A peptide encoded by circular form of LINC-PINT suppresses oncogenic transcriptional elongation in glioblastoma

Abstract

Circular RNAs (circRNAs) are a large class of transcripts in the mammalian genome. Although the translation of circRNAs was reported, additional coding circRNAs and the functions of their translated products remain elusive. Here, we demonstrate that an endogenous circRNA generated from a long noncoding RNA encodes regulatory peptides. Through ribosome nascent-chain complex-bound RNA sequencing (RNC-seq), we discover several peptides potentially encoded by circRNAs. We identify an 87-amino-acid peptide encoded by the circular form of the long intergenic non-protein-coding RNA p53-induced transcript (LINC-PINT) that suppresses glioblastoma cell proliferation in vitro and in vivo. This peptide directly interacts with polymerase associated factor complex (PAF1c) and inhibits the transcriptional elongation of multiple oncogenes. The expression of this peptide and its corresponding circRNA are decreased in glioblastoma compared with the levels in normal tissues. Our results establish the existence of peptides encoded by circRNAs and demonstrate their potential functions in glioblastoma tumorigenesis.

Fig. 1

Translatome sequencing and proteome profiling of potential coding circRNAs in normal and cancer cells. a Illustration of the screening protocol. Briefly, total RNAs or RNC-RNAs were isolated separately from NHA or U251 cells. Equal amounts of total RNA or RNC-RNA were reverse-transcribed and subjected to deep RNA sequencing. Identified differentially expressed circRNAs were annotated in the genome, and the host genes were cross-matched between NHA and U251. b RNA-seq read abundance distribution of identified circRNAs. Upper, total RNA seq; Lower, RNC-RNA seq. X-axis: the back-spliced read numbers of circRNAs detected by sequencing. Y-axis: the abundance of circRNAs classified by different read numbers. The majority of called circRNAs in the study were supported by more than 10 reads. c Venn plot showing the number of circRNAs derived from different genomic regions. Upper, total RNA seq; lower, RNC-RNA seq. d Length distribution of the identified circRNAs. Upper, total RNA seq; lower, RNC-RNA seq. X-axis: the length of circRNAs detected in this study. Y-axis: the abundance of circRNAs classified by different lengths. e The Venn plot of the numbers of called circRNAs in NHA and U251 by RNA-seq or RNC-seq. RNC-seq identified more circRNAs due to the higher sequencing depth (see Supplementary Figure 1). f Scatter plot of all differentially expressed circRNAs between NHA and U251 cells. Upper, total RNA-seq; lower, RNC-seq (x and y axes represent circRNA expression value, RPKM). g Upper, differentially expressed circRNAs between NHA and U251 cells in total RNA or RNC-RNA were cross-matched. A total of 320 differentially expressed circRNAs were identified, generated from 274 host genes. Lower, the host genes were subjected to GO enrichment analysis (The gene expression value in heatmap was normalized by Z score in each row.)

Fig. 2

Identification of exon 2 of LINC-PINT as a circRNA. a Upper, visualization of the forward reads within the exon 2 region in the LINC-PINT junction site of NHA cell in RNA-seq and RNC-seq. These junction reads are specific for circular form of LINC-PINT exon 2. Lower, IGV plot of all reads located on exon 2 of LINC-PINT in RNA-seq and RNC-seq. The IGV plot also included the reads on exon 1 and 3 of LINC-PINT. b Illustration of the annotated genomic region of LINC-PINT (Ensembl number: ENSG00000231721), the putative different mRNA splicing forms (linear splicing and head-to-tail splicing) and the validation strategy for LINC-PINT circular exon 2 (circPINTexon2). Divergent primers detected the circular form of circPINTexon2 in cDNA but not in gDNA. Convergent primers spanning exon 1 and exon 2 of LINC-PINT (variants LINC-PINT-208, shown in a) specifically detected the linear splicing form. β-actin was used as a linear RNA control. c Sanger sequencing was performed following PCR using the indicated divergent flanking primers to confirm the head-to-tail splicing of circPINTexon2 in 293T cells. d Northern blots of 293T total RNA with the exon probe and the junction-specific circular probe for circPINTexon2. Lanes 1–4 detected circPINTexon2 with circular probes. Lanes 5–8 detected circPINTexon2 and LINC-PINT with exon probes. CircPINTexon2-overexpression plasmid was shown in Fig. 3f. e Q-PCR followed by with junction-specific primers was used to detect the expression of circPINTexon2 in vitro. Primers specific for linear LINC-PINT were also used to detect LINC-PINT expression. RNase R treatment was used to validate circPINTexon2. Data are presented as mean ± s.e.m. from three independent experiments. **P < 0.01; ns, P > 0.05, determined by two-tailed Student’s t-tests. f FISH with junction-specific probes specific to circPINTexon2, whereas linear specific probes specific to linear LINC-PINT indicated their cellular localization in vitro. Normal brain tissues and GBM samples were stained with indicated probes. Overexpressed or knocked-down circPINTexon2 or LINC-PINT using corresponding plasmids or siRNA/ASOs in 293T cells to indicate the specificity of these probes. Scale bars, 10 μM. EV empty vector, si-NC random scrambled siRNA, circPINTexon2 circPINTexon2 overexpression vector, ASO LINC-PINT anti-sense oligos

Fig. 3

circPINTexon2 encodes an 87-aa peptide. a Full-length or truncated circPINTexon2 IRES (478, 231, and 209 bp) were cloned between mCherry and GFP as indicated to construct several reporter plasmids. These plasmids were transfected into 293T cells as indicated, with or without 4EGI-1 treatment. IF was performed to determine mCherry and GFP signals. Scale bars, 50 μM. b Rluc and Luc were tandemly cloned into the luciferase reporter plasmid, with or without the indicated truncated IRES between them. Luc/Rluc activities were measured in each transfected plasmid. c RNC-RNA or total RNA from 293T cells was extracted and reverse-transcribed using oligo-dT or random primers as indicated. Specific primers for circPINTexon2 or LINC-PINT were analyzed by using q-PCR. d Upper, antibody recognition test for the predicted 87-aa peptide. Lanes 1 and 2, Coomassie blue staining of the GST-PINT87aa fusion protein; lanes 3 and 4, Western blot performed with GST antibody; lanes 5 and 6, Western blot performed with PINT87aa antibody. Lower, the predicted 87-aa peptide sequence and antibody generation region was shown as indicated. e Endogenous immunoprecipitation using anti-PINT87aa antibodies in 293T cells. LC-MS/MS analysis following SDS-PAGE was performed to identified peptide sequences of PINT87aa. f Upper panel, endogenous circPINTexon2: endogenous formation of circPINTexon2; CircPINTexon2 vector: the artificial circPINTexon2 overexpression plasmid. Note the junction was moved inside the 87-aa ORF. CicrPINTexon2 Del-IRES vector: negative control, in which the IRES sequence was deleted from the artificial circPINTexon2 plasmid. 87-aa overexpression vector: positive control, in which the 87-aa ORF was cloned downstream of a linear CMV promoter. Lower panel, PINT87aa expression was tested in 293T cells after transfection of the plasmids indicated above. g circPINTexon2 and PIN87aa expression were determined using junction-specific siRNA or shRNA specific for circPINTexon2-transfected 293T or hNSC. h Left, LINC-PINT, circPINTexon2, and PIN87aa expression were determined in two ASOs specific for LINC-PINT-transfected 293T. Right, LINC-PINT and PINT87aa were determined in LINC-PINT stably transduced 293T and hNSC. b, c, g, h Data are presented as mean ± s.e.m. from three independent experiments. **P < 0.01; ns, P > 0.05, determined by two-tailed Student’s t-tests

Fig. 4

Localization and expression of PINT87aa in cells and human tissues. a IF microscopy images of the cellular localization of the PINT87aa-RFP-fusion protein in 456 and 4121 BTIC cells showing stable overexpression. Scale bars, 20 μM. b The expression of circPINTexon2 and PINT87aa was detected in human brain, breast, liver, kidney, stomach, intestine, thyroid, and pancreas tissues. c Upper panel, the expression of circPINTexon2 and PINT87aa was detected in hNSC, 293T, Hs683, SW1783, 4121, 456, 387, and H2S cells. Lower panel, circPINTexon2 and PINT87aa levels were determined in normal brain and gliomas with different WHO grades. d Establishment of linear vector-transduced stable PINT87aa-GFP 456 and 4121 cells and artificial circPINTexon2-transduced stable PINT87aa 456 and 4121 cells. Data are presented as mean ± s.e.m. from three independent experiments. **P < 0.01; ns, P > 0.05, determined by two-tailed Student’s t-tests. e Establishment of PINT87aa K.O. SW1783 and Hs683 cells using CRISPR/Cas9 technology. Schematic illustration showing that the genomic regions of the LINC-PINT exon2 contained the 87-aa ORF. gRNAs designed to target the PINT87aa ORF are shown. Genomic K.O. effects were confirmed via Sanger sequencing, as shown in Supplementary Fig. 7. Western blotting revealed the effects of PINT87aa K.O. in SW1783 and Hs683 cells

Fig. 5

Biological functions of PINT87aa. a Left, cell cycle was determined in PINT-87aa or circPINTexon2 stably overexpressed 456 and 4121 BTICs. Right, cell cycle analysis of PINT87aa K.O SW1783 and Hs683 cells. b MTT proliferation assay was examined in PINT87aa-GFP or circPINTexon2 stably overexpressed 456 and 4121 cells or PINT87aa K.O SW1783 and Hs683 cells. c Edu proliferation assay was examined in PINT87aa-GFP or circPINTexon2 stably overexpressed 456 and 4121 cells, PINT87aa K.O SW1783 and Hs683 cells and their control cells. d Soft-agar and plate colony assay was performed in PINT87aa K.O SW1783 and Hs683 cells and their control cells. Scale bar, 100 μM. e In vitro extreme limiting dilution assays (ELDAs) were performed to evaluate BTICs self-renewal capacity. Scale bar, 100 μM. Spheres were counted at 14 days. Cell density per well ranged from 1, 10, 25, 50, 100, 250, 500 to 1000. Each condition was tested in 10 independent wells. Neurosphere-forming capability was determined using the ELDA web-based tool (456, P < 0.001; 4121, P < 0.001, by ELDA analysis). f PINT-87aa or circPINTexon2 stably overexpressed 456 and 4121 BTICs and their control cells were subjected to 6 Gy radiation. DNA damage was determined by flow cytometry and γ-H2AX expression. a–d, f Data are presented as mean ± s.e.m. from three independent experiments. *P < 0.05; ns, P > 0.05, determined by two-tailed Student’s t-tests

Fig. 6

PINT87aa directly interacts with the PAF1 complex and inhibits mRNA transcriptional elongation. a Left, immunoprecipitation was performed using an anti-Flag antibody in PINT87aa-3XFlag or empty vector transfected 293T cells. Western blotting using an anti-Flag antibody confirmed PINT87aa overexpression. The precipitates were subjected to LC-MS/MS to identify potential PINT87aa-interacting proteins. Right, PAF1 complex-related proteins were identified in PINT87aa precipitates. b Immunoprecipitation was performed in PINT87aa-3XFlag-transfected cells or control cells. Western blotting was performed using an anti-PAF1 antibody. c Upper, PINT87aa conformations were modeled with PEP-FOLD and docked to PAF1 based on the ATTRACT2 force field using the PEP-SiteFinder pipeline. PINT87aa was split into three segments: 1–36 aa, 27–62 aa, and 53–87 aa. The top ten peptides in complex with PAF1 were visualized with different colors using PyMOL (The PyMOL Molecular Graphics System, Version 1.8 Schrödinger, LLC.), while the protein–peptide interface residues (within 5 Å for each peptide) were labeled. Lower, the direct mutual interactions of PINT87aa with different domains of HA-tagged PAF1 were tested using purified proteins (Flag-tagged PINT87aa and HA-tagged PAF1). d IF was performed to determine PINT87aa-PAF1 colocalization in PINT87aa-GFP transfected 293T cells or hNSC. Scale bar, 20 μM. e The expression of PAF1 downstream genes was determined by performing Western blotting and q-PCR in PINT87aa- or circPINTexon2-overexpressed 456 and 4121 BTIC and their respective controls. Data are resented as mean ± s.e.m. from three independent experiments. **P < 0.01, ns, P > 0.05, determined by two-tailed Student’s t-tests

Fig. 7

PINT87aa decides the localization of PAF1 complex on target genes promoter. a 2nd ChIP assay was used to determine the co-occupation of PAF1 complex/PINT87aa in CPEB1's promoter region by using RT-PCR or q-PCR. At least three independent experiments were performed. b ChIP assay was performed in circPINTexon2 stably overexpressed 456 and 4121 cells by using indicated antibodies. PCR products of indicated genes’ promoter were analyzed by using RT-PCR or q-PCR. c PINT87aa was overexpressed or knocked-down in 293T or hNSC cells by using indicated plasmid or shRNA, respectively. ChIP assay was performed by using indicated antibodies and PCR products of CPEB1 promoter region were analyzed by using RT-PCR or q-PCR. b, c At least three independent experiments were performed

Fig. 8

Clinical implications of circPINTexon2 and PINT87aa. a Expression of circPINTexon2 and PINT87aa in human glioma samples with different WHO grades and matched adjacent normal tissues. b Expression of circPINTexon2 and PINT87aa in breast cancer, hepatic cellular carcinoma, gastric cancer, and adjacent normal tissues. c Anti-cancer effects of PINT87aa in vivo; left, mice with brain xenograft tumors were sacrificed 60 days after implantation; right, Kaplan–Meier survival curves of mice implanted with 456 and 4121 cells stably overexpressing PINT87aa and their respective parental cells (N = 8 mice/treatment group, P < 0.001, two-tailed log-rank test). d Nude mice were subcutaneously injected with 5 × 10⁶ Hs683 and SW1783 PINT87aa K.O. cells and their parental cells using Matrigel and were then monitored daily. Day 0 indicates the first day of injection. Tumor progression was monitored using a small animal imaging system or through caliper measurements. Error bars indicate standard error (N = 5 mice/treatment group, P < 0.01, two-tailed log-rank test)