The RNA-binding protein RRP1 brakes macrophage one-carbon metabolism to suppress autoinflammation

Abstract

RNA-binding proteins (RBP) are important for the initiation and resolution of inflammation, so better understanding of RBP-RNA interactions and their crosstalk with metabolism may provide alternate targets to controlling inflammation. Here we establish global RNA-protein interactome purification (GRPIp) to profile the RBP landscape in inflammatory primary macrophages and identify ribosomal RNA processing 1 (RRP1) as a suppressor of inflammatory innate responses. Mechanistically, RRP1 binds nuclear thymidylate synthetase (Tyms) transcript and decreases TYMS expression post-transcriptionally in inflammatory macrophages, consequently suppressing folate metabolism cycle and inhibiting one-carbon metabolism-driven inflammation. Myeloid-specific RRP1-deficient mice develop severe experimental arthritis with increased pro-inflammatory cytokines and immunologic injury. Meanwhile, in patients with rheumatoid arthritis, RRP1 expression in peripheral blood monocytes negatively correlates with TYMS expression and serum IL-1β levels. Our results thus suggest that RRP1 acts as an anti-inflammatory factor through braking one-carbon metabolism post-transcriptionally, thereby implicating potential strategies for controlling autoinflammation.

Fig. 1. Global RNA-Protein Interactome purification (GRPIp) screening identifies RRP1 in IL-1-stimulated primary macrophages.

a Schematic of the GRPIp method (See Methods). b Distribution of RNA-RBP complex extracts after proteinase K or RNase A digestion analyzed by agarose gel electrophoresis. Line 1: the purified total RNA from the BMDMs. Line 2 to Line 4: RNA-RBP extracts purified from the interphase. (QC test ① mentioned in the Schematic). c Distribution of samples after RNase digestion, trypsin/Lys-C fragmentation by HPLC analysis (QC test ② mentioned in the Schematic). GO analysis (d) and Venn diagram (e) of potential RBPs identified in BMDMs stimulated by IL-1β for 0 h, 1 h and 12 h. f, g RT-qPCR detection of Il6 and Il-1b mRNA levels in BMDMs with siRNA-mediated RRP1 silence followed by IL-1β stimulation for the indicated hours. h ELISA of IL-6 levels in the supernatant of BMDMs with siRNA-mediated RRP1 silence followed by IL-1β stimulation for the indicated hours. Data are representative of three independent experiments (b, c). Data are of technical replicates from three independent experiments (n = 3) (d, e) with hypergeometric test (one-tailed) (d). Data are presented as means ± SD of (f, g, n = 3; h, n = 4) biologically replicates from three independent experiments with student’s t test (two-tailed unpaired). Source data are provided as a Source Data file.

Fig. 2. The pro-inflammatory phenotype of RRP1-deficient macrophages and the phenotype rescue by RRP1 NOP52 domian.

RT-qPCR detection of IL1B (a), IL6 (b) mRNA levels in THP-1 cells upon RRP1 silencing and IL-1β stimulation for the indicated hours. RT-qPCR detection of IL-1b (c) and Il6 (d) mRNA levels in wild-type (WT) or RRP1 knock out (RRP1 KO) RAW 264.7 cells with IL-1β stimulation for the indicated hours. e RT-qPCR detection of Tnfα mRNA levels in WT or RRP1 KO RAW 264.7 cells with murine TNFα (10 ng/ml) stimulation for the indicated hours. f Cell death ratio of WT or RRP1 KO RAW 264.7 cells treated with TCZ. TCZ: Combination of murine TNFα (20 ng/ml), cycloheximide (CHX) (10 μg/ml) and Z-IETD-FMK (2 μM). See methods. Western blotting of the phosphorylation (p-)of the key inflammatory signal pathway molecules in WT and RRP1 KO RAW 264.7 cells respectively stimulated by murine IL-1β (50 ng/ml) (g), TNFα (10 ng/ml) (h), IL-6 (20 ng/ml) (i) for indicated hours. GAPDH serves as loading control. j Western blot (left) and quantitative analysis (right) of the Flag tagged full-length RRP1 (FL), Nop52 and C terminal (Cter) expression in RRP1 KO RAW 264.7 cells. Empty vector (Vector) was transfected as a control. The Flag band intensities were normalized to GAPDH. RT-qPCR detection of IL-1b (k), Il6 (l) mRNAs after over-expressing the Flag tagged full-length RRP1 (FL), Nop52 (Nop52) and C terminal (Cter) in RRP1 KO RAW 264.7 cells followed by IL-1β stimulation. WT and the RRP1 KO cells transfected with empty vector (Vector) were used as controls. Western blotting data are representative of three independent experiments. RT-qPCR and Western bloting quantitative data are presented as means ± SD of (a–e, j right, k, l, n = 3) biologically replicates from three independent experiments with student’s t test (two-tailed unpaired). Cell death data are presented as means ± SD of (n = 4) biologically replicates from three independent experiments with ANOVA-test (two way). Source data are provided as a Source Data file.

Fig. 3. One-carbon metabolism related RNAs increasingly bind to RRP1 in IL-1β mediated inflammation.

a Volcano plot showing of RNA molecules that differentially bind to RRP1 in BMDMs between 0 h and 12 h of IL-1β stimulation. b Heatmap showing metabolites in BMDMs stimulated by IL-1β for 12 h or not (n = 6). c KEGG analysis of metabolic pathway-related gene sets of RIP-seq identified RNAs that increasingly binding to RRP1 following IL-1β treatment. d Schematic of 1C-metabolism, including the folate cycle and the downstream methionine cycle. Met methionine, SAM S-adenosyl methionine, SAH S-adenosyl homocysteine, Hcy homocysteine, VB12 vitamin B12, THF tetrahydrofolate, 5, 10-mTHF N^5,10-methylene- tetrahydrofolate, 5-mTHF N⁵-methyltetrahydrofolate, DHF dihydrofolate, DHFR dihydrofolate reductase, TYMS thymidylate synthetase. e Enrichment of 1C-metabolic-related gene sets of RIP-seq identified mRNAs that differentially binding to RRP1 following IL-1β treatment. f RIP-qPCR detects RRP1 binding to the Tyms mRNA, as assessed by two pairs of primers in BMDMs stimulated by IL-1β for indicated hours. g Western blot detects the Flag tagged full-length RRP1 (FL), Nop52 and C terminal (Cter) expression in NIH/3T3 cells as long as the immunoprecipitation (IP) effect by using the anti-Flag beads. The data were quantified with Image J. h RIP-qPCR detects the Flag tagged full-length RRP1 (FL), Nop52 and C-terminal (Cter) binding to Tyms mRNA in NIH/3T3 cells followed with IL-1β treatment for indicated hours. The cells transfected with empty Vector were used as controls. The data were normalized as (2^{^}-ΔCt _{RIP RNA} / 2^{^}-ΔCt _{Input RNA}) / (integrated density _{IP protein} / integrated density _{Input protein}). RIP-seq data are presented with normalized FPKM values of technical replicates (n = 3) from three independent experiments (a, e). Metabolomics data are presented with normalized response peak area values (z-score) (b). KEGG enrichment are used by hypergeometric test (one-tailed) with FDR correction (Q-values) (c). Western blotting data is representative of three independent experiments. RT-qPCR data are presented as means ± SD of (f, h, n = 3) biologically replicates from three independent experiments with student’s t test (two-tailed unpaired). Source data are provided as a Source Data file.

Fig. 4. One-Carbon metabolism is activated to support autoinflammation in RRP1 knock-out macrophages.

a Heatmap of 1C-metabolites in wild-type (WT) and RRP1 KO RAW 264.7 cells upon IL-1β induction for 12 h. Color indicates normalized intensities (z-score). b ELISA detection of SAM, methionine (Met), THF levels in RRP1 KO RAW 264.7 cells after over-expressing the Flag tagged full-length RRP1 (FL), Nop52 and C-terminal (Cter) followed by IL-1β stimulation for 8 h. The RRP1 KO cells transfected with empty Vector were used as controls. c Ratio of intracellular serine (M + 3) uptake from exogenous L-serine-[¹³C₃], out of the total serine pools in WT and RRP1 KO RAW 264.7 cells upon IL-1β induction for 4 h or not. The ratio is show as Mass Distribution Vector (MDV). d Abundance of intracellular SAM total pools in WT and RRP1 KO RAW 264.7 cells upon IL-1β induction for 4 h or not. The ratio is show as Mole Percentage Enrichment (MPE). Ratio of intracellular SAM (M + 3) (e), methionine (M + 1- M + 3) (f), methionine (M + 4) (g) derived from L-serine-[¹³C₃], out of their respective total pools in WT and RRP1 KO RAW 264.7 cells upon IL-1β induction for 4 h or not. The ratio is show as MDV. h Schematic of L-serine-[¹³C₃] labeling patterns. i–l RT-qPCR quantification of Il-1b, Il6 mRNAs in WT or RRP1 KO RAW 264.7 cells stimulated with IL-1β for indicated hours in medium replenishment with exogenous SAM, methionine (Met) and serine. ELISA, Isotope Tracers and RT-qPCR data are presented as means ± SD of (b, i–l, n = 3; c–g, n = 4) biologically replicates from more than two independent experiments with student’s t test (two-tailed unpaired). Source data are provided as a Source Data file.

Fig. 5. Defect of TYMS rescues the pro-inflammatory phenotype of RRP1 -deficient macrophages.

RT-qPCR detection of Il-1b (a) and Il6 (b) mRNA levels in RRP1 KO RAW 264.7 cells upon TYMS silencing and IL-1β stimulation for the indicated hours. c Western blot detection of RRP1 and TYMS protein levels in RRP1 KO RAW 264.7 cells upon TYMS silencing targeted by two distinct siRNAs. GAPDH serves as loading control. d RT-qPCR detection of IL-1B mRNA levels in negative control (NC)- and RRP1, TYMS or double-silenced THP-1 cells stimulated with IL-1β for the indicated hours. e ELISA detection of IL-1β levels in the supernatant of NC- and RRP1, TYMS or double-silenced THP-1 cells stimulated with IL-1β for the indicated hours. f Heatmap of 1C-metabolites in RRP1 KO RAW 264.7 cells upon TYMS silencing and IL-1β stimulation for the indicated hours. Color indicates normalized intensities (z-score). g ELISA detection of SAM, methionine (Met), THF levels in RRP1 KO RAW 264.7 cells upon TYMS silencing and IL-1β stimulation for the indicated hours. h RT-qPCR detection of Il-1b and Il6 mRNA levels in RRP1 KO RAW 264.7 cells pre-treated with Raltitrexed in the indicated concentrations and IL-1β stimulation for the indicated hours. i Western blotting for TMYS in RRP1-silenced BMDMs stimulated by IL-1β for the indicated hours. GAPDH serves as loading control. Western blotting for TMYS in the WT and RRP1 KO RAW 264.7 cells upon poly(I:C) (j) or murine IL-6 (k) stimulation for the indicated hours. GAPDH serves as loading control. Western blotting data are representative of three independent experiments (c, i, j, k). Metabolome data are from three biologically independent samples. RT-qPCR data are presented as means ± SD (n = 3) of biologically samples from three independent experiments with student’s t test (two-tailed unpaired) (a, b, d, e, g, h). Source data are provided as a Source Data file.

Fig. 6. Nuclear RRP1 suppresses TYMS expression post-transcriptionally in IL-1β-stimulated macrophages.

a Spectrophotometer (A260) detection of the fractions in RAW 264.7 cells after sucrose density gradient centrifugation (see Methods section). b RT-qPCR analysis of Tyms mRNA levels residing on the ribosome (mainly in Fraction 11). c RT-qPCR analysis of the cytoplasmic/nuclear ratio of mature Tyms mRNA (Across Exon2 and Exon3) in wild-type and RRP1 KO RAW 264.7 cells stimulated by IL-1β for the indicated time. RT-qPCR detection of the mature mRNA (across exon2 and exon3, d) and the newly transcribed nascent transcription (intron3, e) of Tyms in RRP1 KO RAW 264.7 cells upon IL-1β stimulation or not. f Schematic of the reads coverage of Tyms mRNA binding to Flag-tagged RRP1 by seCLIPs performed in RAW 264.7 cells upon IL-1β stimulation or not. g Surface plasmon resonance analysis of binding between the in vitro transcribed RNAs corresponding to the first intron of Tyms mRNA identified by CLIP-seq and the recombinant RRP1 proteins at different concentrations. KD, equilibrium dissociation constant. Immunofluorescence images (h) of RRP1 in RAW 264.7 cells stimulated by IL-1β for the indicated hours and the quantitative analysis (i, >20 cells per field, see methods section). Data are presented as means ± SD of (b, n = 4; d, e, n = 3) technical replicates from three independent experiments with student’s t test (two-tailed unpaired). Data are presented as means ± SEMs of (c, n = 3) technical replicates from three independent experiments with ANOVA test (two way). Data are representative of three independent experiments (a, g, h). Source data are provided as a Source Data file.

Fig. 7. RRP1 restrains autoinflammation in experimental autoimmune model and rheumatoid arthritis patients.

a Images of the ankle joint and the entire paw of Rrp1^fl/fl Csflr-IRES-Cre+ (condition knockout, cKO) mice and the Rrp1^fl/fl Csflr-IRES-Cre- (wilde-type, WT) littermates on day13 in the CAIA model. Intravenous injection with Normal saline (NS) was as control. b Clinical scores of Rrp1^fl/fl Csflr-IRES-Cre+ mice and the Rrp1^fl/fl Csflr-IRES-Cre- littermates of CAIA model mice (n = 10 mice per group). c ELISA of the indicated cytokines in the sera of Rrp1^fl/fl Csflr-IRES-Cre+ mice and Rrp1^fl/fl Csflr-IRES-Cre- littermates of CAIA model on day13 (n = 7 mice per group). H&E staining of the carpal joints (d) or knee joints (e) of Rrp1^fl/fl Csflr-IRES-Cre+ (condition knockout, cKO) mice and Rrp1^fl/fl Csflr-IRES-Cre- (wilde-type, WT) littermates on day13. The black arrows indicate the sites of inflammation. f Schematic for pharmacological inhibition of TYMS in established CAIA models (see methods section). g H&E staining of the carpal joints or knee joints of Rrp1^fl/fl Csflr-IRES-Cre+ mice treated with Raltitrexed (right) and the vehicle-treated control littermates (left) on day13. The black arrows indicate the sites of inflammation. h Clinical scores of Rrp1^fl/fl Csflr-IRES-Cre+ mice treated with Raltitrexed (10 mg/kg mouse body weight) and the vehicle-treated control littermates in CAIA model (n = 5 mice per group). Correlation analysis of the correlation between RRP1 (i) or TYMS (j) relative mRNA level in PBMCs and IL-1β serum levels in patients with RA. k Immunofluorescence analysis of RRP1 protein in PBMC nuclei from patients with RA comparing the RRP1-high-expression-group (H_1, H_2) and RRP1-low-expression-group (L). PBMC samples were classified into high and low expression based on the median value of relative RRP1 mRNA fold change. Data are presented as means ± SEMs from (b, n = 10; h, n = 5) biological mouse samples from two independent experiments with ANOVA test (two way). Data are presented as means ± SEMs from (n = 7) biological mouse samples from two independent experiments with student’s t test (two-tailed unpaired) (c). Data are representative of more than three biological samples from two independent experiments (a, d, e). Data are representative of (n = 6 per group) biological samples (k). The correlation between different variables was computed with nonparemetric Spearman correlation (two-tail) (i, j) (n = 50). Source data are provided as a Source Data file.