent Bioanalyzer (Agilent, Santa Clara, CA). The TruSeq Stranded mRNA Sample Preparation Kit (Illumina, San Diego, CA) was used for library construction following the manufacturer’s protocols. Briefly, mRNA was purified from 200 ng of total-cell RNA using oligo(dT) magnetic beads, and 150–350-nucleotide fragments were generated by exposure to divalent cations at elevated temperature. First-strand cDNA synthesis was performed using random hexamers, followed by second-strand cDNA synthesis with dUTP incorporation for strand marking. End repair and 3'-end adenylation was then performed on the double-stranded cDNA. Illumina adaptors were ligated to both cDNA ends. Ligated cDNAs were purified using AMPureXP beads (Beckman Coulter) and then PCR-amplified using primers specific to the adaptor sequences, generating cDNA amplicons of ~200–500 basepairs. Amplified libraries were sequenced using a NovaSeq 6000 DNA sequencer (Illumina, San Diego, CA), generating single-end reads of 100 nucleotides with at least 30 million raw reads per sample.

Computational analyses of RIP-seq and RNA-seq data, GO term enrichment, biological pathway, and biological network analysis

Analyses of RIP-seq data were as previously reported (Kurosaki et al., 2021b). For analyses of RNA-seq data, bcltofastq version 2.19.0 was used to demultiplex raw reads generated from the Illumina base calls. Quality filtering and adapter removal were undertaken using FastP version 0.20.0 with parameters "–length_required 35 –cut_front_window_size 1 –cut_front_mean_quality 13 –cut_front –cut_tail_window_size 1 –cut_tail_mean_quality 13 –cut_tail –w 8 -y –r -j". Cleaned reads were then mapped to the Homo Sapiens reference genome (GRCh38.p13 + Gencode-31 Annotation) using STAR_2.7.0f and parameters "–twopassMode Basic –runMode alignReads –outSAMtype BAM Unsorted –outSAMstrandField intronMotif –outFilterIntronMotifs RemoveNoncanonical". Gene-level read quantitation was derived using the subread-1.6.4 package (featureCounts) with a GTF annotation file (Gencode-31) and parameters "-s 2 -t exon -g gene_name". Differential expression analysis was performed using DESeq2-1.22.1 with an adjusted P-value threshold of 0.05 within R version 3.5.1 (https://www.R-project.org/). Gene Ontology (GO) term enrichment analysis for biological processes employed the Protein Analysis Through Evolutionary Relationships (PANTHER) classification system (Mi et al., 2013) version 16.0 (http://www.pantherdb.org/) and a false discovery rate (FDR) P-value threshold of 0.05 (Fisher’s exact test). Cytoscape (Shannon et al., 2003) ver. 3.6.1 with an edge cutoff value of 0.375 were used to draw biological interaction networks.

Meta-analyses of our RIP-seq and published CLIP-seq data

Raw anti-FMRP RIP-seq data derive from Kurosaki et al., 2021a. For CLIP-seq data (Li et al., 2020), fastq files from the sequence read archive (SRA; Leinonen et al., 2011; accession ID: SRP189455) were downloaded using the fastq-dump command in the SRA-Toolkit. After removing adapters using cutadapt (Martin, 2011), reads were mapped to hg19 using STAR_2.5.2b (Dobin et al., 2013).

For the FMRP-eCLIP-seq dataset (Van Nostrand et al., 2016) in ENCODE, bam files mapped to hg19 were downloaded from the ENCODE website (accession ID: ENCFF328KAL, ENCFF436TNC, ENCFF736XNI) for mRNA analyses.

Peak-calling analyses

Peak calling was performed using the CLIPper peak-calling algorithm (Lovci et al., 2013) using the bam files of FMRP RIP-seq footprinting. Since normalization to size-matched input control reduces analytical artifacts (Van Nostrand et al., 2016), normalization to Input sample was performed. For each peak, the number of overlapping reads in the α-FMRP IP and Input samples was counted. Peaks with a significantly higher number of reads in the FMRP-IP samples (P < 0.05; two-sided unpaired t-test) were extracted and used for subsequent analyses. Overlapping peaks in each region of mRNAs were extracted using the “intersect” command in the BEDTools (Quinlan and Hall, 2010). The RNA sequences of peaks were extracted using the "getfasta" command in the BEDTools.

Custom annotation and format conversions

For mRNA, custom annotations were defined in Kurosaki et al. (2021a). UTR features were divided into 5′UTRs and 3′UTRs based on the strand and position relative to the CDS feature of the same transcript.

Calculation of FMRP binding to full-length mRNAs, 5′UTRs, CDSs, or 3′UTRs

The number of reads mapped to each mRNA was counted using the –t exon option in featureCounts (Liao et al., 2014). Transcripts per million (TPM) for each mRNA were calculated. Finally, the log₂ fold-change of RNA abundance (FMRP-IP/Input) was calculated. The number of reads mapped to either a 5′UTR, CDS, or 3′UTR was counted using the –t CDS or –t UTR option in featureCounts. Subsequently, the TPM for each region of each mRNA was calculated.

Calculation of G4-structure density

The number of G4 structures in each mRNA, or each region of each mRNA, was counted using G4Hunter (Bedrat et al., 2016). G4-structure density was calculated by dividing the number of G4 structures by the length of the mRNA or each of its regions.

Calculation of mRNA structuredness

The free energy of the most stable structure for each mRNA was calculated using RNALfold (Hofacker et al., 2004) in the ViennaRNA Package (Lorenz, 2011). G4 structure was incorporated into the structure prediction algorithm using the -g option. The structuredness of mRNA was defined as the free energy divided by mRNA length as described (Fischer et al., 2020).

Calculation of codon optimality

Relative synonymous codon usage (RSCU) for each codon, a value of the observed frequency divided by the expected frequency of that codon, assuming that codons encoding the same amino acid, i.e., synonymous codons, have equivalent usage, was mathematically calculated as described (Sharp and Li 1987). The codon adaption index (cAI) was then calculated as a geometric mean of normalized codon-based RSCUs (Sharp and Li 1987).

Calculations of tri- or tetra-nucleotide densities in mRNAs

The number of tri- or tetra-nucleotides in each mRNA was counted using the str_count function in the stringr package (Wickham, 2010). In the analysis of tri-nucleotide densities, we considered the reading frames of mRNAs based on CDS and UTR annotations in Gencode.v19. Subsequently, the density of each tri- or tetra-nucleotide for each mRNA was determined by dividing the number of tri- or tetra-nucleotides in each mRNA by mRNA length.

mRNA half-life changes upon FMRP-KD in SH-SY5Y cells

TRIC-seq data (Kurosaki et al., 2021a) were analyzed using nonlinear regression. Normalized RPKM (called remaining RNA, hereafter) at each time point was used in calculations. The degradation constant (k_d) for each mRNA was determined by fitting the remaining RNA at each time point, using the nonlinear least-squares (nls) function in R (https://www.rdocumentation.org/packages/stats/versions/3.6.2/topics/nls), to the following formula:

Remaining RNA = b + (1 - b) \times exp (- k_{d} \times time)

b is the basal RNA level, which was constrained to 0 < b < 1, and the kd was constrained to > 0 using the port algorithm (http://www.netlib.org/port/).

Subsequently, the half-life of each mRNA (t_1/2) was determined using the formula:

t_{{}^{1}∕_{2}} = \frac{\log 2}{k_{d}}

mRNA half-life changes upon FMRP-KD were determined by dividing the mean half-life in the presence of each FMRP siRNA by the mean half-life in the presence of Ctl siRNA.

Correlation analyses

Spearman’s correlation coefficients (rho) with 95% confidence intervals were calculated using the Spearman Rho function in DescTools (Signorell et al., 2021) after removing those pairs for which data are not available or have infinite values.

Meta-analysis of poly(A) RIP-seq reads

The number of poly(A) sequences of each length in each trimmed fastq file deriving from RIP-seq experiments was counted using the str_dup function and str_count function of the stringr package (Wickham, 2010). The number of poly(A) sequences in RIP-seq was normalized using the Input sample, and the averaged ratio for biological replicates was calculated.

Sample preparation for SILAC protein quantitation

Wild-type (WT) SH-SY5Y cells (3×10⁶), cultured in either light or heavy medium, and one of the FMR1-KO cells (3×10⁶), cultured in the alternative medium relative to WT cells, were mixed and subsequently lysed by adding 300 μl of 5% SDS, 100 mM triethylammonium bicarbonate (TEAB), vortexing, and then sonication (QSonica) for five cycles with 1-min rest on ice between cycles. Lysates were cleared by centrifuging at 15,000 x g for 5 min, after which protein concentrations were determined using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific). Samples were diluted to 1 mg/mL in 5% SDS and 50 mM TEAB. Protein (25 μg) was reduced by adding dithiothreitol to 2 mM, incubated at 55°C for 60 min, and subsequently alkylated by adding iodoacetamide to 10 mM and incubating in the dark at room temperature for 30 min. Phosphoric acid was added to 1.2%, after which six volumes of 90% methanol and 100 mM TEAB were added. Samples were loaded onto S-Trap Micros Spin Columns (ProtiFi), centrifuged at 4,000 x g for 1 min, and washed twice with 90% methanol and 100 mM TEAB. Trypsin (1 ug) in 20 uL of 100 mM TEAB was added to each S-Trap, then 20 μL of TEAB was added, and samples were incubated in a humidity chamber at 37°C overnight. Digested peptides were collected from the S-Traps by centrifuging at 4,000 x g for 1 min after sequentially adding 0.1% trifluoroacetic acid (TFA) in acetonitrile and 0.1% TFA in 50% acetonitrile. Samples were lyophilized using a Speed Vac (Labconco) and resuspended in 0.1% TFA.

LC-MS/MS for SILAC protein quantitation

Peptides were loaded onto a C18 nano-column (100 μm x 30 cm) packed with 1.8 μm beads (Sepax Technologies, Inc) using an Easy nLC-1200 HPLC (Thermo Fisher Scientific) connected to an Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher Scientific). A Nanospray Flex source operating at 2 kV delivered ions to the mass spectrometer. Peptides were eluted from the column using a flow rate of 300 nl/min and a multi-step gradient for a total run-time of 120 min. The gradient began with a mixture of 97% Solvent A (0.1% formic acid in water) and 3% Solvent B (0.1% formic acid in 80% acetonitrile) and was followed sequentially by a 2-min hold, a ramp-up over 6 min to 90% Solvent A and 10% Solvent B, an increase over 95 min to 62% Solvent A and 38% Solvent B, a ramp-up over 5 min to 10% Solvent A and 90% Solvent B, a 3-min hold, and then a return after 2 min to starting conditions. The column was re-equilibrated for 7 min. The Fusion Lumos was operated in data-dependent mode, with a cycle time of 2 sec. The full scan had a 375-400 m/z range, a 120,000 resolution at m/z of 200, a 4e5 AGC target, and a 50 msec maximum injection time. Peptides with a charge state between 2 and 5 were fragmented at a collision energy of 30 using higher-energy collisional dissociation (HCD) and a 1.5 mass-to-charge ratio (m/z isolation width). MS2 scans were collected in the Orbitrap with a resolution of 15,000 at 200 m/z, a maximum injection time of 30 msec, and a 5e4 AGC setting. Dynamic exclusion was 30 sec.

Data analysis for SILAC protein quantitation

The SwissProt human database and SEQUEST (Eng et al., 1994) within the Proteome Discoverer software platform, version 2.4 (Thermo Fisher Scientific), were used to search the raw data, selecting trypsin and allowing for no more than two missed cleavages, with a mass tolerance of 10 ppm for MS1 and 0.025 Da for MS2. Carbamidomethylation on cysteine and oxidation of methionine were selected as, respectively, fixed and variable modifications. Percolator was used to filter out peptides with a q-value greater than 0.01, i.e., the false discovery rate. Both Minora Feature Detector and Precursor Ions Quantifier nodes were used to quantitate abundances in SILAC light relative to SILAC heavy peptides. Heavy-to-light protein ratios for a given protein were determined using the median peptide ratio.

Fluorescent EGFP and mRFP quantitations

EGFP and mRFP fluorescence in SH-SY5Y cells cultured in 6-well plates at 50% confluency was measured using a SpectraMax M4 Microplate Reader (Molecular Devices), averaging nine fluorescent-point measurements per well.

Sample preparation for LC-MS/MS using FMRP-IP samples

After SDS-PAGE, lanes were divided into three regions. Each region was cut into ~2 mm sections, and sections were washed overnight in 50% methanol/water. After washing again in 47.5/47.5/5 % methanol/water/acetic acid for 2 hours, sections were dehydrated using acetonitrile and dried in a Vacuum Concentrator (Labconco). Disulfide bonds were reduced by the addition of 30 μl of 10 mM dithiothreitol (DTT) in 100 mM ammonium bicarbonate for 30 min. The resulting free cysteines were alkylated by incubating with 50 mM iodoacetamide in the dark for 30 min. They were then sequentially washed with aliquots of acetonitrile, 100 mM ammonium bicarbonate, and acetonitrile, and dried in a Vacuum Concentrator. Proteins in the dried gel pieces were enzymatically digested by the addition of 300 ng of trypsin in 50 mM ammonium bicarbonate. Depending on the volume of the pieces, excess ammonium bicarbonate was added to cover them, and digestion was extended overnight at 37 °C. The resulting peptides were extracted by the addition of 50% acetonitrile and 0.1% TFA, dried in a Vacuum Concentrator, desalted using homemade C18 spin columns, dried down again, and resuspended in 0.1% TFA.

LC-MS/MS using FMRP-IP samples

Peptides were injected and bound to a 30-cm C18 column, packed with 1.8 μm beads (Sepax), using an Easy nLC 1000 HPLC coupled to a Q Exactive Plus Hybrid Quadrupole-Orbitrap Mass Spectrometer (Thermo Fisher Scientific). A multi-step gradient was employed using Solvent B (0.1% formic acid in acetonitrile) starting at 3% Solvent B and holding for 2 min, increasing to 30% Solvent B over 41 min, then increasing to 70% Solvent B over 3 min, and holding for 4 min, before returning to 3% B over 2 min. The column was then re-equilibrated with Solvent A (0.1% formic acid in water) for 8 min for a total run time of 60 min. The Q Exactive Plus was operated in data-dependent mode while using a dynamic exclusion list to identify lower abundant proteins.

Data analyses for LC-MS/MS using FMRP-IP samples

Data were analyzed using the SEQUEST search engine within the Proteome Discoverer 1.4 software program, using the SwissProt human database. MSI mass tolerance was set to 10 ppm, while MS2 mass tolerance was set to 0.025 Da. Carbamidomethyl was set as a fixed modification, and methionine oxidation was set as a variable modification. Percolator was used as the FDR calculator, filtering out peptides with a q-value greater than 0.01.

Immunofluorescence

Either differentiated or undifferentiated SH-SY5Y cells (0.5 × 10⁶ cells) were cultured overnight on BioCoat poly-D-lysine/laminin coverslips (Corning) and subsequently fixed for 15 min at room temperature using 4% paraformaldehyde in phosphate-buffered saline (PBS) (Affymetrix). Coverslips were washed five times with PBS-T (PBS containing 0.1% Triton X-100), treated with 3% bovine serum albumin in TBS-T (PBS containing 0.1% Tween20) for 30 min at room temperature, washed once with TBS-T, and incubated in primary antibody diluted in TBS-T overnight at 4 °C. Coverslips were then washed extensively with TBS-T and then incubated with 1:1,000 (v/v) Alexa Fluor 488-labeled goat anti-mouse or Alexa Fluor 594-labeled goat anti-rabbit IgG (Thermo Fisher Scientific) and 1 μg/ml 4',6-Diamidine-2'-phenylindole dihydrochloride (DAPI) for 2 h at room temperature. After further extensive washing, coverslips were mounted using ProLong Gold antifade reagent (Thermo Fisher Scientific). Images were captured by an Olympus FV-1000 confocal laser scanning microscope.

Purification of human PABPC1, human FMRP and EGFP from E. coli

PABPC1, FMRP, and EGFP proteins were generated in bacteria and purified as previously reported (Park et al., 2013; Elbarbary et al., 2017; Kurosaki et al., 2021a).

RT-PCR, poly(A) primer-extension analysis, and splint RT-qPCR

The DNA adaptor, the DNA adaptor RT primer, the splint RT primer, and transcript-specific PCR primers are described in Key Resources Table. All three methods utilized total-cell RNA that was treated with RQ01 DNase I (Promega).

RT-PCR followed by gel staining using SYBR Gold (Thermo Fisher Scientific) was used to quantitate 18S rRNA (Kurosaki et al., 2014; Kurosaki et al., 2018a). Poly(A) primer-extension analysis for poly(A) tail length determinations utilized the DNA adaptor, the RT primer for poly(A) length analysis, and α-[³²P]-TTP (Perkin Elmer) as previously described (Kurosaki et al., 2018a). Densitometric images were captured using Image Studio Lite ver. 4.0.

In splint RT-qPCR analysis for mRNA-specific poly(A) abundance, RNA was ligated to the DNA adaptor, cDNA synthesis by Superscript III was primed using either the DNA adaptor RT primer or the splint RT primer, and PCR was subsequently performed using transcript-specific PCR primers.

QUANTIFICATION AND STATISTICAL ANALYSIS

Details of statistical analyses, the sample size for each experiment, and software utilized are described in the main text, figure legends, and methods.

Supplementary Material

NIHMS1849220-supplement-1.pdf^{(2.1MB, pdf)}

Table S1: FMRP targets in human SH-SY5Y cells, related to Figure 1 and S1G

NIHMS1849220-supplement-2.xlsx^{(47.5KB, xlsx)}

Table S2: Not FMRP targets in human SH-SY5Y cells, related to Figure 1 and S1G

NIHMS1849220-supplement-3.xlsx^{(68KB, xlsx)}

Table S3: mRNA half-life changes to SH-SY5Y FMRP targets deriving from siFMRP relative to siControl treatments, related to Figure 2B

NIHMS1849220-supplement-4.xlsx^{(34.4KB, xlsx)}

Table S4: FBI correlation coefficients for mRNA tetra-nucleotide densities, related to Figure 2G

NIHMS1849220-supplement-5.xlsx^{(17.7KB, xlsx)}

Table S5: Quantitative proteomics for WT and FMR1-KO SH-SY5Y cells, related to Figure 3C

NIHMS1849220-supplement-6.xlsx^{(284.5KB, xlsx)}

Table S6: Quantitative RNA-seq for the WT and FMR1-KO SH-SY5Y cells used in SILAC analyses, related to Figure 4 and S4A

NIHMS1849220-supplement-7.xlsx^{(1.5MB, xlsx)}

Table S7: LC-MS/MS after anti-FMRP IP or mouse IgG IP in the presence of RNase I to detect FMRP-interacting proteins in SH-SY5Y cells, related to Figure 6A and 6B

NIHMS1849220-supplement-8.xlsx^{(13.3KB, xlsx)}

Highlights.

FMRP binding to mRNA generally protects that mRNA from translation and decay
To protect, FMRP binds 5'UTR, CDS and/or 3'UTR sequences and poly(A)-bound PABPC1
mRNA protection from decay manifests as protection from poly(A)-tail shortening
FMRP co-immunoprecipitates in an RNase I-resistant complex with granule constituents\

ACKNOWLEDGEMENTS

We thank Manuel Ascano for pFRT/TO/Flag-HA-FMRP WT, Britt Glaunsinger for pcDEF-HA-PABPC1, Keita Miyoshi for purifying EGFP from E. coli, Naoto Imamachi for early-stage computational analyses, and Xavier Rambout, Dmitri Ermolenko and Xin Li for comments on the manuscript. Library generation and sequencing for RNA-seq and anti-FMRP RIP-seq were performed by the University of Rochester Medical Center (URMC) Genomics Core, with special thanks to John Ashton and Cal Palumbo for advice. Mass spectrometry was undertaken by the URMC Mass Spectrometry Resource Lab, with special thanks to Sina Ghaemmaghami and Kevin Welle for advice and funding from NIH grant S10OD025242. This work was supported by NIH R01 GM059614 to L.E.M., FRAXA Research Foundation to L.E.M. and T.K., and MEXT KAKENHI No. 221S0002 to N. A.

Footnotes

DECLARATION OF INTERESTS

The authors declare no competing financial interests.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

REFERENCES

Altimiras F, García JA, Palacios-García I, Hurley MJ, Deacon R, González B, et al. (2021). Altered gut microbiota in a Fragile X syndrome mouse model. Front. Neurosci 15.653120. [DOI] [PMC free article] [PubMed] [Google Scholar]
Anderson BR, Chopra P, Suhl JA, Warren ST, and Bassell GJ (2016). Identification of consensus binding sites clarifies FMRP binding determinants. Nucleic Acids Res. 44, 6649–6659. [DOI] [PMC free article] [PubMed] [Google Scholar]
Antar LN, Dictenberg JB, Plociniak M, Afroz R, and Bassell GJ (2005). Localization of FMRP-associated mRNA granules and requirement of microtubules for activity-dependent trafficking in hippocampal neurons. Genes, Brain Behav. 4, 350–359. [DOI] [PubMed] [Google Scholar]
Ascano M Jr., Mukherjee N, Bandaru P, Miller JB, Nusbaum JD, Corcoran DL, Langlois C, Munschauer M, Dewell S, Hafner M, et al. (2012). FMRP targets distinct mRNA sequence elements to regulate protein expression. Nature 492, 382–386. [DOI] [PMC free article] [PubMed] [Google Scholar]
Bedrat A, Lacroix L, and Mergny JL (2016). Re-evaluation of G-quadruplex propensity with G4Hunter. Nucleic Acids Res. 44, 1746–1759. [DOI] [PMC free article] [PubMed] [Google Scholar]
Berry-Kravis E, Levin R, Shah H, Mathur S, Darnell JC, and Ouyang B (2015). Cholesterol levels in Fragile X syndrome. Am. J. Med. Genet. Part A 167, 379–384. [DOI] [PMC free article] [PubMed] [Google Scholar]
Bowling H, Bhattacharya A, Zhang G, Alam D, Lebowitz JZ, Bohm-levine N, Lin D, Singha P, Mamcarz M, Puckett R, et al. (2019). Altered steady state and activity-dependent de novo protein expression in fragile X syndrome. Nat. Commun 10, 1710. [DOI] [PMC free article] [PubMed] [Google Scholar]
Brown V, Jin P, Ceman S, Darnell J,C, Donnell WTO, Tenenbaum SA, Jin X, Feng Y, Wilkinson KD, Keene JD, et al. (2001). Microarray identification of FMRP-associated brain mRNAs and altered mRNA translational profiles in Fragile X Syndrome. Cell 107, 477–487. [DOI] [PubMed] [Google Scholar]
Bricout VA, Flore P, Eberhard Y, Faure P, Guinot M, and Favre-Juvin A (2008). Maximal and submaximal treadmill tests in a young adult with fragile-X syndrome. Ann. Readapt. Med. Phys 51, 683–691. [DOI] [PubMed] [Google Scholar]
Çaku A, Seidah NG, Lortie A, Gagné N, Perron P, Dubé J, and Corbin F (2017). New insights of altered lipid profile in Fragile X Syndrome. PLoS One 12, e0174301. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ceman S, Donnell WTO, Reed M, Patton S, Pohl J, and Warren ST (2003). Phosphorylation influences the translation state of FMRP-associated polyribosomes. Hum. Mol. Genet 12, 3295–3305. [DOI] [PubMed] [Google Scholar]
Chen E, Sharma MR, Shi X, Agrawal RK, and Joseph S (2014). Fragile X mental retardation protein regulates translation by binding directly to the ribosome. Mol. Cell 54, 407–417. [DOI] [PMC free article] [PubMed] [Google Scholar]
Corbin F, Bouillon M, Fortin A, Morin S, Rousseau F, and Khandjian EW (1997). The fragile X mental retardation protein is associated with poly(A)⁺ mRNA in actively translating polyribosomes. Hum. Mol. Genet 6, 1465–1472. [DOI] [PubMed] [Google Scholar]
Darnell JC, Jensen KB, Jin P, Brown V, Warren ST, and Darnell RB (2001). Fragile X mental retardation protein targets G quartet mRNAs important for neuronal function. Cell 107, 489–499. [DOI] [PubMed] [Google Scholar]
Darnell JC, Mostovetsky O, and Darnell RB (2005). FMRP RNA targets: identification and validation. Genes Brain Behav. 4, 341–349. [DOI] [PubMed] [Google Scholar]
Darnell JC, Van Driesche SJ, Zhang C, Hung KYS, Mele A, Fraser CE, Stone EF, Chen C, Fak JJ, Chi SW, et al. (2011). FMRP stalls ribosomal translocation on mRNAs linked to synaptic function and autism. Cell 146, 247–261. [DOI] [PMC free article] [PubMed] [Google Scholar]
Dionne O and Corbin F (2021). An “Omic” overview of fragile X syndrome. Biology. 10, 433. [DOI] [PMC free article] [PubMed] [Google Scholar]
Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M, and Gingeras TR (2013). STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
Dölen G, Osterweil E, Rao BSS, Smith GB, Auerbach BD, Chattarji S, and Bear MF (2007). Correction of fragile X syndrome in mice. Neuron 56, 955–962. [DOI] [PMC free article] [PubMed] [Google Scholar]
Elbarbary RA, Li W, Tian B, and Maquat LE (2013). STAU1 binding 3'UTR IRAlus complements nuclear retention to protect cells from PKR-mediated translational shutdown. Genes Dev. 27, 1495–1510. [DOI] [PMC free article] [PubMed] [Google Scholar]
Elbarbary RA, Miyoshi K, Hedaya O, Myers JR, and Maquat LE (2017). UPF1 helicase promotes TSN-mediated miRNA decay. Gene Dev 31, 1483–1493. [DOI] [PMC free article] [PubMed] [Google Scholar]
El Fatimy R, Davidovic L, Tremblay S, Jaglin X, Dury A, Robert C, De Koninck P, and Khandjian EW (2016). Tracking the fragile X mental retardation protein in a highly ordered neuronal RiboNucleoParticles population: A link between stalled polyribosomes and RNA granules. PLoS Genet. 12, e1006192. [DOI] [PMC free article] [PubMed] [Google Scholar]
Encinas M, Iglesias M, Liu Y, Wang H, Muhaisen A, Cen V, et al. (2000). Sequential treatment of SH-SY5Y cells with retinoic acid and brain-derived neurotrophic factor gives rise to fully differentiated, neurotrophic factor-dependent, human neuron-like cells. J Neurochem. 75, 991–1003. [DOI] [PubMed] [Google Scholar]
Eng JK, McCormack AL, and Yates JR (1994). An approach to correlate tandem mass spectral data of peptides with amino acid sequences in a protein database. J. Am. Soc. Mass Spectrom 5, 976–989. [DOI] [PubMed] [Google Scholar]
Ermolenko DN, and Mathews DH (2021) Making ends meet: New functions of mRNA secondary structure. Wiley Interdiscip. Rev. RNA 12, e1611. [DOI] [PMC free article] [PubMed] [Google Scholar]
Feng Y, Absher D, Eberhart DE, Brown V, Malter HE, and Warren ST (1997). FMRP associates with polyribosomes as an mRNP, and the I304N mutation of severe fragile X syndrome abolishes this association. Mol. Cell 1, 109–118. [DOI] [PubMed] [Google Scholar]
Fischer JW, Busa VF, Shao Y, and Leung AKL (2020). Structure-mediated RNA decay by UPF1 and G3BP1. Mol. Cell 78, 70–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
Fulks JL, Obryhim BE, Wenzel SK, Fowler SC, Vorontsova E, Pinkston JW, et al. (2010). Dopamine release and uptake impairments and behavioral alterations observed in mice that model fragile X mental retardation syndrome. ACS Chem. Neurosci 1, 679–690. [DOI] [PMC free article] [PubMed] [Google Scholar]
Goering R, Hudish LI, Guzman BB, Raj N, Bassell GJ, Russ HA, Dominguez D, and Taliaferro JM (2020). FMRP promotes RNA localization to neuronal projections through interactions between its RGG domain and G-quadruplex RNA sequences. eLife 9, e52621. [DOI] [PMC free article] [PubMed] [Google Scholar]
Greenblatt EJ, and Spradling AC (2018). Fragile X mental retardation 1 gene enhances the translation of large autism-related proteins. Science 712, 709–712. [DOI] [PMC free article] [PubMed] [Google Scholar]
Hall D, Pickler L, Riley K, Tassone F, and Hagerman R (2011). Parkinsonism and cognitive decline in a Fragile X mosaic male. Mov. Disord 25, 1523–1524. [DOI] [PMC free article] [PubMed] [Google Scholar]
Hagerman RJ, Berry-Kravis E, Hazlett HC, Bailey DB, Moine H, Kooy RF, Tassone F, Gantois I, Sonenberg N, Mandel JL, et al. (2017). Fragile X syndrome. Nat. Rev. Dis. Prim 3, 17065. [DOI] [PubMed] [Google Scholar]
Hofacker IL, Priwitzer B, and Stadler PF (2004). Prediction of locally stable RNA secondary structures for genome-wide surveys. Bioinformatics 20, 186–190. [DOI] [PubMed] [Google Scholar]
Hu Y, Chen Z, Fu Y, He Q, Jiang L, Zheng J, Gao Y, Mei P, Chen Z, and Ren X (2015). The amino-terminal structure of human fragile X mental retardation protein obtained using precipitant-immobilized imprinted polymers. Nat. Commun 6, 6634. [DOI] [PubMed] [Google Scholar]
Ishizuka A, Siomi MC, and Siomi H (2002). A Drosophila fragile X protein interacts with components of RNAi and ribosomal proteins. Genes Dev. 16, 2497–2508. [DOI] [PMC free article] [PubMed] [Google Scholar]
Khandjian EW, Corbin F, Woerly S, and Rousseau F (1996). The fragile X mental retardation protein is associated with ribosomes. Nat. Genet 12, 91–93. [DOI] [PubMed] [Google Scholar]
Kang JY, Wen Z, Pan D, Zhang Y, Li Q, Zhong A, Yu X, Wu YC, Chen Y, Zhang X, Kou PC, Geng J, Wang YY, Hua MM, Zong R, Li B, et al. (2022). LLPS of FXR1 drives spermiogenesis by activating translation of stored mRNAs. Science 377, 6607. [DOI] [PubMed] [Google Scholar]
Kang Y, Zhou Y, Li Y, Han Y, Xu J, Niu W, Li Z, Liu S, Feng H, Huang W, et al. (2021). A human forebrain organoid model of fragile X syndrome exhibits altered neurogenesis and highlights new treatment strategies. Nat. Neurosci 24, 1377–1391. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kikin O, D’Antonio L, and Bagga PS (2006). QGRS Mapper: A web-based server for predicting G-quadruplexes in nucleotide sequences. Nucleic Acids Res. 34, 676–682. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kim TH, Tsang B, Vernon RM, Sonenberg N, Kay LE, and Forman-kay JD (2019). Phospho-dependent phase separation of FMRP and CAPRIN1 recapitulates regulation of translation and deadenylation. Science 829, 825–829. [DOI] [PubMed] [Google Scholar]
Kumar GR, Shum L, and Glaunsinger BA (2011). Importin alpha-mediated nuclear import of cytoplasmic poly(A) binding protein occurs as a direct consequence of cytoplasmic mRNA depletion. Mol. Cell. Biol 31, 3113–3125. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kurosaki T, Miyoshi K, Myers JR, and Maquat LE (2018a). NMD-degradome sequencing reveals ribosome-bound intermediates with 3'-end non-templated nucleotides. Nat. Struct. Mol. Biol 25, 940–950. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kurosaki T, Hoque M, and Maquat LE (2018b). Identifying cellular nonsense-mediated mRNA decay (NMD) targets: Immunoprecipitation of phosphorylated UPF1 followed by RNA sequencing (p-UPF1 RIP-Seq). Methods Mol. Biol 1720, 175–186. [DOI] [PubMed] [Google Scholar]
Kurosaki T, Imamachi N, Pröschel C, Mitsutomi S, Nagao R, Akimitsu N, and Maquat LE (2021a). Loss of the fragile X syndrome protein FMRP results in misregulation of nonsense-mediated mRNA decay. Nat. Cell Biol 23, 40–48. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kurosaki T, Li W, Hoque M, Popp MW, Ermolenko DN, Tian B, and Maquat LE (2014). A post-translational regulatory switch on UPF1 controls targeted mRNA degradation. Genes Dev. 28, 1900–1916. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kurosaki T, Sakano H, Pröschel C, Wheeler J, Hewko A, and Maquat LE (2021b) NMD abnormalities during brain development in the Fmr1-knockout mouse model of fragile X syndrome. Genome Biol. 22, 317. [DOI] [PMC free article] [PubMed] [Google Scholar]
Lai A, Valdez-Sinon AN, and Bassell GJ (2020). Regulation of RNA granules by FMRP and implications for neurological diseases. Traffic 21, 454–462. [DOI] [PMC free article] [PubMed] [Google Scholar]
Leboucher A, Pisani DF, Martinez-Gili L, Chilloux J, Bermudez-Martin P, Van Dijck A, Ganief T, Macek B, Becker JAJ, Le Merrer J, et al. (2019). The translational regulator FMRP controls lipid and glucose metabolism in mice and humans. Mol. Metab 21, 22–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
Leinonen R, Sugawara H, and Shumway M (2011). The sequence read archive. Nucleic Acids Res. 39, 2010–2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
Li M, Shin J, Risgaard RD, Parries MJ, Wang J, Chasman D, Liu S, Roy S, Bhattacharyya A, and Zhao X (2020). Identification of FMR1-regulated molecular networks in human neurodevelopment. Genome Res. 30, 361–374. [DOI] [PMC free article] [PubMed] [Google Scholar]
Liao L, Park SK, Xu T, Vanderklish P, and Yates JR (2008). Quantitative proteomic analysis of primary neurons reveals diverse changes in synaptic protein content in frm1-knockout mice. Proc. Natl. Acad. Sci. U.S.A 105, 15281–15286. [DOI] [PMC free article] [PubMed] [Google Scholar]
Liao Y, Smyth GK, and Shi W (2014). Sequence analysis featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30, 923–930. [DOI] [PubMed] [Google Scholar]
Lisik MZ, Gutmajster E, and Sieroń AL (2016). Low levels of HDL in fragile X syndrome patients. Lipids 51, 189–192. [DOI] [PMC free article] [PubMed] [Google Scholar]
Liu B, Li Y, Stackpole EE, Novak A, Gao Y, Zhao Y, Zhao X, and Richter JD (2018). Regulatory discrimination of mRNAs by FMRP controls mouse adult neural stem cell differentiation. Proc. Natl. Acad. Sci. U.S.A 115, E111397–E11405. [DOI] [PMC free article] [PubMed] [Google Scholar]
Lorenz R, Bernhart SH, Höner Zu, Siederdissen C, Tafer H, Flamm C, Stadler PF, and Hofacker IL (2011). ViennaRNA Package 2.0. Algorithms Mol Biol. 6, 26. [DOI] [PMC free article] [PubMed] [Google Scholar]
Lovci MT, Ghanem D, Marr H, Arnold J, Gee S, Parra M, Liang TY, Stark TJ, Gehman LT, Hoon S, et al. (2013). Rbfox proteins regulate alternative mRNA splicing through evolutionarily conserved RNA bridges. Nat. Struct. Mol. Biol 20, 1434–1442. [DOI] [PMC free article] [PubMed] [Google Scholar]
Lumaban JG, and Nelson DL (2015). The Fragile X proteins Fmrp and Fxr2p cooperate to regulate glucose metabolism in mice. Hum. Mol. Genet 24, 2175–2184. [DOI] [PMC free article] [PubMed] [Google Scholar]
Maquat LE, and Li X (2001). Mammalian heat shock p70 and histone H4 transcripts, which derive from naturally intronless genes, are immune to nonsense-mediated decay. RNA 7, 445–456. [DOI] [PMC free article] [PubMed] [Google Scholar]
Martin M (2011). Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet J. 17, 10–12. [Google Scholar]
Matic K, Eninger T, Bardoni B, Davidovic L, and Macek B (2014). Quantitative phosphoproteomics of murine Fmr1-KO cell lines provides new insights into FMRP-dependent signal transduction mechanisms. J. Proteome Res 13, 4388–4397. [DOI] [PubMed] [Google Scholar]
Maurin T, Lebrigand K, Castagnola S, Paquet A, Jarjat M, Popa A, Grossi M, Rage F, and Bardoni B (2018). HITS-CLIP in various brain areas reveals new targets and new modalities of RNA binding by fragile X mental retardation protein. Nucleic Acids Res. 46, 6344–6355. [DOI] [PMC free article] [PubMed] [Google Scholar]
Mazroui R, Huot ME, Tremblay S, Boilard N, Labelle Y, and Khandjian EW (2003). Fragile X Mental Retardation protein determinants required for its association with polyribosomal mRNPs. Hum. Mol. Genet 12, 3087–3096. [DOI] [PubMed] [Google Scholar]
Mi H, Muruganujan A, and Thomas PD (2013). PANTHER in 2013: Modeling the evolution of gene function, and other gene attributes, in the context of phylogenetic trees. Nucleic Acids Res. 41, 377–386. [DOI] [PMC free article] [PubMed] [Google Scholar]
Musco G, Kharrat A, Stier G, Fraternali F, Gibson Toby, J., Nilges M, and Pastore A (1997). The solution structure of the first KH domain of FMR1, the protein responsible for the fragile X syndrome. Nat. Struct. Biol 4, 712–716. [DOI] [PubMed] [Google Scholar]
Myrick LK, Hashimoto H, Cheng X, and Warren ST (2015). Human FMRP contains an integral tandem Agenet (Tudor) and KH motif in the amino terminal domain. Hum. Mol. Genet 24, 1733–1740. [DOI] [PMC free article] [PubMed] [Google Scholar]
Nalavadi VC, Muddashetty RS, Gross C, and Bassell GJ (2012). Dephosphorylation-induced ubiquitination and degradation of FMRP in dendrites: A role in immediate early mGluR-stimulated translation. J. Neurosci 32, 2582–2587. [DOI] [PMC free article] [PubMed] [Google Scholar]
Napoli I, Mercaldo V, Boyl PP, Eleuteri B, Zalfa F, De Rubeis S, Di Marino D, Mohr E, Massimi M, Falconi M, et al. (2008). The fragile X syndrome protein represses activity-dependent translation through CYFIP1, a new 4E-BP. Cell 134, 1042–1054. [DOI] [PubMed] [Google Scholar]
Narayanan U, Nalavadi V, Nakamoto M, Pallas DC, Ceman S, Bassell GJ, Warren ST (2007). FMRP phosphorylation reveals an immediate-early signaling pathway triggered by group I mGluR and mediated by PP2A. J. Neurosci 27, 14349–14357. [DOI] [PMC free article] [PubMed] [Google Scholar]
Otero DDY, Severijnen L-A, van Cappellen G, Schrier M, Oostra B, and Willemsen R (2002). Transport of fragile X mental retardation protein via granules in neurites of PC12 cells. Mol. Cell. Biol 22, 8332–8341. [DOI] [PMC free article] [PubMed] [Google Scholar]
Park E, Gleghorn ML, and Maquat LE (2013). Staufen2 functions in Staufenl-mediated mRNA decay by binding to itself and its paralog and promoting UPF1 helicase but not ATPase activity. Proc. Natl. Acad. Sci. USA 110, 405–412. [DOI] [PMC free article] [PubMed] [Google Scholar]
Pivesan A, Caracausi M, Antonaros F, Pelleri MC, and Vitale L (2016). GeneBase 1.1: A tool to summarize data from NCBI gene datasets and its application to an update of human gene statistics. Database 2016, baw153. [DOI] [PMC free article] [PubMed] [Google Scholar]
Qin M, Kang J, Burlin TV, Jiang C, and Smith CB (2005). Postadolescent changes in regional cerebral protein synthesis: An in vivo study in the Fmr1 null mouse. J. Neurosci 25, 5087–5095. [DOI] [PMC free article] [PubMed] [Google Scholar]
Quinlan AR, and Hall IM (2010). BEDTools: A flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841–842. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ramos A, Hollingworth D, and Pastore A (2003). G-quartet-dependent recognition between the FMRP RGG box and RNA. RNA 9, 1198–1207. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ray D, Kazan H, Cook KB, Weirauch MT, Najafabadi HS, Li X, Gueroussov S, Albu M, Zheng H, Yang A, et al. (2013). A compendium of RNA-binding motifs for decoding gene regulation. Nature 499, 172–127. [DOI] [PMC free article] [PubMed] [Google Scholar]
Richter JD, and Zhao X (2021). The molecular biology of FMRP: new insights into fragile X syndrome. Nat. Rev. Neurosci 22, 209–222. [DOI] [PMC free article] [PubMed] [Google Scholar]
Robinson MD, McCarthy DJ, and Smyth GK (2010). edgeR: A Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–140. [DOI] [PMC free article] [PubMed] [Google Scholar]
Rosario R, Filis P, Tessyman V, Kinnell H, Childs AJ, Gray NK, and Anderson RA (2016). FMRP associates with cytoplasmic granules at the onset of meiosis in the human oocyte. PLoS One 11, e0163987. [DOI] [PMC free article] [PubMed] [Google Scholar]
Sahoo PK, Lee SJ, Jaiswal PB, Alber S, Kar AN, Miller-Randolph S, Taylor EE, Smith T, Singh B, Ho TS, Urisman A, Chand S, Pena EA, Burlingame AL, Woolf CJ, Fainzilber M, English AW, Twiss JL. (2018). Axonal G3BP1 stress granule protein limits axonal mRNA translation and nerve regeneration. Nat Commun. 9, 3358. [DOI] [PMC free article] [PubMed] [Google Scholar]
Sawicka K, Hale CR, Park CY, Fak JJ, Gresack JE, Van Driesche SJ, Kang JJ, Darnell JC, and Darnell RB (2019). FMRP has a cell-type-specific role in CA1 pyramidal neurons to regulate autism-related transcripts and circadian memory. eLife 8, e46919. [DOI] [PMC free article] [PubMed] [Google Scholar]
Schaeffer C, Bardoni B, Mandel JL, Ehresmann B, Ehresmann C, and Moine H (2001). The fragile X mental retardation protein binds specifically to its mRNA via a purine quartet motif. EMBO J. 20, 4803–4813. [DOI] [PMC free article] [PubMed] [Google Scholar]
Sharp PM, and Li W-H (1987). The codon adaptation index-a measure of directional synonymous codon usage bias, and its potential applications. Nucleic Acids Res. 15, 1281–1295. [DOI] [PMC free article] [PubMed] [Google Scholar]
Shah S, Molinaro G, Liu B, Wang R, Huber KM, and Richter JD (2020). FMRP control of ribosome translocation promotes chromatin modifications and alternative splicing of neuronal genes linked to autism. Cell Rep. 30, 4459–4472. [DOI] [PMC free article] [PubMed] [Google Scholar]
Shannon P, Markiel A, Ozier O, Baliga NS, Jonathan T Wang DR, Amin N, Schwikowski B, and Ideker T (2003). Cytoscape: A software environment for integrated models. Genome Res. 13, 2498–2504. [DOI] [PMC free article] [PubMed] [Google Scholar]
Signorell A, Aho K, Alfons A, Anderegg N, Aragon T, Arachchige C, Arppe A, Baddeley A, Barton K, Bolker B, et al. (2021). DescTools: Tools for descriptive statistics. R package version 0.99.44 [Google Scholar]
Shu H, Donnard E, Liu B, Jung S, Wang R, and Richter JD (2020). FMRP links optimal codons to mRNA stability in neurons. Proc. Natl. Acad. Sci. USA 117, 30400–30411. [DOI] [PMC free article] [PubMed] [Google Scholar]
Singh G, Ricci EP, and Moore MJ (2014). RIPiT-Seq: a high-throughput approach for footprinting RNA:protein complexes. Methods 65, 320–332. [DOI] [PMC free article] [PubMed] [Google Scholar]
Stefani G, Fraser CE, Darnell JC, and Darnell RB (2004). Fragile X mental retardation protein Is associated with translating polyribosomes in neuronal cells. J. Neurosci 24, 7272–7276. [DOI] [PMC free article] [PubMed] [Google Scholar]
Subramanian M, Rage F, Tabet R, Flatter E, Mandel JL, and Moine H (2011). G-quadruplex RNA structure as a signal for neurite mRNA targeting. EMBO Rep. 12, 697–704. [DOI] [PMC free article] [PubMed] [Google Scholar]
Suhl JA, Chopra P, Anderson BR, Bassell GJ, and Warren ST (2014). Analysis of FMRP mRNA target datasets reveals highly associated mRNAs mediated by G-quadruplex structures formed via clustered WGGA sequences. Hum. Mol. Genet 23, 5479–5491. [DOI] [PMC free article] [PubMed] [Google Scholar]
Tabet R, Moutin E, Becker JAJ, Heintz D, Fouillen L, Flatter E, Krȩzel W, Alunni V, Koebel P, Dembélé D, et al. (2016). Fragile X mental retardation protein (FMRP) controls diacylglycerol kinase activity in neurons. Proc. Natl. Acad. Sci. USA 113, E3619–E3628. [DOI] [PMC free article] [PubMed] [Google Scholar]
Taha MS, Haghighi F, Stefanski A, Nakhaei-Rad S, Kazemein Jasemi NS, Al Kabbani MA, Görg B, Fujii M, Lang PA, Häussinger D, et al. (2021). Novel FMRP interaction networks linked to cellular stress. FEBS J. 288, 837–860. [DOI] [PubMed] [Google Scholar]
Tang B, Wang T, Wan H, Han L, Qin X, Zhang Y, Wang J, Yu C, Berton F, Francesconi W, et al. (2015). Fmr1 deficiency promotes age-dependent alterations in the cortical synaptic proteome. Proc. Natl. Acad. Sci. U.S.A 112, E4697–706. [DOI] [PMC free article] [PubMed] [Google Scholar]
Tran SS, Jun H, Bahn JH, Azghadi A, Ramaswami G, Van Nostrand EL, Nguyen TB, Hsiao YE, Lee C, Pratt GA, et al. (2019). Widespread RNA editing dysregulation in brains from autistic individuals. Nat. Neurosci 22, 25–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
Tsang B, Arsenault J, Vernon RM, Lin H, Sonenberg N, Wang LY, Bah A, and Forman-Kay JD (2019). Phosphoregulated FMRP phase separation models activity-dependent translation through bidirectional control of mRNA granule formation. Proc. Natl. Acad. Sci. USA 116, 4218–4227. [DOI] [PMC free article] [PubMed] [Google Scholar]
Van Nostrand EL, Pratt GA, Shishkin AA, Gelboin-Burkhart C, Fang MY, Sundararaman B, Blue SM, Nguyen TB, Surka C, Elkins K, et al. (2016). Robust transcriptome-wide discovery of RNA-binding protein binding sites with enhanced CLIP (eCLIP). Nat. Methods 13, 508–514. [DOI] [PMC free article] [PubMed] [Google Scholar]
Van Nostrand EL, Pratt GA, Yee BA, Wheeler EC, Blue SM, Mueller J, Park SS, Garcia KE, Gelboin-Burkhart C, Nguyen TB, et al. (2020). Principles of RNA processing from analysis of enhanced CLIP maps for 150 RNA binding proteins. Genome Biol. 21, 90. [DOI] [PMC free article] [PubMed] [Google Scholar]
Vasilyev N, Polonskaia A, Darnell JC, Darnell RB, Patel DJ, and Serganov A (2015). Crystal structure reveals specific recognition of a G-quadruplex RNA by a β-turn in the RGG motif of FMRP. Proc. Natl. Acad. Sci. USA 112, E5391–5400. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wheeler EC, Van Nostrand EL, and Yeo GW (2018). Advances and challenges in the detection of transcriptome-wide protein-RNA interactions. WIRES RNA 9, 1436. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wickham H (2010). Stringr: Modern, consistent string processing. R J. 2, 38–40. [Google Scholar]
Yang L, Duff MO, Graveley BR, Carmichael GG, and Chen LL (2011). Genomewide characterization of non-polyadenylated RNAs. Genome Biol. 12, R16. [DOI] [PMC free article] [PubMed] [Google Scholar]
Yi H, Park J, Ha M, Lim J, Chang H, and Kim VN (2018). PABP cooperates with the CCR4-NOT complex to promote mRNA deadenylation and block precocious decay. Mol. Cell 70, 1081–1088. [DOI] [PubMed] [Google Scholar]
Youn J, Dunham WH, Hong SJ, Knight JDR, Bashkurov M, Chen GI, Bagci H, Rathod B, MacLeod G, Eng SW, et al. (2018). High-density proximity mapping reveals the subcellular organization of mRNA-associated granules and bodies. Mol Cell 69, 517–532. [DOI] [PubMed] [Google Scholar]
Zhao J, Qin B, Nikolay R, Spahn CMT, and Zhang G (2019). Translatomics: The global view of translation. Int. J. Mol. Sci 20, 212. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

NIHMS1849220-supplement-1.pdf^{(2.1MB, pdf)}

Table S1: FMRP targets in human SH-SY5Y cells, related to Figure 1 and S1G

NIHMS1849220-supplement-2.xlsx^{(47.5KB, xlsx)}

Table S2: Not FMRP targets in human SH-SY5Y cells, related to Figure 1 and S1G

NIHMS1849220-supplement-3.xlsx^{(68KB, xlsx)}

Table S3: mRNA half-life changes to SH-SY5Y FMRP targets deriving from siFMRP relative to siControl treatments, related to Figure 2B

NIHMS1849220-supplement-4.xlsx^{(34.4KB, xlsx)}

Table S4: FBI correlation coefficients for mRNA tetra-nucleotide densities, related to Figure 2G

NIHMS1849220-supplement-5.xlsx^{(17.7KB, xlsx)}

Table S5: Quantitative proteomics for WT and FMR1-KO SH-SY5Y cells, related to Figure 3C

NIHMS1849220-supplement-6.xlsx^{(284.5KB, xlsx)}

Table S6: Quantitative RNA-seq for the WT and FMR1-KO SH-SY5Y cells used in SILAC analyses, related to Figure 4 and S4A

NIHMS1849220-supplement-7.xlsx^{(1.5MB, xlsx)}

Table S7: LC-MS/MS after anti-FMRP IP or mouse IgG IP in the presence of RNase I to detect FMRP-interacting proteins in SH-SY5Y cells, related to Figure 6A and 6B

NIHMS1849220-supplement-8.xlsx^{(13.3KB, xlsx)}

Data Availability Statement

Sequencing datasets (FASTQ files), including TRIC-seq and anti-FMRP RIP-seq (Kurosaki et al., 2021a), have been deposited in the DNA Data Bank of Japan (DDBJ) Sequence Read Archive: DRA005644. Datasets from RNA-seq used for SILAC–LC-MS/MS analysis are available through the Gene Expression Omnibus (GEO): GSE197221. Proteomics data obtained using SILAC–LC-MS/MS or α-FMRP IP have been deposited in the ProteomeXchange Consortium: PXD032341 or PXD032339, respectively. Unprocessed and uncompressed imaging data are available in Mendeley Data: https://dx.doi.org/10.17632/5k773sxkcw.1.
This paper does not report original code.
Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

[R1] Altimiras F, García JA, Palacios-García I, Hurley MJ, Deacon R, González B, et al. (2021). Altered gut microbiota in a Fragile X syndrome mouse model. Front. Neurosci 15.653120. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R2] Anderson BR, Chopra P, Suhl JA, Warren ST, and Bassell GJ (2016). Identification of consensus binding sites clarifies FMRP binding determinants. Nucleic Acids Res. 44, 6649–6659. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R3] Antar LN, Dictenberg JB, Plociniak M, Afroz R, and Bassell GJ (2005). Localization of FMRP-associated mRNA granules and requirement of microtubules for activity-dependent trafficking in hippocampal neurons. Genes, Brain Behav. 4, 350–359. [DOI] [PubMed] [Google Scholar]

[R4] Ascano M Jr., Mukherjee N, Bandaru P, Miller JB, Nusbaum JD, Corcoran DL, Langlois C, Munschauer M, Dewell S, Hafner M, et al. (2012). FMRP targets distinct mRNA sequence elements to regulate protein expression. Nature 492, 382–386. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R5] Bedrat A, Lacroix L, and Mergny JL (2016). Re-evaluation of G-quadruplex propensity with G4Hunter. Nucleic Acids Res. 44, 1746–1759. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R6] Berry-Kravis E, Levin R, Shah H, Mathur S, Darnell JC, and Ouyang B (2015). Cholesterol levels in Fragile X syndrome. Am. J. Med. Genet. Part A 167, 379–384. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R7] Bowling H, Bhattacharya A, Zhang G, Alam D, Lebowitz JZ, Bohm-levine N, Lin D, Singha P, Mamcarz M, Puckett R, et al. (2019). Altered steady state and activity-dependent de novo protein expression in fragile X syndrome. Nat. Commun 10, 1710. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R8] Brown V, Jin P, Ceman S, Darnell J,C, Donnell WTO, Tenenbaum SA, Jin X, Feng Y, Wilkinson KD, Keene JD, et al. (2001). Microarray identification of FMRP-associated brain mRNAs and altered mRNA translational profiles in Fragile X Syndrome. Cell 107, 477–487. [DOI] [PubMed] [Google Scholar]

[R9] Bricout VA, Flore P, Eberhard Y, Faure P, Guinot M, and Favre-Juvin A (2008). Maximal and submaximal treadmill tests in a young adult with fragile-X syndrome. Ann. Readapt. Med. Phys 51, 683–691. [DOI] [PubMed] [Google Scholar]

[R10] Çaku A, Seidah NG, Lortie A, Gagné N, Perron P, Dubé J, and Corbin F (2017). New insights of altered lipid profile in Fragile X Syndrome. PLoS One 12, e0174301. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R11] Ceman S, Donnell WTO, Reed M, Patton S, Pohl J, and Warren ST (2003). Phosphorylation influences the translation state of FMRP-associated polyribosomes. Hum. Mol. Genet 12, 3295–3305. [DOI] [PubMed] [Google Scholar]

[R12] Chen E, Sharma MR, Shi X, Agrawal RK, and Joseph S (2014). Fragile X mental retardation protein regulates translation by binding directly to the ribosome. Mol. Cell 54, 407–417. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R13] Corbin F, Bouillon M, Fortin A, Morin S, Rousseau F, and Khandjian EW (1997). The fragile X mental retardation protein is associated with poly(A)⁺ mRNA in actively translating polyribosomes. Hum. Mol. Genet 6, 1465–1472. [DOI] [PubMed] [Google Scholar]

[R14] Darnell JC, Jensen KB, Jin P, Brown V, Warren ST, and Darnell RB (2001). Fragile X mental retardation protein targets G quartet mRNAs important for neuronal function. Cell 107, 489–499. [DOI] [PubMed] [Google Scholar]

[R15] Darnell JC, Mostovetsky O, and Darnell RB (2005). FMRP RNA targets: identification and validation. Genes Brain Behav. 4, 341–349. [DOI] [PubMed] [Google Scholar]

[R16] Darnell JC, Van Driesche SJ, Zhang C, Hung KYS, Mele A, Fraser CE, Stone EF, Chen C, Fak JJ, Chi SW, et al. (2011). FMRP stalls ribosomal translocation on mRNAs linked to synaptic function and autism. Cell 146, 247–261. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R17] Dionne O and Corbin F (2021). An “Omic” overview of fragile X syndrome. Biology. 10, 433. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R18] Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M, and Gingeras TR (2013). STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R19] Dölen G, Osterweil E, Rao BSS, Smith GB, Auerbach BD, Chattarji S, and Bear MF (2007). Correction of fragile X syndrome in mice. Neuron 56, 955–962. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R20] Elbarbary RA, Li W, Tian B, and Maquat LE (2013). STAU1 binding 3'UTR IRAlus complements nuclear retention to protect cells from PKR-mediated translational shutdown. Genes Dev. 27, 1495–1510. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R21] Elbarbary RA, Miyoshi K, Hedaya O, Myers JR, and Maquat LE (2017). UPF1 helicase promotes TSN-mediated miRNA decay. Gene Dev 31, 1483–1493. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R22] El Fatimy R, Davidovic L, Tremblay S, Jaglin X, Dury A, Robert C, De Koninck P, and Khandjian EW (2016). Tracking the fragile X mental retardation protein in a highly ordered neuronal RiboNucleoParticles population: A link between stalled polyribosomes and RNA granules. PLoS Genet. 12, e1006192. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R23] Encinas M, Iglesias M, Liu Y, Wang H, Muhaisen A, Cen V, et al. (2000). Sequential treatment of SH-SY5Y cells with retinoic acid and brain-derived neurotrophic factor gives rise to fully differentiated, neurotrophic factor-dependent, human neuron-like cells. J Neurochem. 75, 991–1003. [DOI] [PubMed] [Google Scholar]

[R24] Eng JK, McCormack AL, and Yates JR (1994). An approach to correlate tandem mass spectral data of peptides with amino acid sequences in a protein database. J. Am. Soc. Mass Spectrom 5, 976–989. [DOI] [PubMed] [Google Scholar]

[R25] Ermolenko DN, and Mathews DH (2021) Making ends meet: New functions of mRNA secondary structure. Wiley Interdiscip. Rev. RNA 12, e1611. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R26] Feng Y, Absher D, Eberhart DE, Brown V, Malter HE, and Warren ST (1997). FMRP associates with polyribosomes as an mRNP, and the I304N mutation of severe fragile X syndrome abolishes this association. Mol. Cell 1, 109–118. [DOI] [PubMed] [Google Scholar]

[R27] Fischer JW, Busa VF, Shao Y, and Leung AKL (2020). Structure-mediated RNA decay by UPF1 and G3BP1. Mol. Cell 78, 70–84. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R28] Fulks JL, Obryhim BE, Wenzel SK, Fowler SC, Vorontsova E, Pinkston JW, et al. (2010). Dopamine release and uptake impairments and behavioral alterations observed in mice that model fragile X mental retardation syndrome. ACS Chem. Neurosci 1, 679–690. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R29] Goering R, Hudish LI, Guzman BB, Raj N, Bassell GJ, Russ HA, Dominguez D, and Taliaferro JM (2020). FMRP promotes RNA localization to neuronal projections through interactions between its RGG domain and G-quadruplex RNA sequences. eLife 9, e52621. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R30] Greenblatt EJ, and Spradling AC (2018). Fragile X mental retardation 1 gene enhances the translation of large autism-related proteins. Science 712, 709–712. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R31] Hall D, Pickler L, Riley K, Tassone F, and Hagerman R (2011). Parkinsonism and cognitive decline in a Fragile X mosaic male. Mov. Disord 25, 1523–1524. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R32] Hagerman RJ, Berry-Kravis E, Hazlett HC, Bailey DB, Moine H, Kooy RF, Tassone F, Gantois I, Sonenberg N, Mandel JL, et al. (2017). Fragile X syndrome. Nat. Rev. Dis. Prim 3, 17065. [DOI] [PubMed] [Google Scholar]

[R33] Hofacker IL, Priwitzer B, and Stadler PF (2004). Prediction of locally stable RNA secondary structures for genome-wide surveys. Bioinformatics 20, 186–190. [DOI] [PubMed] [Google Scholar]

[R34] Hu Y, Chen Z, Fu Y, He Q, Jiang L, Zheng J, Gao Y, Mei P, Chen Z, and Ren X (2015). The amino-terminal structure of human fragile X mental retardation protein obtained using precipitant-immobilized imprinted polymers. Nat. Commun 6, 6634. [DOI] [PubMed] [Google Scholar]

[R35] Ishizuka A, Siomi MC, and Siomi H (2002). A Drosophila fragile X protein interacts with components of RNAi and ribosomal proteins. Genes Dev. 16, 2497–2508. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R36] Khandjian EW, Corbin F, Woerly S, and Rousseau F (1996). The fragile X mental retardation protein is associated with ribosomes. Nat. Genet 12, 91–93. [DOI] [PubMed] [Google Scholar]

[R37] Kang JY, Wen Z, Pan D, Zhang Y, Li Q, Zhong A, Yu X, Wu YC, Chen Y, Zhang X, Kou PC, Geng J, Wang YY, Hua MM, Zong R, Li B, et al. (2022). LLPS of FXR1 drives spermiogenesis by activating translation of stored mRNAs. Science 377, 6607. [DOI] [PubMed] [Google Scholar]

[R38] Kang Y, Zhou Y, Li Y, Han Y, Xu J, Niu W, Li Z, Liu S, Feng H, Huang W, et al. (2021). A human forebrain organoid model of fragile X syndrome exhibits altered neurogenesis and highlights new treatment strategies. Nat. Neurosci 24, 1377–1391. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R39] Kikin O, D’Antonio L, and Bagga PS (2006). QGRS Mapper: A web-based server for predicting G-quadruplexes in nucleotide sequences. Nucleic Acids Res. 34, 676–682. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R40] Kim TH, Tsang B, Vernon RM, Sonenberg N, Kay LE, and Forman-kay JD (2019). Phospho-dependent phase separation of FMRP and CAPRIN1 recapitulates regulation of translation and deadenylation. Science 829, 825–829. [DOI] [PubMed] [Google Scholar]

[R41] Kumar GR, Shum L, and Glaunsinger BA (2011). Importin alpha-mediated nuclear import of cytoplasmic poly(A) binding protein occurs as a direct consequence of cytoplasmic mRNA depletion. Mol. Cell. Biol 31, 3113–3125. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R42] Kurosaki T, Miyoshi K, Myers JR, and Maquat LE (2018a). NMD-degradome sequencing reveals ribosome-bound intermediates with 3'-end non-templated nucleotides. Nat. Struct. Mol. Biol 25, 940–950. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R43] Kurosaki T, Hoque M, and Maquat LE (2018b). Identifying cellular nonsense-mediated mRNA decay (NMD) targets: Immunoprecipitation of phosphorylated UPF1 followed by RNA sequencing (p-UPF1 RIP-Seq). Methods Mol. Biol 1720, 175–186. [DOI] [PubMed] [Google Scholar]

[R44] Kurosaki T, Imamachi N, Pröschel C, Mitsutomi S, Nagao R, Akimitsu N, and Maquat LE (2021a). Loss of the fragile X syndrome protein FMRP results in misregulation of nonsense-mediated mRNA decay. Nat. Cell Biol 23, 40–48. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R45] Kurosaki T, Li W, Hoque M, Popp MW, Ermolenko DN, Tian B, and Maquat LE (2014). A post-translational regulatory switch on UPF1 controls targeted mRNA degradation. Genes Dev. 28, 1900–1916. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R46] Kurosaki T, Sakano H, Pröschel C, Wheeler J, Hewko A, and Maquat LE (2021b) NMD abnormalities during brain development in the Fmr1-knockout mouse model of fragile X syndrome. Genome Biol. 22, 317. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R47] Lai A, Valdez-Sinon AN, and Bassell GJ (2020). Regulation of RNA granules by FMRP and implications for neurological diseases. Traffic 21, 454–462. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R48] Leboucher A, Pisani DF, Martinez-Gili L, Chilloux J, Bermudez-Martin P, Van Dijck A, Ganief T, Macek B, Becker JAJ, Le Merrer J, et al. (2019). The translational regulator FMRP controls lipid and glucose metabolism in mice and humans. Mol. Metab 21, 22–35. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R49] Leinonen R, Sugawara H, and Shumway M (2011). The sequence read archive. Nucleic Acids Res. 39, 2010–2012. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R50] Li M, Shin J, Risgaard RD, Parries MJ, Wang J, Chasman D, Liu S, Roy S, Bhattacharyya A, and Zhao X (2020). Identification of FMR1-regulated molecular networks in human neurodevelopment. Genome Res. 30, 361–374. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R51] Liao L, Park SK, Xu T, Vanderklish P, and Yates JR (2008). Quantitative proteomic analysis of primary neurons reveals diverse changes in synaptic protein content in frm1-knockout mice. Proc. Natl. Acad. Sci. U.S.A 105, 15281–15286. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R52] Liao Y, Smyth GK, and Shi W (2014). Sequence analysis featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30, 923–930. [DOI] [PubMed] [Google Scholar]

[R53] Lisik MZ, Gutmajster E, and Sieroń AL (2016). Low levels of HDL in fragile X syndrome patients. Lipids 51, 189–192. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R54] Liu B, Li Y, Stackpole EE, Novak A, Gao Y, Zhao Y, Zhao X, and Richter JD (2018). Regulatory discrimination of mRNAs by FMRP controls mouse adult neural stem cell differentiation. Proc. Natl. Acad. Sci. U.S.A 115, E111397–E11405. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R55] Lorenz R, Bernhart SH, Höner Zu, Siederdissen C, Tafer H, Flamm C, Stadler PF, and Hofacker IL (2011). ViennaRNA Package 2.0. Algorithms Mol Biol. 6, 26. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R56] Lovci MT, Ghanem D, Marr H, Arnold J, Gee S, Parra M, Liang TY, Stark TJ, Gehman LT, Hoon S, et al. (2013). Rbfox proteins regulate alternative mRNA splicing through evolutionarily conserved RNA bridges. Nat. Struct. Mol. Biol 20, 1434–1442. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R57] Lumaban JG, and Nelson DL (2015). The Fragile X proteins Fmrp and Fxr2p cooperate to regulate glucose metabolism in mice. Hum. Mol. Genet 24, 2175–2184. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R58] Maquat LE, and Li X (2001). Mammalian heat shock p70 and histone H4 transcripts, which derive from naturally intronless genes, are immune to nonsense-mediated decay. RNA 7, 445–456. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R59] Martin M (2011). Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet J. 17, 10–12. [Google Scholar]

[R60] Matic K, Eninger T, Bardoni B, Davidovic L, and Macek B (2014). Quantitative phosphoproteomics of murine Fmr1-KO cell lines provides new insights into FMRP-dependent signal transduction mechanisms. J. Proteome Res 13, 4388–4397. [DOI] [PubMed] [Google Scholar]

[R61] Maurin T, Lebrigand K, Castagnola S, Paquet A, Jarjat M, Popa A, Grossi M, Rage F, and Bardoni B (2018). HITS-CLIP in various brain areas reveals new targets and new modalities of RNA binding by fragile X mental retardation protein. Nucleic Acids Res. 46, 6344–6355. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R62] Mazroui R, Huot ME, Tremblay S, Boilard N, Labelle Y, and Khandjian EW (2003). Fragile X Mental Retardation protein determinants required for its association with polyribosomal mRNPs. Hum. Mol. Genet 12, 3087–3096. [DOI] [PubMed] [Google Scholar]

[R63] Mi H, Muruganujan A, and Thomas PD (2013). PANTHER in 2013: Modeling the evolution of gene function, and other gene attributes, in the context of phylogenetic trees. Nucleic Acids Res. 41, 377–386. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R64] Musco G, Kharrat A, Stier G, Fraternali F, Gibson Toby, J., Nilges M, and Pastore A (1997). The solution structure of the first KH domain of FMR1, the protein responsible for the fragile X syndrome. Nat. Struct. Biol 4, 712–716. [DOI] [PubMed] [Google Scholar]

[R65] Myrick LK, Hashimoto H, Cheng X, and Warren ST (2015). Human FMRP contains an integral tandem Agenet (Tudor) and KH motif in the amino terminal domain. Hum. Mol. Genet 24, 1733–1740. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R66] Nalavadi VC, Muddashetty RS, Gross C, and Bassell GJ (2012). Dephosphorylation-induced ubiquitination and degradation of FMRP in dendrites: A role in immediate early mGluR-stimulated translation. J. Neurosci 32, 2582–2587. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R67] Napoli I, Mercaldo V, Boyl PP, Eleuteri B, Zalfa F, De Rubeis S, Di Marino D, Mohr E, Massimi M, Falconi M, et al. (2008). The fragile X syndrome protein represses activity-dependent translation through CYFIP1, a new 4E-BP. Cell 134, 1042–1054. [DOI] [PubMed] [Google Scholar]

[R68] Narayanan U, Nalavadi V, Nakamoto M, Pallas DC, Ceman S, Bassell GJ, Warren ST (2007). FMRP phosphorylation reveals an immediate-early signaling pathway triggered by group I mGluR and mediated by PP2A. J. Neurosci 27, 14349–14357. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R69] Otero DDY, Severijnen L-A, van Cappellen G, Schrier M, Oostra B, and Willemsen R (2002). Transport of fragile X mental retardation protein via granules in neurites of PC12 cells. Mol. Cell. Biol 22, 8332–8341. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R70] Park E, Gleghorn ML, and Maquat LE (2013). Staufen2 functions in Staufenl-mediated mRNA decay by binding to itself and its paralog and promoting UPF1 helicase but not ATPase activity. Proc. Natl. Acad. Sci. USA 110, 405–412. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R71] Pivesan A, Caracausi M, Antonaros F, Pelleri MC, and Vitale L (2016). GeneBase 1.1: A tool to summarize data from NCBI gene datasets and its application to an update of human gene statistics. Database 2016, baw153. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R72] Qin M, Kang J, Burlin TV, Jiang C, and Smith CB (2005). Postadolescent changes in regional cerebral protein synthesis: An in vivo study in the Fmr1 null mouse. J. Neurosci 25, 5087–5095. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R73] Quinlan AR, and Hall IM (2010). BEDTools: A flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841–842. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R74] Ramos A, Hollingworth D, and Pastore A (2003). G-quartet-dependent recognition between the FMRP RGG box and RNA. RNA 9, 1198–1207. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R75] Ray D, Kazan H, Cook KB, Weirauch MT, Najafabadi HS, Li X, Gueroussov S, Albu M, Zheng H, Yang A, et al. (2013). A compendium of RNA-binding motifs for decoding gene regulation. Nature 499, 172–127. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R76] Richter JD, and Zhao X (2021). The molecular biology of FMRP: new insights into fragile X syndrome. Nat. Rev. Neurosci 22, 209–222. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R77] Robinson MD, McCarthy DJ, and Smyth GK (2010). edgeR: A Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–140. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R78] Rosario R, Filis P, Tessyman V, Kinnell H, Childs AJ, Gray NK, and Anderson RA (2016). FMRP associates with cytoplasmic granules at the onset of meiosis in the human oocyte. PLoS One 11, e0163987. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R79] Sahoo PK, Lee SJ, Jaiswal PB, Alber S, Kar AN, Miller-Randolph S, Taylor EE, Smith T, Singh B, Ho TS, Urisman A, Chand S, Pena EA, Burlingame AL, Woolf CJ, Fainzilber M, English AW, Twiss JL. (2018). Axonal G3BP1 stress granule protein limits axonal mRNA translation and nerve regeneration. Nat Commun. 9, 3358. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R80] Sawicka K, Hale CR, Park CY, Fak JJ, Gresack JE, Van Driesche SJ, Kang JJ, Darnell JC, and Darnell RB (2019). FMRP has a cell-type-specific role in CA1 pyramidal neurons to regulate autism-related transcripts and circadian memory. eLife 8, e46919. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R81] Schaeffer C, Bardoni B, Mandel JL, Ehresmann B, Ehresmann C, and Moine H (2001). The fragile X mental retardation protein binds specifically to its mRNA via a purine quartet motif. EMBO J. 20, 4803–4813. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R82] Sharp PM, and Li W-H (1987). The codon adaptation index-a measure of directional synonymous codon usage bias, and its potential applications. Nucleic Acids Res. 15, 1281–1295. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R83] Shah S, Molinaro G, Liu B, Wang R, Huber KM, and Richter JD (2020). FMRP control of ribosome translocation promotes chromatin modifications and alternative splicing of neuronal genes linked to autism. Cell Rep. 30, 4459–4472. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R84] Shannon P, Markiel A, Ozier O, Baliga NS, Jonathan T Wang DR, Amin N, Schwikowski B, and Ideker T (2003). Cytoscape: A software environment for integrated models. Genome Res. 13, 2498–2504. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R85] Signorell A, Aho K, Alfons A, Anderegg N, Aragon T, Arachchige C, Arppe A, Baddeley A, Barton K, Bolker B, et al. (2021). DescTools: Tools for descriptive statistics. R package version 0.99.44 [Google Scholar]

[R86] Shu H, Donnard E, Liu B, Jung S, Wang R, and Richter JD (2020). FMRP links optimal codons to mRNA stability in neurons. Proc. Natl. Acad. Sci. USA 117, 30400–30411. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R87] Singh G, Ricci EP, and Moore MJ (2014). RIPiT-Seq: a high-throughput approach for footprinting RNA:protein complexes. Methods 65, 320–332. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R88] Stefani G, Fraser CE, Darnell JC, and Darnell RB (2004). Fragile X mental retardation protein Is associated with translating polyribosomes in neuronal cells. J. Neurosci 24, 7272–7276. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R89] Subramanian M, Rage F, Tabet R, Flatter E, Mandel JL, and Moine H (2011). G-quadruplex RNA structure as a signal for neurite mRNA targeting. EMBO Rep. 12, 697–704. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R90] Suhl JA, Chopra P, Anderson BR, Bassell GJ, and Warren ST (2014). Analysis of FMRP mRNA target datasets reveals highly associated mRNAs mediated by G-quadruplex structures formed via clustered WGGA sequences. Hum. Mol. Genet 23, 5479–5491. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R91] Tabet R, Moutin E, Becker JAJ, Heintz D, Fouillen L, Flatter E, Krȩzel W, Alunni V, Koebel P, Dembélé D, et al. (2016). Fragile X mental retardation protein (FMRP) controls diacylglycerol kinase activity in neurons. Proc. Natl. Acad. Sci. USA 113, E3619–E3628. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R92] Taha MS, Haghighi F, Stefanski A, Nakhaei-Rad S, Kazemein Jasemi NS, Al Kabbani MA, Görg B, Fujii M, Lang PA, Häussinger D, et al. (2021). Novel FMRP interaction networks linked to cellular stress. FEBS J. 288, 837–860. [DOI] [PubMed] [Google Scholar]

[R93] Tang B, Wang T, Wan H, Han L, Qin X, Zhang Y, Wang J, Yu C, Berton F, Francesconi W, et al. (2015). Fmr1 deficiency promotes age-dependent alterations in the cortical synaptic proteome. Proc. Natl. Acad. Sci. U.S.A 112, E4697–706. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R94] Tran SS, Jun H, Bahn JH, Azghadi A, Ramaswami G, Van Nostrand EL, Nguyen TB, Hsiao YE, Lee C, Pratt GA, et al. (2019). Widespread RNA editing dysregulation in brains from autistic individuals. Nat. Neurosci 22, 25–36. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R95] Tsang B, Arsenault J, Vernon RM, Lin H, Sonenberg N, Wang LY, Bah A, and Forman-Kay JD (2019). Phosphoregulated FMRP phase separation models activity-dependent translation through bidirectional control of mRNA granule formation. Proc. Natl. Acad. Sci. USA 116, 4218–4227. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R96] Van Nostrand EL, Pratt GA, Shishkin AA, Gelboin-Burkhart C, Fang MY, Sundararaman B, Blue SM, Nguyen TB, Surka C, Elkins K, et al. (2016). Robust transcriptome-wide discovery of RNA-binding protein binding sites with enhanced CLIP (eCLIP). Nat. Methods 13, 508–514. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R97] Van Nostrand EL, Pratt GA, Yee BA, Wheeler EC, Blue SM, Mueller J, Park SS, Garcia KE, Gelboin-Burkhart C, Nguyen TB, et al. (2020). Principles of RNA processing from analysis of enhanced CLIP maps for 150 RNA binding proteins. Genome Biol. 21, 90. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R98] Vasilyev N, Polonskaia A, Darnell JC, Darnell RB, Patel DJ, and Serganov A (2015). Crystal structure reveals specific recognition of a G-quadruplex RNA by a β-turn in the RGG motif of FMRP. Proc. Natl. Acad. Sci. USA 112, E5391–5400. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R99] Wheeler EC, Van Nostrand EL, and Yeo GW (2018). Advances and challenges in the detection of transcriptome-wide protein-RNA interactions. WIRES RNA 9, 1436. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R100] Wickham H (2010). Stringr: Modern, consistent string processing. R J. 2, 38–40. [Google Scholar]

[R101] Yang L, Duff MO, Graveley BR, Carmichael GG, and Chen LL (2011). Genomewide characterization of non-polyadenylated RNAs. Genome Biol. 12, R16. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R102] Yi H, Park J, Ha M, Lim J, Chang H, and Kim VN (2018). PABP cooperates with the CCR4-NOT complex to promote mRNA deadenylation and block precocious decay. Mol. Cell 70, 1081–1088. [DOI] [PubMed] [Google Scholar]

[R103] Youn J, Dunham WH, Hong SJ, Knight JDR, Bashkurov M, Chen GI, Bagci H, Rathod B, MacLeod G, Eng SW, et al. (2018). High-density proximity mapping reveals the subcellular organization of mRNA-associated granules and bodies. Mol Cell 69, 517–532. [DOI] [PubMed] [Google Scholar]

[R104] Zhao J, Qin B, Nikolay R, Spahn CMT, and Zhang G (2019). Translatomics: The global view of translation. Int. J. Mol. Sci 20, 212. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Integrative Omics Indicate FMRP Sequesters mRNA From Translation and Deadenylation In Human Neuronal Cells

Tatsuaki Kurosaki

Shuhei Mitsutomi

Alexander Hewko

Nobuyoshi Akimitsu

Lynne E Maquat

SUMMARY

Graphical Abstract

eTOC Blurb

INTRODUCTION

RESULTS

FMRP-bound transcripts function in neurodevelopmental and signaling pathways in human neuronal cells

Figure 1. FMRP footprints in SH-SY5Y cells are enriched on mRNAs with neuronal functions whose misregulation is associated with FXS and autism.

FMRP binding occurs throughout mRNA bodies, primarily at structured and GC-rich sequences without sensing protein-coding potential, and correlates with increased mRNA stabilization

Figure 2. SH-SY5Y-cell anti-FMRP RIP-seq footprints correlate with structured and GC-rich sequences in mRNA 5′UTRs, CDSs and 3′UTRs.

Globally defining changes in protein abundance upon loss of FMRP

Figure 3. Amino acid metabolic labeling and quantitative mass spectrometry identify differentially expressed proteins in FMR1-KO relative to WT SH-SY5Y cells.

Globally defining the amount of protein made per mRNA upon loss of FMRP

Figure 4. FMR1-KO generally reduces mRNA abundance and increases the amount of protein made per mRNA.

FMRP binding to an mRNA 5'UTR, CDS, or 3'UTR inhibits mRNA degradation and translation

Figure 5. Functional consequences of FMRP binding to an mRNA 5'UTR, CDS or 3'UTR.

FMRP binding to poly(A)-bound PABPC1 protects against mRNA deadenylation

Figure 6. Evidence that FMRP binds mRNA poly(A) tails via PABPC1 to inhibit deadenylation.

DISCUSSION

Figure 7. Model for FMRP-mediated effects on mRNA translation and decay.

Limitations of the study

STAR★METHODS

RESOURCE AVAILABILITY

Lead contact

Material availability

Data and code availability

EXPERIMENTAL MODEL AND SUBJECT DETAILS

METHOD DETAILS

Plasmid constructions

Cell lysis, and protein and RNA preparations

siRNAs

Western blotting

Immunoprecipitations (IPs)

RT-qPCR

cDNA library constructions for RIP-seq and RNA-seq

Computational analyses of RIP-seq and RNA-seq data, GO term enrichment, biological pathway, and biological network analysis

Meta-analyses of our RIP-seq and published CLIP-seq data

Peak-calling analyses

Custom annotation and format conversions

Calculation of FMRP binding to full-length mRNAs, 5′UTRs, CDSs, or 3′UTRs

Calculation of G4-structure density

Calculation of mRNA structuredness

Calculation of codon optimality

Calculations of tri- or tetra-nucleotide densities in mRNAs

mRNA half-life changes upon FMRP-KD in SH-SY5Y cells

Correlation analyses

Meta-analysis of poly(A) RIP-seq reads

Sample preparation for SILAC protein quantitation

LC-MS/MS for SILAC protein quantitation

Data analysis for SILAC protein quantitation

Fluorescent EGFP and mRFP quantitations

Sample preparation for LC-MS/MS using FMRP-IP samples

LC-MS/MS using FMRP-IP samples

Data analyses for LC-MS/MS using FMRP-IP samples

Immunofluorescence

Purification of human PABPC1, human FMRP and EGFP from E. coli

RT-PCR, poly(A) primer-extension analysis, and splint RT-qPCR

QUANTIFICATION AND STATISTICAL ANALYSIS

Supplementary Material

Highlights.

ACKNOWLEDGEMENTS

Footnotes

REFERENCES

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases