RNA-based translation activators for targeted gene upregulation

Abstract

Technologies capable of programmable translation activation offer strategies to develop therapeutics for diseases caused by insufficient gene expression. Here, we present "translation-activating RNAs" (taRNAs), a bifunctional RNA-based molecular technology that binds to a specific mRNA of interest and directly upregulates its translation. taRNAs are constructed from a variety of viral or mammalian RNA internal ribosome entry sites (IRESs) and upregulate translation for a suite of target mRNAs. We minimize the taRNA scaffold to 94 nucleotides, identify two translation initiation factor proteins responsible for taRNA activity, and validate the technology by amplifying SYNGAP1 expression, a haploinsufficiency disease target, in patient-derived cells. Finally, taRNAs are suitable for delivery as RNA molecules by lipid nanoparticles (LNPs) to cell lines, primary neurons, and mouse liver in vivo. taRNAs provide a general and compact nucleic acid-based technology to upregulate protein production from endogenous mRNAs, and may open up possibilities for therapeutic RNA research.

Fig. 1. taRNAs built from an array of IRESs increase reporter gene translation.

a Schematic overview of taRNA technology. taRNAs recruit initiation factors (eIFs, blue) to increase targeted mRNA translation. A taRNA molecule is made of a target-specific guide RNA domain (gRNA, green), a linker (gray) and a translation machinery-recruitment domain (purple). Poly(A)-binding protein (PABP, light gray) interacts with both eIFs and mRNA poly-A tail. b Schematic of vectors for taRNA and reporter used in dual-luciferase assay. The taRNA (green and purple) targets Firefly luciferase mRNA (Fluc, blue), while Renilla luciferase (Rluc, gray) is the internal control. The nucleotides at position −4 to −1 before the start codon of Fluc mRNA are ATTG. The hU6 (human U6), PGK (phosphoglycerate kinase 1) and SV40 (simian vacuolating virus 40) indicate different promoters. c Evaluation of viral IRESs as taRNA effector domains in HEK293T cells by dual-luciferase assay. Each IRES was attached to either a non-targeting gRNA (gray), or an Fluc-targeting gRNA (g3′-1, blue) on the taRNA vector, and co-transfected into HEK293T cells with the reporter. Viral origin and classification are listed for each IRES. Data were normalized to empty vector group. n = 8 biological replicates for empty vector group. n = 4 biological replicates for all the other groups. d RNA level of Fluc relative to Rluc in HEK293T cells after taRNA treatment (gRNA: g3′ -1, blue) measured by RT-qPCR. Data were normalized to empty vector group. n = 3 biological replicates. e PTV-based taRNAs with gRNAs that bind to different regions on Fluc transcript, including 5′ UTR (g5′), CDS (gCDS-1) and 3′ UTR (g3′-1), were evaluated by dual-luciferase assay. n = 4 biological replicates. All data are shown as mean ± SEM with individual data points. c Statistical analyses were performed using two-way ANOVA with Sidak’s multiple comparisons test between non-targeting and Fluc-targeting within each IRES group. Statistical analyses were performed using one-way ANOVA with Dunnett’s multiple comparison test d vs. vector; e vs. NT. **P < 0.01, ****P < 0.0001. No asterisk = not significant. The P value and source data are provided as a Source Data file.

Fig. 2. taRNAs enhance endogenous gene expression.

Two gRNAs were tested in PTV-based taRNAs for each endogenous mRNA target, including a human PTEN; b human PPIB; c human CDKN1A; d human ABCA7. Each gRNA is annotated with where their last nucleotide binds relative to the stop codon on the target transcript. All experiments were done in HEK293T cells transfected with plasmids expressing the indicated taRNAs and harvested for western blots after 48 h. GAPDH or α-tubulin was used as the loading control. Representative blots were shown as top panels, and quantifications normalized to non-targeting control (NT) were shown below. All bar-graph values are shown as mean ± SEM with data points. n = 4 biological replicates for (a) and (c). n = 3 biological replicates for (b) and (d). Statistical analyses were performed using one-way ANOVA with Dunnett’s multiple comparison test vs. NT. *P < 0.05, **P < 0.01. No asterisk = not significant. The P value and source data are provided as a Source Data file.

Fig. 3. taRNAs constituted of truncated effector domains increase target protein level and manipulate cellular function.

a Schematic of HCV IRES domains. Superscript numbers on nucleotides indicate their position on the HCV genome. U228 (red) is critical for eIF3 binding. b Evaluation of effector domains via dual-luciferase assay. The conserved apical domains from HCV, CSFV, and PTV-1 IRES (blue) were attached to Fluc-targeting gRNA (g3′-1). HCV-IIIabc^U228C (light gray) has an inactivating U228C mutation in the HCV-IRES IIIabc domain. Data were normalized to empty vector. n = 4 biological replicates. c Schematic of PTV-1 IIIab-based taRNA, with annotated domains and nucleotide sequence. d GAG-to-UGU mutations were introduced at the middle of Fluc-targeting gRNA in the PTV-IIIab taRNA (mis-g3′-1). This mismatched taRNA (light gray) was evaluated by dual-luciferase assay, along with original positive control (g3′-1, blue) and negative control (NT, dark gray). n = 4 biological replicates. e Representative western blot showing increased PTEN expression in HEK293T cells following transfection with g2(PTEN)-PTV-IIIab taRNA. GAPDH is the loading control. n = 7 biological replicates. f Representative western blot showing PTV-IIIab taRNA-mediated increase of SYNGAP1 protein in N2a cells. Two mSYNGAP1-targeting gRNAs (g1 and g2) were tested. α-tubulin is the loading control. n = 4 biological replicates. g Representative western blot showing PTV-IIIab taRNA-mediated increase of PMP22 protein in NIH/3T3 cells. Two mPMP22-targeting gRNAs (g1 and g2) were tested. GAPDH is the loading control. n = 3 biological replicates. h Dual upregulation of anti-proliferative proteins PTEN and p21 (gene: CDKN1A) by PTV-IIIab taRNAs inhibits MDA-MB-231 cell growth by 54%, measured by CCK-8 assay at 72 h after plating. n = 6 biological replicates. All bar-graph values are shown as mean ± SEM with data points. Statistical analyses were performed using one-way ANOVA with Sidak’s multiple comparisons test b vs. vector, and HCV-IIIabc vs HCV-IIIabc^U228C; d vs. vector, mis-g3′-1 vs. NT, and mis-g′-1 vs. g3′-1; f–h Dunnett’s multiple comparison test vs. NT. Unpaired two-tailed Student’s t test was performed in (e). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. The P value and source data are provided as a Source Data file.

Fig. 4. LNP delivery of taRNAs in vitro and in vivo.

a Nucleotide sequence and secondary structure of the PTV-IIIab-based taRNA with stabilizing hairpins added at both 5′ and 3′ ends. All taRNAs in this figure utilized this stabilized form with PTV-IIIab as their effector domain, and were in vitro transcribed and delivered by LNPs. b Non-targeting (NT) or mouse PTEN (mPTEN)-targeting taRNA (500 ng) was delivered to N2a cells. After 12 h, the cells were lysed and PTEN protein levels measured via western blot. n = 4 biological replicates. c, d LNPs containing PTEN-targeting taRNAs or non-targeting taRNAs were injected into the tail vein of mice (c) and liver tissues were collected 12 h later. Hepatic PTEN protein levels were quantified via western blot (d). n = 4 biological replicates. LNP-packaged taRNAs targeting SYNGAP1 or non-targeting control (NT) were delivered to N2a cells (e) or rat primary neurons (f). Levels of SYNGAP1 protein and phosphorylated ERK1/2 (p-ERK1/2) were evaluated at 12 h post delivery. α-tubulin and total ERK1/2 were used for normalization as indicated. n = 4 biological replicates for (e), and n = 3 biological replicates for (f). All bar-graph values are shown as mean ± SEM with data points. Unpaired two-tailed Student’s t tests were performed between each group and its NT control. *P < 0.05, **P < 0.01. The P value and source data are provided as a Source Data file. Some elements created with BioRender.com.

Fig. 5. Optimized mini taRNA rescues SYNGAP1 expression in haploinsufficiency-disease cells.

a PTV-IIIab-based taRNAs with various mouse SYNGAP1-targeting gRNAs (g1, g3, g4 and g5) were transfected into N2a cells as plasmids. The SYNGAP1 protein levels were measured by western blotting. The gRNAs are labeled with the number of nucleotides between stop codon and their landing sites (gray). The lengths of 5′ UTR, CDS and 3′ UTR of SYNGAP1 mRNA are labeled. n = 3 biological replicates. b Schematic illustrating the engineering of the PTV-IIIab-based taRNA to a minimized taRNA (mini taRNA). The detachable stabilizing hairpins on both ends (5′ hp and 3′ hp) are shadowed in gray. Nucleotides with yellow background were known to be protected upon 48S complex binding. c mini- or PTV-IIIab-based taRNAs with NT or g4 were transfected into N2a cells as plasmids, and mouse SYNGAP1 upregulation levels were compared. g4 is the optimized guide RNA from (a). n = 4 biological replicates. d mini taRNA with NT or human SYNGAP1-targeting gRNA (hSYNGAP1) was transfected to HEK293T cells. n = 4 biological replicates. e Expression and functional rescue of SYNGAP1 in iPSC-derived SYNGAP1^+/− neurons. The hSYNGAP1-targeting mini taRNA and NT control were in vitro transcribed and delivered by LNPs to iPSC-neurons. At 12-h post delivery, the levels of SYNGAP1 and phosphorylated ERK1/2 were assessed by western blots. Matched iPSC-neurons from heterozygous mutant (+/−, purple) or homozygous normal (+/+, brown) individuals treated with DPBS were used as reference level (dashed lines). n = 3 biological replicates. For western blots, α-tubulin and total ERK1/2 were used as loading controls. Representative blots were shown, and quantifications were normalized to non-targeting control (NT). All bar-graphs are shown as mean ± SEM with data points. Statistical analyses were performed using (a) one-way ANOVA with Sidak’s multiple comparisons test vs. NT, and g1 vs g4; c two-way ANOVA with Sidak’s multiple comparisons test between g4 vs. NT in each group. Unpaired two-tailed Student’s t tests were performed in (d, e). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. The P value and source data are provided as a Source Data file.