Nuclear RNA cap-chaperones eIF4E and NCBP2 govern distinct fates for 1000s of mRNAs uncovering an unexpected regulatory point in gene expression

Abstract

mRNA processing constitutes a series of steps in the nucleus to generate mature mRNAs that can be translated into protein. This relies on the methyl-7-guanosine (m⁷G) "cap" on the 5'end of mRNAs which is bound by the nuclear cap-binding protein NCBP2 with its cofactor NCBP1. The NCBP1/2 complex chaperones capped mRNA through these processing steps. NCBP2 is considered the sole nuclear cap-binding factor and thus its cap-chaperone role is thought to be a constitutive, housekeeping activity. However, another cap-binding protein, the eukaryotic translation initiation factor eIF4E, is also found in the nucleus. Two cap-binding factors co-existing in the nucleus intimate an undiscovered regulatory point in gene expression or, alternatively, redundancy to ensure gene expression fidelity. Consistent with the former possibility, eIF4E and NCBP2 drove distinct gene expression, transcriptomic, and splicing signatures impacting ~2500 transcripts involved in distinct biological programmes with only ~360 transcripts in common and of these only 79 common splicing events. Thus, each cap-chaperone designates distinct mRNA populations for specific processing revealing a new step in gene expression. We denote this mRNA specification of cap-chaperones (SOCCS). We uncovered multiple molecular mechanisms that contribute to SOCCS: distinct spatial localization of eIF4E and NCBP2 within the nucleus, identification of sequence motifs within targeted mRNAs segregated by eIF4E or NCBP2 sensitivity, distinct protein partners for these cap-chaperones and differential impacts on the production of key spliceosome components e.g. U2AF1, PRP31, SF3B1 and SNRNP200 indicative of distinct transcriptomic landscapes produced by eIF4E or NCBP2 overexpression. In all, the realization that multiple cap-binding proteins coexist in the nucleus led us to identify an unexpected gene-expression regulatory point which engaged distinct biological programmes.

Figure 1:. EIF4E and NCBP2 alter the splicing of distinct targets.

A) The potential models of cap chaperones interacting with nuclear mRNA. Created with https://BioRender.com B) Western blot depicting the expression levels of EIF4E and NCBP2 overexpression in U2OS cells from two clones (a and b). C) Volcano plot depicting the genes that are significantly upregulated (blue) or downregulated (purple) upon NCBP2 overexpression in U2OS cells. D) The top gene ontology terms of genes undergoing differential expression in EIF4E (red circles) or NCBP2 (blue triangles). Size of the point indicates the number of genes. Dashed line indicates p-value = 0.05. E) Left Bar graph illustrating the number of alternative splicing events in EIF4E (red) and NCBP2 (blue). Right Network analysis of enriched gene ontology terms from genes with skipped exon (SE) and mutually exclusive exon (MXE) alternative splicing events in NCBP2 overexpression. F) Heatmap of significantly changing skipped exon events showing inclusion (red) or exclusion (blue) of the exon in each sample. G) Position of significantly changing skipped exons in following overexpression of NCBP2 (blue, left) or EIF4E (red, right). H) Top gene ontology terms of genes undergoing alternative splicing in NCBP2 only (triangle, blue), EIF4E only (circle, red), or both (square, purple) overexpression. Size of the point indicates the number of genes. Dashed line indicates p-value = 0.05. I) Euler plot comparing the genes that have significantly changing alternative splicing in eIF4E (red) or NCBP2 (blue) overexpression. The number of skipped exons (SE) and mutually exclusive exon (MXE) events that are changing in both overexpression are indicated.

Figure 2:. eIF4E and NCBP2 Interact with Distinct RBPs.

A) Quantitative PCR showing alternative splicing following NCBP2 or eIF4E overexpression. B) Euler plot showing the overlap between eIF4E RIP targets from LY1 cells (blue) and NCBP2 iCLIP targets from HeLa cells (green). C) Significantly enriched motifs found in the introns of skipped exons from genes with alternative splicing only in eIF4E (top) or NCBP2 (bottom). RNA binding proteins recognition motifs similar to the enriched motifs are listed. D) RNA binding proteins with recognition motifs similar to motifs enriched in the introns of genes undergoing alternative splicing in eIF4E (left) or NCBP2 (right) overexpressing cells. E) Western blot showing the impact of eIF4E or NCBP2 overexpression on production of each other and selected splice. Left Representative blots probed as indicated. ENDO refers to endogenous NCBP2 or eIF4E and FLAG, the overexpressed form. Right Quantitation of 3–6 biological replicates for each condition are shown. For all panels, means, standard deviations and p-values relative to vector (Welch t-test) were calculated in PRISM. * = p-value < 0.05, ** = p-value < 0.005, n.s. = non-significant

Figure 3.. A fraction of eIF4E interacts with active sites of transcription.

A) Endogenous eIF4E co-immunoprecipitated with POLR2A-S2P in the nuclear fraction of U2OS cells. Immunoprecipitations (IP) were carried out using nuclear lysates and anti-eIF4E antibody (eIF4E-IP) or appropriate IgG control (IGG), carried out at least three independent times with one representative experiment shown. Samples were analyzed by western blot (WB) using antibodies as indicated. Fractionation controls for HSP90 for the cytoplasmic fraction (Cyto), histone 2B (H2B) for the nuclear fraction (Nuc) are shown. Whole Cell lysis (WCL). B) Endogenous eIF4E co-immunoprecipitated with POLR2A-S2P in crosslinked nuclear fractions using NEXSON, one representative experiment is shown of three biological replicates. C) The impact of DNAse and RNAse on the eIF4E-POLR2A-S2P interaction were analyzed by western blot. D) Quantification of three or more biological replicates from C) showing the relative signal of each protein in control (black), RNase-treated (green), and DNase-treated conditions (blue) (p-values are provided using Student t-test in PRISM, all comparisons non indicated were non-significant). E) Confocal micrographs of U2OS cells stained with antibodies against eIF4E (red) and POLR2A-S2P (green), with nuclei counterstained with DAPI (blue) and overlay in yellow (merge). Single section through the plane of the cell is shown. Arrows indicate examples of eIF4E-only (red), POLR2A-S2P-only (green), or co-localized (yellow) signals H) Quantitation of colocalization using Imaris, data were collected as Z-stacks for >25 cells. Percentage colocalization over the cell volume and the Manders coefficient were calculated in Imaris and processed in PRISM. Each data point represents a single nuclei.

Figure 4.. Mechanisms for eIF4E and NCBP2 dependent splicing.

A) Endogenous eIF4E co-immunoprecipitated with NCBP1 and NCBP2 in nuclear fractions from U2OS cells. Immunoprecipitations (IP) were carried out using U2OS nuclear lysates and anti-eIF4E antibody (eIF4E-IP) or appropriate IgG control (IGG), carried out at least three independent times, with a representative experiment shown. Fractionation controls included Lamin (nuclear fraction, Nuc), MEK1 served as controls for the cytoplasmic fractions (Cyto). Whole Cell lysis (WCL). B) The impact of the RNAse and DNAse treatment on of the interaction between endogenous eIF4E and NCBP2 and NCBP1. IPs were performed using nuclear lysates from U2OS cells with or without RNase A/T1 or DNase I treatment prior to IP (as in Figure 3C). IP samples, along with input, were analyzed by western blot using the indicated antibodies. C) Quantification of three biological replicate experiments showing the relative signal of each protein in control (black), RNase-treated (green), and DNase-treated conditions (blue) (p-values are provided using Student t-test in PRISM, all comparisons non indicated were non-significant). D) Top Confocal micrographs of a single section through the plane of U2OS cells stained with antibodies against endogenous eIF4E (red) and NCBP2 (green), overlay (yellow) and nuclei with DAPI (blue). Arrows indicate examples of eIF4E-only (red), NCBP2-only (green), or co-localized (yellow) signals. Bottom Quantitation of colocalization using Imaris, data were collected as Z-stacks. Percentage colocalization over the cell volume and the Manders coefficient are shown. Each point represents a single nuclei. Data were collected as Z-stacks for >25 cells and processed in PRISM. E) FLAG Ips for nuclear lysates of NCBP2- or eIF4E-overexpressing cells showing overlapping but distinct complexes were formed. Representative experiments of two different eIF4E and NCBP2 clones in biological replicates are shown. Vector control cells and probing for NOPP140 served as negative controls.

Figure 5:. Multiple mechanisms underpin distinct splicing profiles for eIF4E and NCBP2.

Left eIF4E and NCBP2 have some nuclear co-localization, but are largely separate from one another. However, they can both binding splicing factors and markers of active transcription with notable differences such as ARS2 only binding NCBP2. Together with the unique effects on the splicing factor landscapes, this leads to changes in the splicing profile of distinct sets of mRNAs, with small overlap.