Roles of RNA-binding proteins in neurological disorders, COVID-19, and cancer

Abstract

RNA-binding proteins (RBPs) have emerged as important players in multiple biological processes including transcription regulation, splicing, R-loop homeostasis, DNA rearrangement, miRNA function, biogenesis, and ribosome biogenesis. A large number of RBPs had already been identified by different approaches in various organisms and exhibited regulatory functions on RNAs’ fate. RBPs can either directly or indirectly interact with their target RNAs or mRNAs to assume a key biological function whose outcome may trigger disease or normal biological events. They also exert distinct functions related to their canonical and non-canonical forms. This review summarizes the current understanding of a wide range of RBPs’ functions and highlights their emerging roles in the regulation of diverse pathways, different physiological processes, and their molecular links with diseases. Various types of diseases, encompassing colorectal carcinoma, non-small cell lung carcinoma, amyotrophic lateral sclerosis, and Severe acute respiratory syndrome coronavirus 2, aberrantly express RBPs. We also highlight some recent advances in the field that could prompt the development of RBPs-based therapeutic interventions.

Introduction

RNA-binding proteins (RBPs) represent an important class of post-transcriptional regulators implicated in a plethora of biological processes in several organisms. It has become apparent that the binding function of RBPs is not limited to host survival but is implicated in abnormal cellular processes (e.g. cancer) and essential for some pathogens during infection of the host system. As almost every process in a cell is mediated by proteins and RBPs are posttranscriptional regulators of gene expression involved in stability, translation, or other biological events determining mRNAs fate, an approach focused on the deep characterization of RBPs for therapeutic drug discovery is required. RBPs have been prioritized in biomedical research due to their implication in gene expression and RNA processing, which are among the factors essential for gene function. Human RBPs are further classified in canonical RBPs containing RNA recognition motif (RRM), ribosomal and K-homology (KH) domains (600 structurally different RBPs, e.g. APOBEC3B, C9orf114, DDX58), and some non-canonical RBPs (Fig. 1) recognized by their intrinsically disordered regions (e.g. breast cancer type 1 susceptibility protein BRCA1, CDC40, SETD3) as not all IDR-containing RBPs are "non-canonical" [1]. These domains raised fundamental questions about the distinct role they might play in normal or altered biological events.

Fig. 1.

Typical domains observed in RNA-binding proteins. A Schematic representation of IMP3 domains. All tandem RBDs are endorsed with RNA-binding activities. RRM1 and RRM2 recognize CA-rich and CA repeats sequences in RNA. KH1-2 specifically recognizes GGC-CA or CA-GGC motifs in RNAs. KH3-4 identify GGC-CA motifs in RNAs [2]. B Schematic representation of IGF2BP domains. All tandem RBDs are identified but only KH domains are specifically used to recognizes m6A-modified mRNAs [3]. C Schematic representation of hnRNP A2/B1 domains with known residue numbers. RNA-binding domain (RBD) composed of tandem RRMs separated by a 15-aa linker, and a C-terminal Gly-rich low complexity (LC) region that includes a prion-like domain (PrLD), an RGG box: arginine-glycine-glycine box, and a PY-motif containing a M9 nuclear localization signal (PY-NLS). RRM1 specifically recognizes AGG motif in RNAs. RRM2 specifically recognizes UAG motif in RNAs but also purine-rich GAG, and pyrimidine-rich UU sequences [4]. D Schematic representation of TDP-43 RRM domains. 414-residue TDP-43 includes: N-terminal domain (NTD) over residues 1–80; two tandemly-tethered RNA recognition motifs (RRM1 and RRM2) over residues 105–261 connected by an unstructured linker over residues 181–192, and C-terminal prion-like domain over residues 274–414 [5]

RBPs can be classified based on their binding preferences [6], as single-stranded RNA (ssRNA) and double-stranded RNA (dsRNA). Amongst RBPs, the double-stranded RBPs TAO kinases (TAOK1, TAOK2, and TAOK3) are particularly critical for antiviral protein expression and/or antiviral signaling cascade regulation in human cells, with TAOK2 also being antiviral against Semliki Forest virus and Herpes-Simplex virus 1 [7]. The single-stranded RNA-binding protein coronaviral nonstructural protein 9 (Nsp9) contains a conserved self-associating GxxxG sequence motif within its C-terminal α-helix, the residues Phe-40, Val-41, as well as Ile-91 prone to perturbations, and N-terminus for high affinity for oligonucleotide [8]. For the ssRNA-binding protein, Nsp9, known to form an obligate homodimer via an essential GxxxG protein-interaction motif, in an unmodified state, it has been established that its ssRNA-binding mechanism used a pyrimidine RNA base-mimic compound [8].

The underlying mechanism implicated in the recruitment and binding of RBPs to the target mRNAs in several diseases is still under investigation. It is crucial to explore the regulation of RBP's interaction with the binding site of target mRNAs so that novel therapeutic strategies for the treatment of diverse RBPs-related diseases could be proposed.

Ribonucleic acid (RNA) therapeutics, exploiting messenger RNAs (mRNAs), small interfering RNAs (siRNAs), and micro RNAs (miRNAs), have garnered considerable interest as an approach to rapidly overcoming several diseases in modern medicine. Besides, recent reports have highlighted the promising role of mRNA vaccines in treating the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic [9, 10]. Further, the messenger RNA-based vaccine candidate encoding the SARS-CoV-2 spike glycoprotein receptor-binding domain (BNT162b1) has been revealed to display safety and immunogenicity results on vaccinated younger (aged 18–55 years) and older adults (aged 65–85 years) Chinese participants [11]. Vaccination with mRNA holds the potential to transform modern medicine by allowing it to overcome the existing bottleneck of conventional vaccines in the prevention of infectious diseases [12, 13]. Yet the reduced efficacy linked to poor RNA stability in solution is among the technical obstacles facing mRNA therapeutics [14, 15]. Recently some cases of infection with SARS-CoV-2 in vaccinated individuals have been reported [16], and post-SARS-CoV-2 mRNA vaccination responses exhibited a highly variable humoral and cell-mediated immunity across the range of immunodeficiency disorders [17].

Additionally, mRNAs have been identified to be presented as complexes with a diverse range of RNA-binding proteins (RBPs) where they contributed to the construction of messenger ribonucleoprotein (mRNP) complexes [18–20]. Therefore, efforts towards the production of therapeutic molecules in addition to vaccines (e.g. therapeutic antibodies) that would neutralize the target pathogen and impede its interaction with cellular receptors to gain entry into cells could be the object of much attention. Besides, it has already been envisioned for the treatment of COVID-19 [21]. RBPs represent an important class of post-transcriptional regulators implicated in a plethora of biological processes in several organisms. It has become apparent that the binding function of RBPs is not limited to host survival but is implicated in abnormal cellular processes (e.g. cancer) and essential for a diverse array of viruses during infection of the host system. The same RBP activity can be shared by at least two diseases. For example, beyond the role notified for the RBPs of Musashi (MSI) protein family in diverse cancer cells, spanning breast, colorectal, lung, and pancreatic cancers, as well as leukemia or glioblastoma [22], an additional role as host factors stimulating the replication of a mosquito-borne Flavivirus Zika virus (ZIKV) has been recognized for them [23]. The recent identification of 9 shared RBPs that connected COVID-19 to neurological disorders and diverse types of cancers at the molecular level prompted us to look into the role of RBPs in neurodegenerative diseases, cancer progression, and viral infection. For example, Ref. [24] unveiled that the RBPs RPS10, LIN28B, MOV10 (Moloney murine leukemia virus infection in mice), RPS3, EIF4B, EIF5A, EEF1A1, RPL18A, and Poly(A) Binding Protein Cytoplasmic 1 (PABPC1) are jointly involved in COVID-19, cancer, and neurological disorders [24]. Furthermore, it has been revealed that the RBP TDP-43 previously known for its importance in neurodegenerative disease progression, also played a key role in hepatocellular carcinoma progression. Indeed, TDP-43 binds to the UG-rich sequence1 of the 3’UTR of ABHD2 (abhydrolase domain containing 2) mRNA to improve the stability of ABHD2 mRNA stimulates the synthesis of free fatty acid and fatty acid oxidation-derived reactive oxygen species in an ABHD2-dependent manner, to oppose apoptosis and improve the proliferation of hepatocellular carcinoma cells [25]. Taking together, the aforementioned roles of RBPs turned them into essential targets to reveal the mechanism of a plethora of diseases.

This review is particularly important as it highlights the strong effect of RBPs in several pathways and mechanisms that can protect from or lead to disease, pinpoints the sequences from RNAs that can be monitored to find long-lasting treatments, offers strategies to control RBP activity in time and space, and reports methods to enrich subcompartment-specific RBPs.

General roles attributed to RNA-binding proteins

RBPs are of interest since their activities are involved in the direct monitoring of a diverse array of events spanning transcription, RNA splicing, polyadenylation, stabilization, localization, and translation [26], and their aberrant functions are implicated in a variety of diseases such as viral infection, cancer, and neurodegenerative disorders. They are among the diverse human innate immune response mechanisms that may trigger signaling cascades in the host to halt or allow pathogen establishment in the cell. A growing body of evidence supports the likelihood that RBPs exert a major influence on cellular events to drive cell function, and that a mechanistic understanding of RBP's function during disease progression is crucial for the development of new disease therapeutics. RBPs affected alternative splicing regulatory events (e.g. TDP-43-mediated mutually exclusive exons, skipped exons, alternative 5' and 3' splice sites, and intron retention [27], translational efficiency events, pre-mRNA processing, mRNA nuclear export and retinal development, sterility mechanisms, stress granules management, R loop formation and are identified in processes occurring in neurodegenerative disease, cancer, and viral infection. For example, in addition to their central role in the monitoring of gene expression through changes in RNA stability and degradation, RBPs are also implicated in DNA damage response regulation and one of them, HNRNPC (heterogeneous nuclear ribonucleoprotein C), is identified as a regulator of chemotherapeutic resistance in human lung cancer [28]. Furthermore, this latter finding has clear clinical relevance for the response of lung cancer to chemotherapy in human patients. The different roles attributed to RBPs are depicted in Fig. 2. Furthermore, in recent times there are many emerging microbial infections with obscure immune invasion properties and no effective therapeutic targets available to limit their fast expansion or outbreaks. Therefore, the design of strategies focusing on the proteins targeted by these pathogens in the host could bring more therapeutic solutions for their control.

Fig. 2.

Numerous cellular events and disease processes regulated by RNA-binding proteins. The RBP PUM1 binds the 3´-untranslated region of Toll-like receptor 4 (TLR4) to clear TLR4 mRNA translation and monitor the activity of nuclear factor-κB (NF-κB), a master regulator of the aging process in human mesenchymal stem cells. PUM1 also protects human chondrocytes to avoid chondrogenic phenotype loss and alleviates cellular osteoarthritis [29]. The loss of functional RBP muscle blind-like protein 1 (MBNL1) in response to its nuclear seclusion by mutant transcripts containing pathogenic expanded CUG repeats (CUGexp) leads to Myotonic dystrophy type 1 [30]. Mitochondrial RNA-binding protein tumor necrosis factor receptor-associated protein 1 (TRAP1)-associated, chaperone-mediated autophagy can stimulate ferroptosis [31]. The RBP human antigen R (HuR) translocation into the cytoplasm orchestrates by its interaction with circular RNAs circStag1 is involved in the regeneration of bone tissue [32]. Inclusion of the RBP TDP-43-regulated alternative exons is altered in skeletal muscles of patients with inclusion body myositis (IBM, [27]). RBP ZFP36 and ZFP36L1 bind transcripts encoding subunits of the NF-κB pathway, Notch1, Irf8 and Il2 to monitor the abundance of their protein products early after activation in CD8 + T cells [33]. Heterozygous frameshift variants in the RBP hnRNPA2/B1 do not augment the propensity of hnRNPA2 protein to fibrillize but diminished affinity for the nuclear import receptor karyopherin β2, leading to cytoplasmic accumulation of hnRNPA2 protein in cells and early-onset form of oculopharyngeal muscular dystrophy [34]. HNRNPC (heterogeneous nuclear ribonucleoprotein C) associated with the newly identified regulator of ferroptosis in Colorectal cancer, CUL9, for resistance to drug-induced ferroptosis [35]. The RBP heterogeneous nuclear ribonucleoprotein RBP (hnRNPH1) is essential for pre-mRNA alternative splicing, spermatogenesis and oogenesis [36]. RBP Tardbp is an important candidate monitoring alternative splicing and alternative polyadenylation to optimize clonal expansion and human CD8 + T cells effector function during an antigen-specific immune response [37]. In embryonic stem cells, RBPs represent half of the chromatin proteome, co-compartmentalize with RNA polymerase (Pol) II at promoters and enhancer, and connect RNA to the transcription machinery [38]. The RBP DDX41 is a tumor suppressor that opposes double-strand DNA breaks, genomic instability and R-loop-dependent replication stress by preferentially binding RNA–DNA hybrids and unwinding RNA–DNA hybrids in R-loops, and also reduces fragility of DNA in promoter regions [39]. RBP HNRNPC is demonstrated to interact with circXRCC5 to encourage circXRCC5 biogenesis that was proved to foster Gastric cancer progression [40]. By binding to hTERT (reverse transcriptase protein catalytic subunit hTERT) mRNAs in intron 8, the RBP NOVA1 (neuro-oncological ventral antigen 1) acts as a splicing enhancer that promotes the inclusion of exons 7 and 8 to enhance the production of full-length (FL) hTERT mRNAs in non-small cell lung cancer cells [41]. The RBP “partner of NOB1” (PNO1)-mediated ribosome biogenesis is negatively regulated by miR-340-5p directly binds the 3′-UTR of PNO1 and assumes an important role in LUAD progression via Notch signaling pathway [42]. RBP (ROD1, so-called PTBP3) interacts with Activation-induced cytidine deaminase (AID), which jointly bind bi-directionally transcribed RNAs to facilitate genome-wide AID targeting to immunoglobulin (Ig) loci in a way to induce DNA rearrangement during immune responses [43]. The RBP KHSRP regulates cell proliferation and cell cycle in Wilms tumor by modulating the expression of PPP2CA and p27 [44]

Human RBPs are reported to frequently contain specific low-complexity motifs and bind various coding along with non-coding RNAs (ncRNAs) types, including messenger RNA (mRNA), transfer RNA (tRNA), small nuclear RNAs (snRNAs), small nucleolar RNAs (snoRNAs), ribosomal RNA (rRNA), internal ribosome RNA (irRNA), the novel class of RNAs circular RNAs (circRNAs), RNA polymerase II [45]. CircRNAs are noticeable for their role in cancer progression. For example, RBP Polypyrimidine tract binding protein 1 (PTBP1) is reported to repress the expression of circ_0076611, whose interaction with diverse proliferation-related transcripts (MYC and VEGFA mRNAs) augments cell proliferation and migration in triple-negative breast cancer cells, the RBP oncogenic splicing factor SRSF1 (Serine And Arginine Rich Splicing Factor 1) encourages this expression after the inactivation of PTBP1 protein by the long non-coding RNAs (lncRNA) MALAT1 (metastasis-associated lung adenocarcinoma transcript 1 [46]. Several novel RBPs regulating circRNA functions in tumorigenesis, such as SRSF5, HNRNPA2B1, HNRNPF, KHSRP, or PSIP1, have also been unveiled [47].

The discovery of different types of RNAs targeted by RBPs makes the overall investigation of RBP's target sites more challenging. It is established that RBPs regions that bind mRNA, rRNA, and snRNA are enriched in disordered proteins, while regions interacting with ncRNA and irRNA, and tRNA are depleted in disorder proteins [48].

Binding preference and activity of RNA-binding proteins

The function of RRMs as well as the residues they contain have been investigated. For example, the truncation in the C-terminal region identified as the RNA-binding domain of the RBP (Hel2), not only altered the ability of Hel2 to bind to 18S rRNA and translated mRNAs but also led to a loss of polysomal association and monitoring of both Ribosome-associated quality control (RQC) pathways and no-go decay [49].

Moreover, emerging pieces of evidence have shown that RBPs deployed their multidomain to recognize specific target sequences in RNAs. For example, the multidomain RBP, insulin-like growth factor 2 mRNA-binding protein 3 (IMP3, also known as IGF2BP3), used six RNA-binding domains (four KH and two RRM domains) to identify CA-rich and GGC-core elements in its targets RNA [2]. Furtherly, the RBP IGF2BP3 can directly augment the mRNA stability of E2F3 by binding the Long noncoding RNAs, LINC00958, and independently bind to the 3’UTR region of E2F3 to stimulate endometrial carcinoma progression [50]. Additionally, IGF2BP3 has the potential to post-transcriptionally control the oncogenic activity of MLL-AF4 protein; an oncogenic driver of MLL-rearranged leukemia [51].

RBPs preferentially bind sequences in 5′ or 3′ untranslated regions (UTRs) or RNA folds in the coding regions of mRNAs and orchestrate stability and translation efficiency [52]. It has been recently displayed that somatic 3′-UTR mutations (3′-UTR single nucleotide variants) in a given sequence motif (e.g. AACAAA) residing in mRNA could disrupt the binding between RBPs (e.g. SRSF1 and PTBP1) and their corresponding RNAs. The abnormal expression of RBPs resulting from this dysregulation is reported broadly in seven cancer types including biliary adenocarcinoma, breast adenocarcinoma, kidney renal cell carcinoma, liver hepatocellular carcinoma, prostate adenocarcinoma, stomach adenocarcinoma, as well as our in-house esophageal squamous cell carcinoma [53]. Mechanistically, the RBP La-related protein 6 (LARP6) is demonstrated to bind 5′ stem-loop (5′SL) of collagen mRNAs with its La domain-containing RNK-motif to stabilize the complex between La domain and 5′SL RNA using its juxtaposed RNA recognition motif and to stimulate the excessive synthesis of type I collagen leading to fibrotic diseases [54].

The binding properties and function of some RBPs on their targets can be sequence-specific. For example, the RBP DHX36 is an indispensable post-transcriptional regulator of stem cell proliferation and muscle regeneration acting via binding and unwinding G-quadruplex (rG4) structures at 5′ UTR of target mRNAs [55]. The identity of some RNAs' motifs has been determined for a subset of RBPs. A report elucidated that the RRM1 and RRM2 domains of the human RBP hnRNP A2/B1-mediating maturation or transport of mRNA as well as gene regulation of long noncoding RNA, recognized the AGG and UAG motifs, respectively, in their target RNAs [4].

Georgakopoulos-Soares and colleagues [56] demonstrated that the RBPs HNRNPK and HNRNPU displayed higher binding to G-quadruplex (G4) motif around splice sites, are positively connected to differentially included exons and have a direct impact on G4-mediated alternative splicing (AS) regulation. On the other side, RBPs such as AQR and RBM15 are depleted around splice sites flanked by G4 motifs and are negatively connected to differentially included exons, and the binding and impact of these regulatory factors over AS are suggested to be repressed by G4 formation [56]. The same authors further revealed that exon–intron junctions in humans exhibited an acute enrichment for G4s motifs, both at the 3′ss and the 5′ss, 31% of human genes displayed a G4 motif near at least one splice site within a distance of 100 bp, alternative splicing modulation by G4s is constricted to mammals and birds, but absent in plants, other tetrapods or fish. This data highlights the key regulatory roles exerted by RBPs on alternative splicing events known as a pivotal step of eukaryotic mRNA processing during gene expression.

Besides binding to nucleic acids [57] identified that the RRM of RBPs can bind to proteins. For example, the RRM2 of the pre-mRNA processing hnRNP proteins member, hnRNP L, forms a ternary complex with RNA as well as with the methyltransferase protein SETD2. For the interaction with SEDT2 to take place, leucine pair within a highly conserved stretch of SETD2 insert their side chains in hydrophobic pockets formed by hnRNP L RRM2. The same authors highlighted how specific interactions between RBPs and other proteins allowed the coupling of transcription and splicing. Indeed, they uncovered that the histone methyltransferase SETD2 collaborates with Pol II and interacts with RRM2 domains of the RBP hnRNP L via its SETD2 hnRNP interaction (SHI) domain to allow the deposition of the histone mark H3K36me3-regulated splicing, thereby mediating the coupling of splicing and transcription in mammalian cells [57].

Additionally, mutations in RNA-binding motifs drive disease progression such as cancer progression. The RBP U2AF1 carries hotspot mutations in its RNA-binding motifs that create new 3′ splice site contacts (at − 3 nucleotides) altering RNA splicing, illuminating how mutations in RBPs’ motifs affect splicing and promote cancer [58].

Some RBPs required a certain modification (e.g. post-transcriptional mRNA modification N6-methyladenosine m6A) of their target mRNAs to act properly. N6-methyladenosine (m6A) is commonly detected methyl-marks on RNA (e.g. mRNA, rRNA) and its deposition is catalyzed by RNA methylation writer enzymes, so-called RNA methyltransferases (e.g. complex METTL3-METTL14 complex acting on m6A deposition; [59]. A study revealed that the RBPs IGF2BPs preferentially recognize m6A-modified mRNAs, promote the stability of thousands of potential mRNA targets in an m6A-dependent manner under normal as well as stress conditions, and their oncogenic roles in cervical, as well as liver cancers, depending on their role as m6A readers [60]. It has been also evidenced that the RBP, Fragile X mental-retardation protein (FMR1), preferentially binds m6A-modified “AGACU”-containing mRNAs. The presence of m6A-modified RNA augments the size of FMR1 granules, improves the partitioning of non-methylated RNA into the granules, and facilitates the role of FMR1 in maternal mRNA decay [61]. On the other hand, in human cell lines, the DEAH-box helicase DHX36/RHAU is reported to preferentially bind the G-rich and G4-forming sequences in mRNAs and unwinds the typically formed rG4s to increase translational efficiency [62].

All the aforementioned events prompt the detailed analysis of RBPs' actions in mammals, especially from a disease development viewpoint.

RBPs assume diverse types of functions in mammalian cells

A thorough report on the role of RBPs in humans has been associated with neurodegenerative disease, viral infection, and cancer.

Contribution of dysregulated RBPs to neurodegenerative diseases

RBPs are involved in the metabolism of neurodegenerative diseases demonstrated recently to share a feature with the neurological effects of COVID-19; the occurrence of cytotoxic aggregated amyloid protein or peptides [63]. RBPs notified in cellular mechanisms altered in neurodegenerative diseases (mislocalization of RBPs), played a role in multiple RNA processing stages spanning transcription (FUS), turnover (MATR3), alternative splicing (TDP-43, FUS, hnRNPA1, hnRNPA2/B1), and stress granules; TDP-43 [64–66].

In addition to the commonly possessed RNA interacting motifs, RBPs often harbor IDRs that have been noticed to be involved in forming membrane-less organelles (such as stress granules SGs), which assemble via liquid–liquid phase separation (LLPS). SGs have a physiologically relevant role as many disease-linked RBPs (FUS, hnRNPA1, and hnRNPA2) identified inside contained intrinsically disordered low-complexity domains that can be prion-like domains (PrLDs) whose mutations accelerated fibrillization and caused disease [67]. For example, the RBP TDP-43 whose activity is dysregulated in neurodegenerative disease contained the two highly conserved RNA recognition motifs (RRM1 and RRM2), a short polyalanine tract and an intrinsically disordered region (IDR) harboring an “AMMAAAQAALQ” amino acid motif prone to form an α-helix promoting condensation. Its IDR differs from hnRNP A1 or A2’s as it is prion-like and possesses a short α-helical region promoting self-association and LLPs. The LLPS is driven by just a few aromatic residues, a feature that has been attributed to the presence of the α-helix stimulating inter-molecular contacts [68]. In healthy cells, TDP-43 proteins normally egress nuclei by passive diffusion, independent of facilitated mRNA export, and their nuclear localization relies on binding to GU-rich nuclear RNAs [69]. Table 1 highlighted some RBPs connected to neurodegenerative disease and the effect of their inhibitors. On the other hand, Fig. 3 shows illustrative examples of representative diseases implicating intrinsically disordered regions.

Table 1.

Contribution of dysregulated RNA-binding proteins to neurodegenerative diseases

Dysregulated RBPs	Biological events	Disease case		References
hnRNPA2/B1	Missense variants in the LCD	MSP or Paget’s disease of bone		[70]
	Heterozygous frameshift variants in the LCD partially impairing nuclear import and augmenting cytoplasmic concentration of protein without changing its intrinsic propensity to fibrillate	Early-onset oculopharyngeal muscular dystrophy
TDP-43	Mislocalization to the cytoplasm due to the impaired structural integrity of the nucleus connected to the translation of repeat expansion in C9orf72-linked neurodegeneration disease into dipeptide repeat proteins	ALS or FTD		[71]
				[72]
	Cytoplasmic mislocalization and LLPS due to RNA depletion	ALS		[73]
FUS	Oligomerize into condensates			[74]
	Synaptic localization of FUS, in absence of aggregation, triggered synaptic impairment	ALS		[75]
MATR3	Prone to phase separation into droplet-like structures			[76]
Inhibitors of dysregulated RBPs	Anti-TDP-43 aggregation activity of TDP-43 bearing a missense mutation A315T		Acridine derivative (AIM4)	[73]
	Impairment of the TDP-43 prion-like domain proneness to aggregation		ATP with concentrations > mM	[5]
	Suppression of aberrant phase separation and solidification of TDP-43	ALS and FTD	C-terminal hyperphosphorylation of TDP-43	[77]
	Inhibition of toxic misfolding effects of both TDP-43 and FUS		Single missense mutation in nucleotide-binding domain 1 (NBD1) or NBD2 of the AAA + protein disaggregate, Hsp104	[78]
	Targeted cNLS (a canonical NLS) to repress and reverse		Nuclear-import receptor Karyopherin-β1 cooperating with Importin-α	[67]
	TDP-43 fibrillization
	Targeted PY-NLSs (PY-nuclear localization signal (NLS)) to repress and reverse TAF15, EWSR1, hnRNPA1, and hnRNPA2 fibrillization		Karyopherin-β2	[67]
	Binding to the C-terminal proline-tyrosine nuclear localization signal (PY-NLS) of FUS and impairment of FUS self-association, blocking LLPS		Karyopherin-β2/transportin-1	[79]
Other inhibition method	Depletion of cytosolic RBP (ATXN2)	ALS and FTLD	Protection against TDP-43 neurotoxicity	[80]

Fig. 3.

Architecture of typical RNA-binding proteins with low complexity domain localized in intrinsically disordered proteins. The figure highlighted RNA recognition motifs 1 and 2 (RRM1 and RRM2), a low complexity domain (LCD), and a nuclear localization signal (NLS) within the LCD. Missense mutations (G287S, A315T, G348C, R361S, A382T, N390D, N390S) in LCD are responsible for amyotrophic lateral sclerosis [81]. Pathogenic mutations (c.785/941A.T, p.D262V/D314V in hnRNPA1 and c.869/905A.T, p.D290V/D302V in hnRNPA2B1) in prion-like domains of hnRNPA2B1 and hnRNPA1 are responsible for multisystem proteinopathy (MSP; [34]. Missense mutations (P362L, A381T, and E384K) impacting the LCD of the RNA-binding protein T cell-restricted intracellular antigen-1 (TIA1) drive ALS and Frontotemporal Dementia (FTD) [82]. RBMX’s low-complexity region consists of a serine and arginine-rich region (SRR) followed by a tyrosine-rich region (TRR) is required for monitoring the expression of CBX5 mRNA encoding CBX5 protein in the hematological malignancy acute myeloid leukemia (AML) cells [83]

Dysregulated role of RBPs in viral replication and infection

Certain viruses are a serious public health threat responsible for frequent morbidity and mortality in humankind. The antiviral activity of RBPs is stimulated by both RNA viruses and RNAs transcribed from DNA viruses [84]. Elucidating the molecular interplay between viral RNAs and host cell proteins during infection can enhance our understanding of viral RNA functions and the host's innate immune defenses. A virus can remodel host metabolism for successful viral infection and transient stimulation of cancer in cancer virus conditions. A set of RBPs involved in viral metabolism has been gathered in Table 2. Based on these data, deploying programmable RBPs is conceivable to strengthen human immune responses.

Table 2.

Contribution of dysregulated RNA-binding proteins to viral infection

Dysregulated RBPs/Inhibitors	Biological events	Virus type	References
Interferon-induced protein with tetratricopeptide repeats 2 (IFIT2)	Targeted by virus to sustain viral gene expression and viral replication	Influenza virus	[85]
Heterogeneous nuclear ribonucleoprotein AB (hnRNPAB)	Impaired influenza A virus replication via the inhibition of the nuclear export of viral mRNA	Influenza A virus	[86]
EIF4G1, EIF4G3, EIF4A1, EIF4A2, EIF4B, PABPC1, HNRNPA1, LSM14A, PATL1, DDX6, DDX60, DHX57, HSP90AA1	Interactors of SARS-CoV-2 RNA	Severe acute respiratory syndrome coronavirus 2	[87]
PPIA or cyclophilin A, PA2G4, ZC3H11A, DDX3, and HSP90AB1	Highly upregulated by SARS-CoV-2 as well as Sindbis infections	SARS-CoV-2 and Sindbis (SINV)	[87]
IGF2BP1	Binds the SARS-CoV-2 RNA genome, stabilizes RNA, encourages the translation of SARS-CoV-2 RNA, and regulates SARS-CoV-2 infection	SARS-CoV-2	[88]
	Promotes the translation of Zika virus RNA
Heat shock protein HSP90AB1	Binds SARS-CoV-2 RNA and monitors infection of SARS-CoV-2	SARS-CoV-2	[88]
KSHV-encoded viral interferon regulatory factor 1 (vIRF1)	Triggered by KS-associated herpesvirus (KSHV) infection, upregulated and recruited E3 ubiquitin ligase Kelch-like 3 (KLHL3) to decompose heterogeneous nuclear ribonuclear protein Q1 (hnRNP Q1) via a ubiquitin–proteasome pathway, which destabilized GDPD1 mRNA	KS-associated herpesvirus	[89]
Poly (rC) binding protein 2 (PCBP2)	Targeted by microRNA-like small RNA (miR-HA-3p) encoded by the H5N1 influenza virus that suppressed its expression and promoted cytokine production in human macrophages infected with the H5N1 virus	H5N1 influenza virus	[90]
Ebola virus viral protein 35 (VP35)	Contained interferon inhibitory domain (IID) and a cryptic pocket allosterically coupled to a key dsRNA-binding interface involved in Ebola infection	Ebola Virus	[91]
GRSF1	Allowed efficient cytosolic accumulation and translation of viral mRNAs	Avian influenza A virus	[92]
DEAD-box helicase 6	Interactor of Zika virus–derived subgenomic flavivirus RNA for compromising host response to viral infection	Zika virus	[93]
The isoform of the canonical translation factor eukaryotic translation initiation factor 4γI (DAP5)	Essential for the initial round of translation of the enteroviral (Coxsackievirus B3) RNA	Coxsackievirus B3	[94]
EB2	Essential for stabilizing viral mRNAs generated from intronless genes in the nucleus and trigger their export to the cytoplasm along with their translation	Virus	[95]
TRIM25, TRIM56, ZC3HAV1/ZAP, DHX36, and GEMIN5	Involved in the anti-viral response against both SARS-CoV-2 and SINV	SARS-CoV-2 and SINV	[87]
HSP90 and IGF2BP1	Drugs targeting these RBPs significantly repress SARS-CoV-2 protein production and infection	SARS-CoV-2 and SINV	[87]
Nuclear Factor 90 (NF90)	Bound to the 5’UTR and regulated the abundance of the Interferon Regulatory Factors (IRFs) IRF3 and IRF9 mRNAs encoding IRF3 and IRF9 proteins that are found vital for the monitoring of the interferon pathway implicated in the immunological response to the SARS-CoV-2 infection	SARS-CoV-2	[96]
La-related protein 1 (LARP1)	Bound to the 5′-leader containing TOP-like sequence motif in SARS-CoV-2 mRNA and impeded replication in infected human cells under dynamic phosphorylation by the mTORC1 complex	SARS-CoV-2	[97]
Interferon-induced zinc-finger antiviral protein (ZAP-S)	Interacted with the SARS-CoV-2 RNA through the SL2 and SL3 regions of the SARS-CoV-2 pseudoknot and interfered with the folding of the frameshift RNA element to repress programmed ribosomal frameshifting and impair viral replication	SARS-CoV-2	[98]
Peptidyl-prolyl cis–trans isomerase A, sodium/potassium -transporting ATPase subunit alpha-1 and actin-related protein 2 complex (ARP2/3 complex)	Decreased viral replication in human cell lines	Virus	[97]
ZFP36 and ZFP36L1	Their defects led to superior cell-intrinsic cytotoxicity and conferred the greatest protection against influenza A virus infection	Influenza A virus	[33]
Human Silencing Hub complex (TASOR)	Facilitating the association of RNA degradation proteins with RNA polymerase II is implicated in the human immunodeficiency virus proviral expression repression	Human immunodeficiency virus	[99]

Furthermore, a recent role demonstrated for RBPs in defending the host system against viral infection and promoting the expression of genes (IFIT1, IFIT2, IFTI3, IFI44, HERC4, OASL) endowed with antitumoral effects (e.g. multifunctional RBP Raly; [100]), led us to highlight the diverse impact of RBPs in tumorigenesis.

Crucial role played by dysregulated RNA-binding proteins in tumorigenesis and cancer progression

Cancers are often characterized by frequent genetic abnormalities, inhibition of tumor suppressor genes, molecular alterations, high resistance to existing therapies, and recently by the overwhelming presence of 3′-UTR posttranscriptional impairment-related single nucleotide variants (3′-UTR piSNVs) that disturbed RBPs binding, APA and m6A RNA modification [53]. Some cancers are associated with the worst prognosis with decreased overall survival rates. For example, it has been demonstrated that left-sided breast cancer is linked to aggressive biology and worse sequels as opposed to right-sided breast cancer [101]. Further investigations revealed that glioblastoma is a highly aggressive brain tumor displaying poor outcomes with a median survival time of approximately 1-year, a 5-year survival rate of only 5.8%, and a major therapeutic challenge [102]. An ideal medical treatment for a given cancer should accurately address its complexity and its resistance to standard therapies. In line with this statement, cancer interacting partners (e.g. RBPs) may represent a therapeutic opportunity for identifying sustainable solutions for patients with advanced or aggressive diseases. RBPs have emerged as important contributors to tumorigenesis and tumor progression. For example, the RBPs PUMILIO, comprising the two human homologs PUM1 and PUM2, have been demonstrated to suppress the function of the cell cycle regulators p21 and to exert a critical role in human colorectal cancer progression. Their inactivation resulted in colorectal tumor growth reduction and a delay in G1/S transition [103]. These data illustrated the oncogenic roles of some RBPs and opened new therapeutic perspectives for targeting cancers in humans.

It has been provided evidence that the RBP ADAR3 (Adenosine deaminases that act on RNA) increased Nuclear factor-κB (NF-κB) activation and a gene expression program favorable for glioblastoma tumors progression [104]. The molecular mechanism underlying the role of other tumor-promoting RBPs has also been established. Indeed, it has been unveiled that the RBP zinc finger CCHC domain-containing protein 4 (ZCCHC4) exerted its antitumor effect by impeding DNA-damage-induced apoptosis in hepatocellular carcinoma (HCC) cells through the inhibition of the pro-apoptotic function of the new long noncoding RNA (AL133467.2) and also interrupts the interaction between AL133467.2 and γH2AX upon DNA-damage agent (DDA) treatment to repress apoptotic signaling and stimulate chemoresistance to DDAs. Its defects stimulate AL133467.2 and γH2AX interaction for improving chemosensitivity in HCC cells [105]. Another study pointed out that the RBP Paraspeckle component 1 (PSPC1), exerted a crucial role in the proliferation of estrogen receptor (ER)-positive breast cancer cells by interacting with the RBP belonging to Drosophila behavior human splicing family, PSF, and affecting the post-transcriptional regulation of PSF target genes, ESR1 and SCFD2 [106].

Additional roles have been attributed to RBPs in tumorigenesis where they exerted multiple functions spanning tumor initiation, chemoresistance, tumor suppression, R-loop clearance, and circular RNAs biogenesis. Yet, the RBPs monitoring the activity of the tumor suppressors (e.g. p53) is particularly important to analyze.

The RBP named RBM28, known as a nucleolar component of spliceosomal small nuclear ribonucleoproteins, promotes the survival and growth of cancer cells by both translocating from the nucleolus to the nucleoplasm upon phosphorylation and interacting with the DNA-binding domain of tumor suppressor p53 to inhibit p53 transcriptional activity [107]. Moreover, the direct binding of the RBP, LIN28B, to the 5′-UTR of p53 mRNA is reported to impede the translation of p53 by preventing the binding of the translation enhancer protein, ribosomal protein L26 (RPL26), to 5′UTR of TP53 mRNA, thereby facilitating the resistance of cancer cells to p53–MDM2 protein–protein interaction inhibitors (Nutlin3a and MI773, [108]. These two RBPs can be therefore the perfect potential biomarkers and therapeutic targets for cancer inhibition. The tumor-suppressive RBP, hnRNP-E1, is known to modulate splicing and translation of the epithelial to mesenchymal transition (EMT)-associated transcripts and to play a major role in the control of epithelial cell plasticity during cancer progression. The ARIH1, an E3 ubiquitin ligase, can target and interfere with the activity of this RBP by degrading it and promoting breast cancer progression [109]. The RBP DDX41 is a tumor suppressor that opposes double-strand DNA breaks, genomic instability, and R-loop-dependent replication stress by preferentially binding RNA–DNA hybrids and unwinding RNA–DNA hybrids in R-loops, and also reduces the fragility of DNA in promoter regions [39]. It joins therefore the repertoire of RBPs capable of suppressing R-loop formation in human cells.

In the disease suppression pathways, RBPs can regulate microRNA’s (miRNA) function in cancer cells. Indeed, the RBP, Heterogeneous nuclear ribonucleoprotein A2/B1 (HNRNPA2B1), binds to the UUAGG motif of the stem structure, formed by the conserved cis-elements of miR-506 (UGCCUUA) and HNRNPA2B1 binding sites (UUAGG) in the CDK6 3′-UTR, to denature it and recruit the RNA helicase DExH-box helicase 9 (DHX9) to the 3’-UTR, which ultimately facilitates miRNAs-mediated suppression of CDK6 protein expression in lung cancer cells [110].

In addition to their involvement in pathways leading to disease suppression, RBPs are found in pathways resulting in disease activation. An intact RNA-binding motif is necessary for the EXOSC9 protein (RNA exosome constituent) to promote tumor growth and mediate stress resistance as well as P-bodies formation in cancer cells [111].

Although RBPs have been linked to different human cancers, the activities of some of them can be cancer-specific or patient-specific. For example, the higher expression of the RBP KHDRBS1 (KH domain-containing, RNA-binding, signal transduction-associated protein 1)/Sam68 (Src substrate associated in mitosis 68 kDa) leads to reduced survival of the patient in kidney renal papillary cell carcinoma and lung adenocarcinoma but not in acute myeloid leukemia and ovarian cancer [112]. This data displayed its critical role in cancer-specific events.

RBPs are implicated in the early stage of cancer manifestation. For example, in the work of [113], the increase in mitochondrial fatty acid beta-oxidation (FAO) and oxidative phosphorylation (OXPHOS) allows adenosine triphosphate (ATP) production and cellular lamellipodia formation during the initial step of High-grade serous ovarian cancer (HGSOC) cell migration and invasion. FAO and OXPHOS activities are linked to the improvement of mRNA stability and expression of carnitine palmitoyltransferase 1A (CPT1A) as well as NADH: Ubiquinone Oxidoreductase Subunit A2 (NDUFA2) resulting from the direct interaction between the Cancer-type organic anion transporting polypeptide 1B3 (Ct-OATP1B3) and the RBP insulin-like growth factor 2 mRNA-binding protein 2 (IGF2BP2). IGF2BP2 binding to N(6)-methyladenosine (m6A) methyltransferase of polo-like kinase 1 (PLK1) 3′UTR, initiated by the methyltransferase-like 3 (METTL3) activity, is also revealed to be essential for maintaining the cell cycle in pancreatic adenocarcinoma [114].

RBPs can be an important cause of chemoresistance in cancer. For example, in colorectal cancer (CRC), an investigation by [115] demonstrated that the RBP, leucine-rich pentatricopeptide repeat-containing protein (LRPPRC) specifically bound to the mRNA of multidrug resistance 1 (MDR1), augmenting MDR1 mRNA stability and protein expression. Fortunately, LRPPRC is more recently found to be impaired by gossypol-acetic acid [115]. RBPs activities have been noticed in autophagosome processes in human cancer. The RBP SQSTM1 is responsible for transporting the lncRNAs ARHGAP5-AS1 for autophagic degradation. The impaired autophagic degradation of ARHGAP5-AS1 is found to stimulate chemoresistance in human gastric cancer [116].

RBPs have been found in breast cancer (e.g. the RBPs RBM20 and PCDH20, [117], and their roles in cancer progression have also been proven. It has been shown that the RBP nucleolin (NCL) bound with Long non-coding RNA LINC01296, mainly located in the nuclei of Neuroblastoma (NB) cells, and formed a complex that triggered SRY-box transcription factor 11 (SOX11) gene transcription and accelerated the tumorigenesis of NB, a known common extracranial solid tumor in childhood [118]. Puf-A, a novel member of Puf-family RBPs, encourages non-small cell lung cancer (NSCLC) and colorectal cancer progression by interacting with nucleolar factors nucleophosmin-mediated ribosome biogenesis in nucleolus via the residues 478RRR480 of Puf-A [119]. This RBP may be useful for an effective mRNA-based treatment that will significantly suppress tumor growth.

RBPs are the targets of proteins promoting cancer invasion and metastasis. In the case of human Gastric cancer (GC), ecdysoneless (ECD) protein promotes cancer invasion and metastasis by binding the RBP hnRNP F via the N-terminal STG1 domain (13-383aa), preventing hnRNP F ubiquitination by the E3 ubiquitin ligase ZFP91, and proteasomal degradation. This effect of ECD turned it into an attractive marker candidate and a potential therapeutic target for GC patients [120].

An expanding set of RBPs is also involved in multiple malignancy events. The frequently overexpressed cancer-associated protein, death-associated protein 3 (DAP3), is an RBP demonstrated to act as a splicing regulatory RBP in cancer via its ability: (i) to mediate the formation of ribonucleoprotein complexes to trigger substrate-specific splicing changes, and (ii) to modulate the splicing of numerous splicing factors (RBM6, HNRNPH1, AKAP17A, and TIA1) to cause an indirect effect on splicing. This RBP triggers the nonsense-mediated decay of RBP (RBM6) and allows the association of splicing factor proline and glutamine-rich (SFPQ) and Non-POU domain-containing octamer binding (NONO) with target RNAs for splicing modulation. Its defects impair tumorigenesis at least partially via stimulating alternative splicing coupled nonsense-mediated decay (AS-NMD) of the WSB1 (WD repeat and SOCS box-containing protein 1) gene [121].

Colorectal cancer (CRC) cell proliferation relies on the activity of the RBP PTBP3 that monitors the expression of the E3 ubiquitin ligase UBE4A by binding its 3′UTR to impair UBE4A mRNA degradation and allows UBE4A to mediate P53 decomposition and stimulate CRC progression [122]. This process exemplifies the oncogenic function of PTBP3 and turns it into a promising therapeutic target for CRC in humans.

The two RNA/DNA-binding proteins, DHX9 helicase known to be overexpressed in cancer, are required for transcriptional termination and are involved in genome stability maintenance, interact with PARP-1 helicase to inhibit R-R-loop accumulation, which triggers DNA damage in human cells [123]. As these two proteins exerted a role in the cancer transcriptional program, they can be a promising target in a cancer treatment program. Given the compelling genetic evidence that RBPs exert a critical role in human disease, approaches using RBPs as a biomarker are of significant therapeutic interest, especially in the diagnosis or treatment of human cancer.

Furthermore, current evidence suggests that RBPs in various cell types serve as platforms for lncRNAs to regulate the cognate signaling pathways. The long non-coding RNA Urothelial carcinoma-associated 1 (UCA1) physically interacts with RBP (PTBP1)-regulated erythroid maturation and recruits PTBP1 to ALAS2 (5-aminolevulinic acid synthase 2) mRNA to stimulate ALAS2 mRNA stability and monitor heme biosynthesis as well as human erythroid maturation [124]. Defects in the regulation of RBPs activities by lncRNA drive genome instability in mammalian cells. The nuclear lncRNA MIR31HG interacts with the RBP YBX1 to allow its phosphorylation at serine 102 by the kinase RSK resulting in the induction of IL1A translation which stimulates the expression and secretion of a subset of senescence-associated secretory phenotype (SASP) components (interleukins and chemokines such as interleukin-6 (IL6), IL8 and CXCL1) during BRAF-induced senescence. Therefore, YBX1 can be used as a potential therapeutic target to promote senescence in cancer cells to repress cancer progression [125].

Interestingly, recent pieces of evidence have shown that RBPs are also responsible for the biogenesis circRNAs (circular RNAs) in cancer. In the line with this evidence, the role of the RBP (RBM3), acting as an oncogene critical for the formation of circRNA derived from the 3’UTR of the stearoyl-CoA desaturase (SCD) gene (so-called SCD-circRNA 2 stearoyl-CoA desaturase (SCD)-circRNA 2) in hepatocellular carcinoma (HCC) cells, has been unveiled [126]. It is worth noting that circRNAs such as hsa_circ_0001394 are suggested to be the key players in the proliferation, migration, and invasion of HCC cells [127].

Tools for identifying and analyzing RNA-interactome of RBPs

In recent years many RBPs have been identified and the mechanism regulating the activity of some of them has been elucidated, yet their precise function in the interactome involving RNAs is still under investigation. The complete elucidation of key elements or structures underlying protein-RNA interactions has the potential to greatly aid in analyzing the functionality of both the RNA and RNA-binding protein implicated as a large proportion of RBPs are still of unknown function.

Conventional approaches for elucidating RNA-interactome of RBPs or motifs in the RNAs and sequence-specific binding by a given RBP included a range of crosslinking and immunoprecipitations (eCLIP-seq) followed by RNA-sequencing [128]. To determine the functional relevance of RNA binding, a UV RNA crosslinking (CRAC) for a target protein and RNA in parallel with ChIP-seq experiments can be performed [129]. For example, individual-nucleotide resolution ultraviolet (UV) crosslinking and immunoprecipitation (iCLIP) are coupled to examine the potential role of RNA binding in CDK11 (cyclin-dependent kinase 11) functions. Crucially, the data revealed that this non-canonical RBP (CDK11) specifically monitors the expression of Replication-dependent histone (RDH)-coding genes. It binds to RDH RNAs and FLASH-containing RDH chromatin via its N-terminal RNA-binding region and associates with the chromatin of RDH genes in a cell-cycle-dependent manner. CDK11 is further demonstrated to phosphorylate Ser2 in the CTD of RNAPII positioned on the RDH genes to specifically control their transcriptional elongation and recruitment of 3′-end processing factors [130].

RBPs can be recognized by combining eCLIP with orthogonal approaches including in vitro assessment of RNA affinity for the same RBPs, chromatin association by ChIP–seq, and functional evaluation of transcriptome changes by RBP depletion and RNA-seq [131]. To identify the complete set of RBPs from yeast to humans, some new high-throughput technologies or models have been employed to detect RBPs during interactions between proteins and nucleic acids. For example, a tool combining peroxidase-catalyzed proximity labeling (PL) with organic-aqueous phase separation of crosslinked protein-RNA complexes (so-called APEX-PS) is employed to identify RBPs with novel functions in specific subcellular compartments. Notably, using this tool, it has been found in an outer mitochondrial membrane (OMM) that the RBP, SYNJ2BP, retains specific nuclear-encoded mitochondrial mRNAs at the OMM during translation stress, facilitating their local translation and channeling of protein products into the mitochondrion during stress recovery [132]. A method named STAMP (Surveying Targets by APOBEC-Mediated Profiling), which efficiently detects RBP–RNA interactions and does not rely on ultraviolet cross-linking or immunoprecipitation, is coupled with single-cell capture to identify RBP-specific and cell-type-specific RNA–protein interactions for multiple RBPs and cell types in single, pooled experiments [133]. It has been reported the use of post-transcriptional regulatory element sequencing (PTRE-seq) to examine post-transcriptional regulation by 3′ untranslated regions (UTRs) elements such as RBPs that often bind sequences in 3′UTR of mRNAs, and are known to regulate different aspects of the RNA lifecycle: RNA stability, translation efficiency, and translation initiation [134]. A deep learning tool, NucleicNet, complementing structural biology experiments, serves to provide quantitative fitness of RNA sequences for given binding pockets or to predict potential binding pockets as well as binding RNAs for previously unknown RBPs. For example, its application identified a diverse set of challenging RNA-binding proteins, including Fem-3-binding-factor 2, Argonaute 2, and Ribonuclease III [135]. A recent 3D2D model to predict binding sites of protein–RNA interaction named PRIME-3D2D has been also reported. It predicts binding sites on PDB data and transcription-wide [136]. In addition, photoswitchable RBP LicV responsive to blue light irradiation has been designed for controlling the spatiotemporal activity of RBP [137]. Additional tools for examining RBPs are highlighted in Table 3. To further comprehend protein-RNA interactions, a method for purifying protein-crosslinked RNA as a physical entity that may serve as a universal starting point for diverse downstream applications has been developed under the name XRNAX. This tool (XRNAX) can be combined with CLIP-seq to validate proteomic data, allow the discovery of the WKF RNA-binding domain, and helped both to pinpoint more than 700 proteins interacting with non-polyadenylated RNA and to observe the dynamic changes to the RNA-bound proteome [45]. A model for constructing protein-binding RNA sequences and motifs using a long short-term memory (LSTM) neural network has been employed to generate protein-binding RNA motifs comparable to the motifs from experimentally validated protein-binding RNA sequences [138]. A tool employing the ultraviolet cross-linking method and named kinetic cross-linking and analysis of cDNAs (χCRAC) has been used to quantitatively measure the global RNA-binding dynamics of the yeast transcription termination factor Nab3 in response to stress (glucose starvation) in vivo on a minute time-scale [139]. The post-transcriptional regulatory element sequencing (PTRE-seq) tool has been applied to assay the target sequence preferences and interactions between miRNAs and RBPs. It highlighted diverse facets of the RNA lifecycle encompassing RNA stability, translation efficiency, and translation initiation [134]. As most of the methods used so far to identify RBPs and their RNA-binding sites took advantage of proteins connected to polyadenylated RNAs and ignored RBPs bound to non-adenylate RNA classes (tRNA, rRNA, pre-mRNA) as well as species lacking poly-A tails in their mRNAs, a method free from specific RNA sequence or motif constraints and relying on the purification of cross-linked ribonucleoproteins (RNPs) based entirely on their physicochemical properties, and named (PTex) protocol has also been designed. It allowed capturing RBPs that bind to RNA as short as 30 nucleotides [140]. All the mentioned tools for RNA biology can serve as a universal starting point for the understanding of post-transcriptional regulatory processes in diverse cell types and organisms.

Table 3.

Additional tools for exploring proteins-RNAs interactions

Tools	Advantages	Targets components	References
Deep-RBPPred	Needs few physicochemical properties based on protein sequences	RNA-binding protein (RBP)	[141]
	Runs much faster
	Good generalization ability
rec-YnH	Detects interactions within protein libraries or between protein libraries and RNA fragment pools	Screen protein libraries against protein libraries or RNA fragment libraries	[142]
	Generate bait–prey fusion libraries
	Maps interactions between multiple RNA-binding proteins and RNAs
NucleicNet	Binding preference of RNA backbone constituents and different bases	Local physicochemical characteristics of protein	[135]
	Identify new RBPs and their binding pockets/preferences	Structure surface
	Predict potential binding pockets	Predict interaction modes for differents RNAs
	Predict binding RNAs for previously unknown RNA-binding proteins	constituents: Phosphate, Ribose, Adenine,
	Accurately recover interaction modes discovered by structural	Guanine, Cytosine, Uracil, and non-site
	Biology experiments
	Score the binding potential of individual RNA sequences
Gemini-561–o-Coral	Improved brightness and photostability	A bright and photostable fluorogen for	[143]
	High affinity	RNA imaging in cells
	Direct fluorescence imaging in live mammalian cells	RNA in living mammalian cells
	Sensitive detection of RNA inside live cells
	RNA-based fluorogenic module
Spliceosome iClip	Identifies crosslinks of endogenous, untagged spliceosomal factors on pre-mRNAs at nucleotide resolution	RBPs	[144]
		Peaks of spliceosomal crosslinking
	Interactions of diverse RBPs on pre-mRNAs during spliceosomal assembly	Around branchpoints (BPs)
		Splice sites
vIPR (in vivo Interactions	In vivo RNA–protein crosslinking and pulldown approach	Common and specific RNA-binding proteins	[145]
by ulldown of RNA)	Enables highly specific and efficient enrichment of an RNA of interest	Identify factors binding to selected mRNAs
	Enables identification of protein interactors of both transgenic and endogenous transcripts and endogenous transcripts	microRNAs (miRNAs) binding to mRNAs
Enhanced RNA interactome	Greater specificity and increased signal-to-noise ratios	RBPs	[146]
capture (eRIC)	Empowers comparative analyses of changes of RNA-bound proteomes
	Identifies m6A-responsive RNA-binding proteins that escape RNA interactome capture (RIC)
RBS-ID	Fully cleave RNA into mono-nucleosides	RNA-binding sites (RBSs)	[147]
	Replace enzymatic digestion and drastically improves the coverage (~ 2000 sites) and resolution (single amino acid level) of RNA-binding sites identification
	Reach single amino acid resolution
	View of residues important for genome editing
	RNA-specific chemical cleavage
Quantitative SILAC-based	Associate individual RNA structures within messenger RNA with their interacting proteins	RNA-binding proteins	[52]
RNA pull-down		Proteins deprive of canonical RNA-binding domains
STAMP(Surveying Targets by	Efficiently detects RBP–RNA interactions	RBP–RNA interactomes and translational landscapes	[133]
APOBEC-Mediated Profiling)	Identify RBP-specific and cell-type–specific RNA–protein interactions for multiple RBPs and cell types in single, pooled experiments
	Identify full-length RBPs that bind both polyadenylated mRNAs (RBFOX2 and TIA1) and non-polyadenylated mRNAs (SLBP
	Identifies RBP–RNA sites at isoform-specific and single-cell resolution
	Allows deconvolution of targets for multiplexed RBPs
	Uncovers cell type–specific binding of an RBP in a heterogeneous mixture
	of cell types
	Allows single-cell detection of ribosome association while simultaneously measuring gene expression
pRBS-ID	Offering direct RNA-binding evidence and identifying RBSs at single amino acid-resolution with base-specificity (U or G)	Human RBSs	[147]
	Profile uridine-contacting RNA-binding sites (RBSs)
	Discovering guanosine-contacting RBSs
	System-wide profiling of PAR-RBSs with a robust identification rate and accurate label-free quantification
RNA interactome capture (RNA-IC) and orthogonal organic phase separation (OOPS)	A global understanding of post-transcriptional gene regulation	Identify canonical and non-canonical RBPs from mouse and human CD4 + T cells RBPomes Elucidate higher-order RBP Roquin-mediated mRNA regulation	[1]
	Identify how the cellular RBPs participate or intervene with post-transcriptional control of target mRNAs
CAPRI (Crosslinked and Adjacent Peptides-based RNA-binding domain Identification)	Simultaneously identifying both crosslinked peptides (in the immediate vicinity of the RNA–protein interface) and adjacent peptides (next to the crosslinked peptide)	Map RNA-binding domains (RBDs)	[148]
	Identification of evolutionary conserved RBDs in globular domains and intrinsically disordered regions (IDRs)
FOREST (folded RNA element profiling with structure library)	Identify functional interactions from transcriptome-wide RNA structure datasets	RNA–protein (RNP) interactions	[149]
		Structure-binding preferences of RBPs
	Reveals different binding landscapes of RNA G-quadruplex (rG4)
	structures-binding proteins and
	Discovery of rG4 structures in the terminal loops of precursor microRNAs
	Discovers translational regulatory elements that function in living cells
	Evaluate a variety of RNA structures regardless of the organism
	Eliminate reverse transcription (RT) and polymerase chain reaction (PCR) amplification bias
Quantitative RNA-interactome capture (qRIC)	Combination with phosphoproteomics allowed systematic comparison of pull-down efficiencies of phosphorylated and nonphosphorylated forms of mRBPs Capture known regulatory phosphorylation sites in ELAVL1, SF3B1, and UPF1 and identify potential regulatory sites Reveal multiple phosphorylation sites in the C-terminal disordered region in the splicing regulator RBM20 affecting nucleocytoplasmic localization, association with cytoplasmic, ribonucleoprotein granules and alternative splicing	Quantify the fraction of mRNA-binding proteins (mRBPs) pulled down with polyadenylated mRNAs	[150]

Conclusion

Collectively, our review advanced our mechanistic understanding of RNA-binding protein diversity, provided a rich resource for functional studies, and identified therapeutic targets. A more comprehensive understanding of the molecular mechanism underlying the activity of the interactors of cancer progression, viral replication, and neurodegenerative disease is pivotal for the design of novel biomarkers and effective therapeutic targets for these diseases. While the biological implications of some RBPs domains remain to be explored, almost all domains reported herein are associated with important and various cellular processes. Furthermore, our review provides the basis for measuring the physiological impact of RBPs. The outcome of this review could aid to identify novel drugs for better treatment of RBPs-related disease.

Abbreviations

ALS

Amyotrophic lateral sclerosis

ATP

Adenosine triphosphate

CRC

Colorectal carcinoma

DDX

DEAD-box RNA helicase

lncRNAs

Long non-coding RNAs

PRC2

Polycomb repressive complex 2

FUS

Fused in sarcoma

hnRNP-K

Heterogeneous nuclear ribonucleoprotein K

HCC

Hepatocellular carcinoma

RBP

RNA-binding protein

NSCLC

Non-small cell lung cancer

LCD

Intrinsically disordered low-complexity domain

miRNAs

MicroRNAs

MLL

Mixed lineage leukemia

ZFP36

Zinc finger protein 36

eIF

Eukaryotic initiation factor

PARP-1

Poly(ADP-ribose) polymerase-1

IGF2BP

Insulin-like growth factor 2 mRNA-binding proteins

RRM

RNA recognition motif

RNP

Ribonucleoprotein

METTL3

Methyltransferase 3, N6-adenosine-methyltransferase complex catalytic subunit

TDP-43

Transactive response DNA-binding protein 43 kDa

Author contributions

DRAS conceived, designed and wrote the review paper. CC and DO analyzed the data and made the proofread of the manuscript. All authors read and approved the manuscript.

Funding

This work received no external funding.

Availability of data and materials

Not applicable.

Declarations

Conflict of interest

The authors declare no competing interests.

Ethical approval and consent to participate

Not applicable.

Consent for publication

Not applicable.