The RNA-binding ubiquitin ligase MKRN1 functions in ribosome-associated quality control of poly(A) translation

Abstract

Background: Cells have evolved quality control mechanisms to ensure protein homeostasis by detecting and degrading aberrant mRNAs and proteins. A common source of aberrant mRNAs is premature polyadenylation, which can result in non-functional protein products. Translating ribosomes that encounter poly(A) sequences are terminally stalled, followed by ribosome recycling and decay of the truncated nascent polypeptide via ribosome-associated quality control.

Results: Here, we demonstrate that the conserved RNA-binding E3 ubiquitin ligase Makorin Ring Finger Protein 1 (MKRN1) promotes ribosome stalling at poly(A) sequences during ribosome-associated quality control. We show that MKRN1 directly binds to the cytoplasmic poly(A)-binding protein (PABPC1) and associates with polysomes. MKRN1 is positioned upstream of poly(A) tails in mRNAs in a PABPC1-dependent manner. Ubiquitin remnant profiling and in vitro ubiquitylation assays uncover PABPC1 and ribosomal protein RPS10 as direct ubiquitylation substrates of MKRN1.

Conclusions: We propose that MKRN1 mediates the recognition of poly(A) tails to prevent the production of erroneous proteins from prematurely polyadenylated transcripts, thereby maintaining proteome integrity.

Keywords: MKRN1; Poly(A); RNA binding; Ribosome-associated quality control; Translation; Ubiquitin remnant profiling; Ubiquitylation; iCLIP.

Fig. 1

MKRN1 interacts with PABPC1 and other regulators of translation and RNA stability. a Protein interactome of GFP-MKRN1^wt in HEK293T cells analyzed by quantitative MS-based proteomics. Combined SILAC ratios (n = 3 replicates) after z-score normalization are plotted against log₁₀-transformed intensities. 1100 protein groups were quantified in at least two out of three replicate experiments. MKRN1 and significant interactors are highlighted (FDR < 5%). b MKRN1 and PABPC1 associate with polysomes. A 10–50% sucrose gradient of cycloheximide-treated HEK293T cell extracts. Shown are the Western blot analyses of individual gradient fractions with antibodies against MKRN1 and PABPC1/3 (n = 3 replicates, plus one technical replicate). UV absorbance was measured at λ = 254 nm. Replicates and uncropped gel images are shown in Additional file 3: Figure S10A-C. c A PAM2 motif similar to the previously reported consensus (shown on top; Additional file 1: Figure S1B) [22] is present in MKRN1 (first amino acid position indicated on the left). Introduced mutations in MKRN1^PAM2mut are indicated in petrol below. Relevant positions are highlighted (Additional file 1: Figure S1B). d Endogenous PABPC1 interacts strongly with MKRN1^wt and MKRN1^RINGmut, but only to a lesser extent with MKRN1^PAM2mut. Western blots for endogenous PABPC1 and GFP after AP of GFP-MKRN1 (wt and mutants). Ratios of PABPC1 signal (normalized to input) in GFP-MKRN1 APs over control (GFP empty vector, EV) are shown below. Replicates 2, 3, and uncropped gel images are shown in Additional file 3: Figure S10D-F. e Quantitative comparison of the interactomes of GFP-MKRN1^wt and GFP-MKRN1^PAM2mut shows that PABPC1 and several other interactors are lost upon PAM2 mutation. Combined ratios of three replicates are shown in a scatter plot. Only proteins detected in at least two out of three replicates are shown. MKRN1^wt significant interactors (from a) are highlighted as in a (FDR < 5% in MKRN1^wt). f MKRN1^WT and MKRN1^PAM2mut, but not MKRN1^RINGmut, efficiently autoubiquitylate. In vitro ubiquitylation assays with recombinant His-tagged MKRN1^wt and mutant proteins that were incubated with or without the E2 enzyme UBC5a, the E1 enzyme UBA1, and ubiquitin. A reaction with UBC5a only served as a control. Autoubiquitylation was analyzed by Western blot. Replicates 2, 3, and uncropped gel images are shown in Additional file 3: Figure S10G,H

Fig. 2

MKRN1 binds upstream of A-rich stretches in 3′ UTRs. a MKRN1 binds upstream of A-rich stretches in the 3′ UTR of the LARP1 gene. Genome browser view of GFP-MKRN1 iCLIP data showing crosslink events per nt (merged replicates) together with binding sites (lilac) and associated A-rich stretches (dark green). b MKRN1 predominantly binds in the 3′ UTR of protein-coding genes. Pie charts summarizing the distribution of MKRN1 binding sites to different RNA biotypes (7331 binding sites, top) and different regions within protein-coding transcripts (6913 binding sites, bottom). c MKRN1 binding sites display a downstream enrichment of AAAA homopolymers. Frequency per nucleotide (nt) for four homopolymeric 4-mers in a 101-nt window around the midpoints of the top 20% MKRN1 binding sites (according to signal-over-background; see the “Materials and methods” section). d MKRN1 crosslink events accumulate upstream of A-rich stretches. Metaprofile (top) shows the mean crosslink events per nt in a 201-nt window around the start position of 1412 MKRN1-associated A-rich stretches in 3′ UTRs. Heatmap visualization (bottom) displays crosslink events per nt (see color scale) in a 101-nt window around the MKRN1-associated A-rich stretches. e MKRN1 binding site strength (signal-over-background, SOB) increases with the number of continuous A’s within the A-rich stretch. Mean and standard deviation of MKRN1 binding site strengths associated with A-rich stretches harboring continuous A runs of increasing length (x-axis). MKRN1 binding sites without associated A-rich stretches are shown for comparison on the left. Number of binding sites in each category indicated as bar chart above

Fig. 3

MKRN1 binds at poly(A) tails. a Unmapped MKRN1 iCLIP reads display increased A-content (more than half of all nucleotides in the read), evidencing poly(A) tail binding. Cumulative fraction of iCLIP reads (y-axis, merged replicates) that could not be mapped to the human genome (see the “Materials and methods” section) and show at least a given A-content (x-axis). iCLIP data for the unrelated RBP HNRNPH [33] are shown for comparison. Cumulative percentage of reads with a minimum number of terminal A’s is displayed as an inset. b MKRN1 crosslink events increase towards 3′ UTR ends. Metaprofile of MKRN1 crosslink events and seven additional RBPs shows the normalized sum of crosslink events per nt in a 2001-nt window around annotated polyadenylation sites of transcripts with > 1 kb 3′ UTRs. Gray bars indicate windows in 3′ UTR body (− 750 to − 650) and close to the poly(A) site (− 150 to − 50), which were used to calculate enrichment factors. c MKRN1 binds near the polyadenylation site of the SRSF4 gene. Genome browser view as in Fig. 2a. d Overall RNA binding of MKRN1 is strongly reduced when abrogating PABPC1 interaction. Autoradiograph (left) of UV crosslinking experiments (replicate 1, with 4SU and UV crosslinking at 365 nm; replicates 2 and 3 in Additional file 1: Figure S6A,B) comparing GFP-MKRN1^PAM2mut with GFP-MKRN1^wt at different dilution steps for calibration. Quantification of radioactive signal of protein-RNA complexes and corresponding Western blot shown on the right. Uncropped gel images are shown in Additional file 3: Figure S11A,B. e MKRN1 is recruited to poly(A) RNA with the help of PABPC1. SDS-PAGE (Coomassie staining) shows recovery of recombinant His-MKRN1^wt and/or His-PABPC1 (marked by petrol and gray arrowheads, respectively) from pulldown of biotinylated RNA oligonucleotides, containing the last 22 nt of the SRSF4 3′ UTR followed by 20 A (A₂₀ RNA) or 20 V nucleotides (A or C or G; control RNA). Beads without RNA served as controls. Replicates and uncropped gel images are shown in Additional file 1: Figure S6C,D and Additional file 3: S11C-E, respectively

Fig. 4

MKRN1 stalls ribosomes at poly(A) sequences. a The dual fluorescence reporter harbors an N-terminal GFP, followed by a FLAG-SR-X linker and a C-terminal RFP, which are separated by P2A sites to ensure translation into three separate proteins [8]. The resulting GFP:RFP ratio was determined using flow cytometry. The inserted fragment K(AAA)₂₀ encodes 20 lysines by repeating the codon AAA. The starting vector without insert (K₀) served as control. Schematic ribosomes illustrate translation of the respective reporter segments. b Ribosomes fail to stall in the absence of MKRN1. HEK293T cells were transfected with control siRNA or siRNAs targeting MKRN1 (KD1 and KD2) or ZNF598 for 24 h, followed by transfection of the reporter plasmids for 48 h. Western blots for KDs are shown in Additional file 1: Figure S8A. RFP and GFP signals were analyzed by flow cytometry. Median RFP:GFP ratios, normalized to K₀ in control, are shown. Error bars represent s.d.m.; P values indicated above (paired two-tailed Student’s t test, Benjamini-Hochberg correction, n ≥ 6 replicates; ns, not significant). Analyses for inserts coding for 12 lysines (K(AAA)₁₂) and ten arginines (R(CGA)₁₀) in the dual fluorescence reporter are shown in Additional file 1: Figure S7A. c Expression of MKRN1^wt can rescue ribosome stalling. HEK293T cell lines with stable integrations of siRNA2-insensitive MKRN1 wild type and mutant constructs, or empty vector, were transfected with MKRN1 siRNA2 for 24 h, followed by transfection of the reporter plasmids for 48 h. RFP and GFP signals were analyzed by flow cytometry. Median RFP:GFP ratios, normalized to K₀ in WT cells, are shown. Error bars represent s.d.m.; P values indicated above (paired two-tailed Student’s t test, Benjamini-Hochberg correction, n = 6 replicates; ns, not significant). Analyses for reporter plasmids with inserts coding for K(AAA)₁₂ or R(CGA)₁₀ are shown in Additional file 1: Figure S7B. d MKRN1 knockout (MKRN1 KO) and wild type (WT) HEK293T cells were transfected with the reporter plasmids for 48 h. Measurements, analyses, and visualization as in c (n = 4 replicates). Analyses for reporter plasmids with inserts coding for K(AAA)₁₂ or R(CGA)₁₀ are shown in Additional file 1: Figure S7C (d)

Fig. 5

MKRN1 ubiquitylates ribosomal protein RPS10 and translational regulators. a Ubiquitin remnant profiling to compare the relative abundance of ubiquitylation sites in MKRN1 KD2 and control HEK293T cells. Ubiquitin remnant peptides were enriched and analyzed by quantitative mass spectrometry, quantifying a total of 15,528 ubiquitylation sites on 4790 proteins. 29 putative MKRN1 target sites with significantly decreased ubiquitylation upon MKRN1 KD2 (FDR < 10%, n = 4 replicates) are highlighted and labeled with the respective protein name. Note that many proteins contain several differentially regulated ubiquitylation sites. b Protein interaction network of 21 proteins with putative MKRN1 ubiquitylation target sites (significantly reduced, shown in a). The functional interactions were obtained from the STRING and BioGrid databases and our study. Visualization by Cytoscape. c Ubiquitin remnant profiling results for significantly regulated ubiquitylation sites (FDR < 10%) in proteins from network in b. Mean and standard deviation of the mean (s.d.m., error bars) are given together with all data points. d Comparison of interactome of GFP-MKRN1^wt (WT over GFP, see Fig. 1a) with putative MKRN1 ubiquitylation substrates from ubiquitin remnant profiling (UB, see a). Protein names are given for all ubiquitylation substrates. e Ubiquitin remnant profiling results for seven quantified ubiquitylation sites in RPS10 and RPS20. Significant changes are shown in black (FDR < 10%) and non-significant changes in gray. Representation as in c. f Comparison of ubiquitylation sites in the target proteins RPS10 (UniProt ID P46783), RPS20 (P60866), PABPC1 (P11940), PABPC4 (B1ANRO), IGF2BP1 (Q9NZI8), IGF2BP2 (F8W930), and IGF2BP3 (O00425) that are modified by ZNF598 and MKRN1 during RQC. g MKRN1 ubiquitylates RPS10 and PABPC1 in vitro. His-RPS10 (left) or His-PABPC1 (right) were incubated with or without His-MKRN1. Ubiquitylation of the target proteins was assessed by Western blot. Replicates 2, 3, and uncropped gel images are shown in Additional file 3: Figure S12I,J for RPS10 and Additional file 3: Figure S12K,L for PABPC1

Fig. 6

MKRN1 is a sensor for poly(A) sequences that stalls ribosomes to initiate ribosome-associated quality control. Proposed model of MKRN1 function: MKRN1 is positioned upstream of (premature) poly(A) tails via interaction with PABPC1. Ribosomes translating the open reading frame run into MKRN1 that acts as a roadblock to prohibit poly(A) translation. Upon contact with the translating ribosome, MKRN1 ubiquitylates the 40S ribosomal protein RPS10. This stalls the ribosome, causing the trailing ribosomes to collide. ZNF598 recognizes the collided ribosomes and ubiquitylates ribosomal proteins to promote RQC