Dual-specificity RNA aptamers enable manipulation of target-specific O-GlcNAcylation and unveil functions of O-GlcNAc on β-catenin

SUMMARY

O-GlcNAc is a dynamic post-translational modification (PTM) that regulates protein functions. In studying the regulatory roles of O-GlcNAc, a major roadblock is the inability to change O-GlcNAcylation on a single protein at a time. Herein, we developed a dual RNA aptamer based approach that simultaneously targeted O-GlcNAc transferase (OGT) and β-catenin, the key transcription factor of the Wnt signaling pathway, to selectively increase O-GlcNAcylation of the latter without affecting other OGT substrates. Using the OGT/β-catenin dual-specificity aptamers, we found that O-GlcNAcylation of β-catenin stabilizes the protein by inhibiting its interaction with β-TrCP. O-GlcNAc also increases β-catenin’s interaction with EZH2, recruits EZH2 to promoters and dramatically alters the transcriptome. Further, by coupling riboswitches or an inducible expression system to aptamers, we enabled inducible regulation of protein-specific O-GlcNAcylation. Together, our findings demonstrate the efficacy and versatility of dual-specificity aptamers for regulating O-GlcNAcylation on individual proteins.

Graphical Abstract

In Brief

• Zhu and Hart report the development of dual-specificity RNA aptamers for altering O-GlcNAcylation of specific proteins; this work provides a versatile tool kit for studying an elusive post-translational modification.

INTRODUCTION

O-linked β-N-acetylglucosamine (O-GlcNAc) is a post-translational modification (PTM) attached to serine and threonine residues of more than 8,000 human nucleocytoplasmic proteins^1–3. O-GlcNAc is a highly dynamic PTM which is actively added and removed by two enzymes: O-GlcNAc transferase (OGT) and O-GlcNAcase (OGA), respectively^4–7. As a fast-cycling, nutrient-sensitive PTM, O-GlcNAc regulates many signaling events, interplays with phosphorylation, and regulates protein functions such as intermolecular interactions and degradation. Abnormal O-GlcNAcylation is associated with many human diseases, including cancer and diabetes⁸.

With current approaches, it is difficult to study the biological functions of O-GlcNAc on a specific protein. OGT and OGA are the sole O-GlcNAc modifying enzymes, any chemical inhibitors or genetic approaches targeting them change O-GlcNAcylation on thousands of proteins simultaneously. O-GlcNAc and phosphate frequently modify the same or proximal sites in a mutually exclusive way. This crosstalk makes studies that use site-directed mutagenesis to convert serine or threonine to other non-modifiable residues difficult to interpret. Moreover, there is no amino acid mimic of O-GlcNAcylation, as there is for phosphorylation. The lack of tools to modify O-GlcNAcylation of a single protein without affecting the many others in the cell has substantially impeded our understanding of the functions of this ubiquitous and essential modification.

To address this problem, Pratt and others have used expressed protein ligation to study O-GlcNAc modified proteins in vitro^9,10. In this method, a chemically synthesized, O-GlcNAc modified peptide is ligated to bacterial expressed, unmodified protein portions to generate a full-length protein with O-GlcNAc on certain sites. Proteins produced in this method are completely modified at defined sites but can only be used for in vitro assays. It also requires cysteine residues at the ligation sites. Woo and colleagues used nanobody-tagged OGT or OGA constructs to selectively modify O-GlcNAcylation of target proteins in cells^11,12. However, the tight and stable binding of these antibodies to the target proteins make sorting out the biological effects of altered O-GlcNAcylation from the effects of the antibody binding difficult.

Wnt signaling plays essential roles in the regulation of cell proliferation, differentiation, and self-renewal of stem cells. Its abnormalities occur in many types of cancer and inheritable disorders^13,14. β-catenin is the key transcription factor in the canonical Wnt signaling pathway. Its functions are regulated by PTMs. In resting cells, β-catenin maintains a low abundance due to the consecutive phosphorylation, ubiquitination, and degradation^15,16. Wnt signaling triggers a cascade of events that ultimately stabilizes β-catenin by inhibiting its ubiquitination, causing its nuclear accumulation and activation of downstream genes¹⁷. Besides phosphorylation and ubiquitination, the functions of β-catenin are regulated by PTMs including O-GlcNAcylation¹⁸.

Here, we developed an RNA-based method that induces protein-specific O-GlcNAcylation under endogenous levels of OGT and OGA, and studied how O-GlcNAc regulates β-catenin. Dual-specificity (DS) aptamers are modular designed RNA that connect two aptamer motifs with a linker domain. In cells, they induce proximity between OGT and a designated protein, and increase O-GlcNAcylation on the substrate. These RNA aptamers have short half-lives in cells and their binding can be controlled by riboswitches. Using DS aptamers, we induced O-GlcNAcylation on GFP-tagged and endogenous proteins and studied the regulatory roles of O-GlcNAc on β-catenin. O-GlcNAc regulates the interactions between β-catenin and a variety of proteins, including the E3 ubiquitin ligase β-TrCP, and the epigenetic modifiers EZH2, KAT2A, KAT5 and EP300. Particularly, O-GlcNAcylated β-catenin recruits EZH2 to promoters and shifts the transcriptome in a Wnt-dependent manner. Finally, by coupling riboswitches, or a Tet-On inducible expression system to DS aptamers, we deployed DS aptamers for inducible regulation of O-GlcNAcylation. Beyond O-GlcNAc, DS aptamers also provide us programable tools for the regulation of other PTMs and interactions in a cell.

RESULTS

A Noninhibiting RNA Aptamer Targeting ncOGT Was Generated From SELEX

Discovered in 1990, aptamers are short, single-stranded DNA or RNA molecules that bind targets with high affinity and specificity. They have been selected to bind a wide variety of targets from metal ions, small organic molecules, proteins to whole cells^19–22. Aptamers are generated from the in vitro selection named Systematic Evolution of Ligands by EXponential enrichment (SELEX, Figure 1A), which is an iterative binding, partitioning and amplification process.

Figure 1. A Noninhibiting RNA Aptamer Targeting ncOGT Was Generated From SELEX.

(A) Schematic of SELEX.

(B) and (C) Predicted secondary (B) and tertiary (C) structures of T1. Nucleotides in the two panels are color coded in the same way.

(D) Radiolabeled Dot-Blot Assay measuring the binding affinity of T1 to full-length ncOGT. Data are fitted to a Michaelis-Menten model (curve).

(E) and (F) Surface Plasmon Resonance (SPR) characterizing the binding affinity and kinetics of T1 to full-length ncOGT (E), and to the TPR domain (F). Sensorgrams represent OGT concentrations from 1 μM to 977 pM (E) or 1 μM to 488 pM (F), in 2× serial dilutions. Data are fitted to a 1:1 binding model.

(G) UDP-Glo assay of OGT activity in the presence of T1. T1 was included in the indicated reactions (orange) from 2 μM to 7.8 nM in 2× serial dilution. Data are normalized to the positive control reaction (red). N=3.

(H) RNA-IP and RT-qPCR quantification of OGT-bound T1 in cells. OGT was immunoprecipitated from T1-expressing cells, using buffers with Mg²⁺ (left panel) or without Mg²⁺ (right panel). Then OGT-bound T1 was quantified by RT-qPCR. Data are normalized to the IgG groups. RT – Reverse Transcriptase. N=3.

(I). Western Blot (WB) analysis of global O-GlcNAcylation in HEK293T cells expressing T1 or AP3. N=3.

(J) Pulse-chase experiments on the half-life of T1 in cells. Cells stably expressing T1 were treated with ActD for the indicated periods. Abundance of T1 was quantified by RT-qPCR (left panel) and normalized to that of 18S rRNA (right panel) and initial timepoint (T=0). Data are fitted to a one-phase decay model (curve). N=3.

One-way ANOVA test in (G), (H). Data are represented as mean ± SD. ns, p ≥ 0.05; ****, p < 0.0001.

To regulate O-GlcNAcylation with DS aptamers, it is a prerequisite that we have a noninhibiting RNA aptamer targeting the nucleocytoplasmic isoform of OGT (ncOGT). Therefore, SELEX was performed. To generate aptamers that bind the non-catalytic domain of OGT, we switched the targets of SELEX between the full-length ncOGT and its tetratricopeptide repeat (TPR) domain (STAR Methods)²³. After 11 rounds of SELEX, the sequence T1 was highly enriched in the final RNA library (Figures 1B, 1C).

We validated T1 as a noninhibiting aptamer targeting ncOGT in vitro. Binding between T1 and ncOGT was confirmed in Radiolabeled Dot-Blot Assays (Figure 1D), and further characterized in Surface Plasmon Resonance (SPR) (Figure 1E). In both experiments, the dissociation constant (K_d) was measured to be ~60 nM. SPR also shows that T1 recognizes the TPR domain of ncOGT (Figure 1F). To study if T1 inhibits the activity of the enzyme, we performed UDP-Glo OGT activity assays (Figure 1G). No inhibition was observed when T1 was included in the reactions from 7.8 nM to 2 μM, which was ~ 33 times higher than the K_d.

We further validated T1 in living cells. The binding between T1 and OGT in HEK293T cells was confirmed in RNA-IP experiments (Figure 1H). When expressed in cells, T1 was highly enriched on immunoprecipitated OGT protein. This binding requires Mg²⁺. In RNA-IP reactions containing no Mg²⁺, T1 was no longer enriched on OGT. The divalent Mg²⁺ ion facilitates RNA folding by shielding the negative charges on the RNA backbone^24,25, and Mg²⁺ was included in the selection buffer in SELEX. In cells, expression of T1 did not change the global O-GlcNAcylation pattern or the abundance of OGT (Figure 1I). Moreover, in pulse-chase experiments with actinomycin D (ActD) treatment, we found T1 to have a short half-life (16.18 min) in cells. In contrast, 18S rRNA was stable during the 8-hr treatment (Figure 1J).

In conclusion, T1 is an RNA aptamer that binds OGT without inhibiting its enzymatic activity both in vitro and in cells. It is a short-lived RNA in cellular context.

Dual-Specificity Aptamers Increase O-GlcNAcylation on GFP-Tagged Proteins

Next, we conjoined T1 with AP3, an RNA aptamer of GFP²⁶. In these DS aptamers, T1 and AP3 were connected by a linker region. We first designed a series of flexible linkers that were single-stranded, GC-neutral, and that did not interfere with the folding of T1 and AP3 (Figure 2A). Like the individual T1 aptamer, the DS aptamer has a short half-life (8.029 min) in cells (Figure 2B). The nomenclature of all aptamers in this research is illustrated in Table S1.

Figure 2. Dual-Specificity Aptamers Increase O-GlcNAcylation on GFP-Tagged Proteins.

(A) Schematic of DS aptamers with flexible linkers (NL1F, T1-linker-AP3). Five adenine residues are inserted upstream of AP3 to stabilize the transcribed U6 terminator.

(B) Pulse-chase experiments on the half-life of NL1F50 (T1–50nt-AP3). Experiments were performed and analyzed in the same way as in Figure 1J. N=3.

(C) and (D) IP-WB on cells expressing GFP-β-catenin and the indicated aptamers. N.A. – Non-aptamer. endo. – endogenous. N=5.

(E) Mass-shift assays on cells expressing GFP-β-catenin and the indicated aptamers. Two blots on β-catenin with short and long exposure time are shown. Intensities of the shifted bands are normalized to that of β-tubulin. N=3.

(F) and (G) IP-WB on cells expressing GFP-β-catenin and the indicated aptamers and treated with 50 μM Ac5S for 20 hr (F) or 2 μM TMG for 6 hr (G). N=3.

(H) and (I) Schematics of the DS aptamers with folded linkers. 3JB1F has a three-way RNA junction (H) and 4JC1RR has a four-way RNA junction (I) as linkers.

(J) Optimization of the folded linker in 3JB1F. IP-WB on cells expressing GFP-β-catenin and the indicated aptamers. N=3.

(K) and (L) IP-WB (K) or Co-IP (L) on HEK293T cells expressing GFP-ERα and the indicated aptamers. N=3.

T1 · AP3 – Individual T1 and AP3 aptamers (control). NL1F30/50 – DS aptamers (T1-linker-AP3) with flexible linkers of 30/50 nt. 3JB1F/4JC1RR – DS aptamers (T1-linker-AP3) with folded linkers. +2/+4/…/+12, serial additions (bp) into the folded linker. Quantitated WB data are normalized to the control groups (T1 · AP3). Aptamers and proteins were expressed from plasmids. One-way ANOVA test in (C), (J), (K), (L). Student’s t-test in other panels. Data are represented as mean ± SD. ns, p ≥ 0.05; *, p < 0.05; **, p < 0.01; ***, p < 0.001.

See also Figures S1 and S2.

Expression of these NL1F DS aptamers significantly increased O-GlcNAcylation of GFP-β-catenin in a specific manner. Protein-specific O-GlcNAcylation was quantified by immunoprecipitation (IP) and Western Blot (WB). In comparison to the individual T1 and AP3 aptamers (T1 · AP3), NL1F30, NL1F35 and NL1F50 up-regulated O-GlcNAc levels on GFP-β-catenin by 2- to 3-fold (Figures 2C, 2D, S2A to S2C). Efficacy of DS aptamers was further confirmed in mass-shift assays, where O-GlcNAc groups were enzymatically labeled with N-azidoacetylgalactosamine (GalNAz), then conjugated to polyethylene glycol (PEG) of 8.5 kDa by strain-promoted alkyne-azide cycloaddition (SPAAC), causing a mobility shift of the protein in WB²⁷. In mass-shift assays, NL1F30 and NL1F50 increased O-GlcNAcylation on GFP-β-catenin by 3.3-fold and 2.9-fold, respectively (Figures 2E, S2F). Increasing the expression of GFP-β-catenin renders NL1F35 ineffective, suggesting that the molar ratio of aptamer to substrate is important for the efficacy of DS aptamers (Figures S2B, S2C). Meanwhile, O-GlcNAc on endogenous β-catenin remained unchanged (Figures 2C, 2D, S2A), and DS aptamers did not affect global O-GlcNAcylation (Figure S2E). These results indicate good selectivity of the DS aptamers. The abundance of GFP-β-catenin was also increased, suggesting a stabilizing effect of O-GlcNAc on this protein (Figure S2E). DS aptamer increased O-GlcNAcylation on GFP-β-catenin in Wnt-stimulated cells (Figure S2G). On GFP-β-catenin, the changes in O-GlcNAcylation arose from its β-catenin portion, as O-GlcNAc on the GFP tag was undetectable (Figure S2D).

Ac₄-5SGlcNAc (Ac5S) and thiamet-G (TMG) are highly potent inhibitors of OGT and OGA, respectively^28,29. DS aptamers did not induce hyper-O-GlcNAcylation on GFP-β-catenin in Ac5S-treated cells but increased this modification in TMG-treated cells (Figures 2F and 2G). Thus, the activity of OGT is required for aptamer-induced O-GlcNAcylation, while the activity of OGA is not. These observations, together with the observation that the DS aptamers do not change the abundance of OGT and OGA (Figures 3C, 3D), rule out the possibility that aptamer-induced O-GlcNAcylation arise from the potential “off-target” effects of RNA aptamers.

Figure 3. O-GlcNAc Stabilizes GFP-β-catenin by Inhibiting Its Interaction With β-TrCP.

(A) Confocal images of Proximity Ligation Assay (PLA) on cells expressing GFP-β-catenin and the indicated aptamers. PLA was performed with antibodies targeting GFP and β-TrCP.

(B) Quantification of (A). PLA puncta are counted and normalized to the number of nuclei in each image. N=10.

(C) to (F) WB on HEK293T cells expressing GFP-β-catenin and the indicated plasmids. In (E), cells were treated with 50 μM Ac5S for 20 hr. (D) and (F) Quantifications of (C) and (E). Intensity of each band is normalized to that of β-tubulin. Np – non-phosphorylated. p – phosphorylated. N=3.

T1 · AP3 – Individual T1 and AP3 aptamers (control). NL1F30 – DS aptamers (T1-linker-AP3) with a flexible linker of 30 nt. Quantitated data are normalized to the control groups (T1 · AP3). Student’s t-test. Data are represented as mean ± SD. ns, p ≥ 0.05; *, p < 0.05; **, p < 0.01; ***, p < 0.001.

See also Figure S3.

In addition to the single-stranded flexible linkers, we designed DS aptamers with folded linkers. Three-way and four-way RNA junctions are naturally occurring RNA structures. In family C of three-way junctions, the P1 and P2 helices are coaxially stacked while the P3 helix “bends” towards P1 (Figure 2H)³⁰. These structural properties result in proximity between the two helices. We picked a structure (PDB 1MFQ, S-domain of 7SL RNA) from this family³¹ and grafted T1 and AP3 onto its helices. By swapping the locations of T1 and AP3, two DS aptamers (3JB1F and 3JB1R) were generated (Figures 2H, S2H). On 3JB1F and 3JB1R, T1 and AP3 were of identical distance but different orientations. To optimize the folded linkers, we introduced serial additions into the P3 helices of 3JB1F and 3JB1R (Figure S1A). Considering that in RNA double helices, addition of 1 bp introduces a rotation of 32.7° but only increases the length by 2.8 ų², these additions cause notable changes in the orientation of T1/AP3, but minor variations on the distance.

The 3JB1F variants significantly promoted O-GlcNAcylation on GFP-β-catenin (Figure 2J). The original 3JB1F increased this PTM by 3.5-fold, which was dramatically improved by serial additions on its P3 helix. Most of the variants increased O-GlcNAc on GFP-β-catenin by more than 10-fold. Among them, the variant with 6 bp addition (3JB1F+6; Figure S1A) increased this PTM by 15-fold. Notably, the impact of serial additions showed some “periodicity”, which peaked at 6 bp and decreased as additions grew longer. This observation agrees with the structural properties of RNA helices, where addition of 11 bp introduces a 360° rotation. Meanwhile, O-GlcNAcylation of endogenous β-catenin remained unchanged (Figure 2J). 3JB1R increased O-GlcNAcylation of GFP-β-catenin by 2-fold, however, serial additions had little effect (Figure S2I). This observation indicates that on 3JB1R, the orientation of T1 is suboptimal, so that OGT is facing away from its target protein when bound to the complex.

DS aptamers increase O-GlcNAc levels on other GFP-tagged proteins. Here we designed another folded linker. In family π of four-way RNA junctions, the helices H3 and H4 are coaxially stacked while H1 and H2 “swings”, and H2 is attracted to H3 (Figure 2I)³³. We picked a structure (PDB 1U9S, specificity domain of A-type ribonuclease P) from this family³⁴, and grafted two copies of T1 and AP3 onto its helices (Figure S1B). This 4JC1RR, as well as 3JB1F+6, increased O-GlcNAc on GFP-tagged Estrogen Receptor α (ERα), Src and AMP-activated protein kinase α2 (AMPKα2) (Figures 2K, S2J, S2L). Co-immunoprecipitation (Co-IP) revealed that 4JC1RR promoted the interactions between OGT and these proteins (Figures 2L, S2K), supporting our hypothesis that DS aptamers increase O-GlcNAcylation by inducing proximity between OGT and its substrates.

In conclusion, DS aptamers with flexible or folded linkers promote interactions between OGT and GFP-tagged proteins and selectively increase O-GlcNAcylation on these substrates.

DS Aptamers Reveal Unmapped O-GlcNAc Sites on β-catenin

Our study reveals unmapped O-GlcNAc sites. Four O-GlcNAc sites have been mapped (Ser23, Thr40, Thr41 and Thr112) on β-catenin¹⁸. When all these sites were mutated to alanine, O-GlcNAcylation on this GFP-β-catenin^4A mutant was still induced by NL1F30 or 3JB1F+6 (Figures S3A, S3E), though to a lesser extent than on the wild-type protein (Figures 2C, 2J). This induction required the activity of OGT (Figures S3C, S3G). Similar to on the wild-type protein, O-GlcNAc stabilized the 4A mutant (Figures S3B, S3F). Interestingly, though NL1F30 did not induce higher O-GlcNAcylation than 3JB1F+6, it was more effective on stabilizing the protein. The stabilization effects required the activity of OGT (Figures S3D, S3H). These observations suggest that NL1F30 and 3JB1F+6 have different site-selectivity when inducing O-GlcNAcylation on GFP-β-catenin, and that O-GlcNAc on the unidentified sites regulates the stability of β-catenin.

In control experiments without aptamers, O-GlcNAc signal on the 4A mutant was reduced by a bacterial homolog of OGA (CpOGA)³⁵, or by competing the O-GlcNAc antibody with 1 M GlcNAc during WB (Figures S3I, S3J). These results confirm the WB bands to be specific to O-GlcNAc, and that O-GlcNAc on the 4A mutant was not artifacts of aptamer expression.

O-GlcNAc Stabilizes GFP-β-catenin by Inhibiting Its Interaction With β-TrCP

As the key transcription factor of the Wnt signaling pathway, β-catenin is activated by Wnt. Without Wnt, newly synthesized β-catenin is recruited to a destruction complex comprised of Axin, APC, GSK3β and CK1. In the complex, β-catenin is consecutively phosphorylated by CK1 and GSK3β on its Ser45, Thr41, Ser37 and Ser33 residues, ubiquitinated by the SCF^β-TrCP complex and degraded by proteasomes^15,16,36. Binding of Wnt to the membrane receptor Frizzled triggers a cascade of signaling events, which ultimately blocks the ubiquitination of phosphorylated β-catenin in the destruction complex¹⁷. Stabilized β-catenin saturates the complex, accumulates in the cytoplasm, enters the nucleus, and activates transcription³⁷.

β-TrCP recruits β-catenin that is phosphorylated on Ser33 and Ser37 to the SCF^β-TrCP E3 ubiquitin ligase, causing its ubiquitination and degradation^15,38. In Proximity Ligation Assays (PLA), we found that DS aptamer inhibited the interaction between GFP-β-catenin and β-TrCP (Figures 3A, 3B). We reasoned that impairing this interaction would inhibit the degradation of phosphorylated GFP-β-catenin, therefore saturate the destruction complex, stabilize the newly synthesized, unphosphorylated GFP-β-catenin, and increase the total abundance of this protein (Figures S2E, 3C, 3D, S3B, S3F). DS aptamer increased the non-phosphorylated moiety of GFP-β-catenin regarding Ser33, Ser37, Thr41 and Ser45 (Figures 3C, 3D), agreeing with our hypothesis. Meanwhile, abundance of phosphorylated GFP-β-catenin was unchanged regarding these sites (Figures 3C, 3D), suggesting that O-GlcNAc inhibit the interaction between GFP-β-catenin and β-TrCP not by competing phosphorylation, but likely by steric hinderance since it has a large Stokes radius³⁹. Additionally, DS aptamer increased phosphorylation on Ser675 (Figures 3C, 3D) that enhances the transcriptional activity of β-catenin⁴⁰. These site-specific phosphorylations on endogenous β-catenin were not changed (Figures 3C, S3K). Changes of site-specific phosphorylation required the activity of OGT (Figures 3E, 3F, S3L), supporting that they were caused by aptamer-induced O-GlcNAcylation.

In conclusion, O-GlcNAc on GFP-β-catenin inhibits its recognition by β-TrCP and stabilizes this protein.

DS Aptamers Increase O-GlcNAcylation on Endogenous β-catenin

Beyond epitope-tagged proteins, we asked if DS aptamers could regulate O-GlcNAc on endogenous proteins. An RNA aptamer (bc339) targeting β-catenin was generated from SELEX (Figure 4A). We characterized its binding properties by SPR (Figure 4B). The dissociation constant (K_d) between bc339 and β-catenin is 16.00 nM. We connected this aptamer (bc339) to T1 with a series of flexible linkers, thus the NL8F DS aptamers were designed (Figure 4C).

Figure 4. DS Aptamers Increase O-GlcNAcylation on Endogenous β-catenin.

(A) Predicted secondary (left panel) and tertiary (right panel) structures of bc339. Nucleotides in the two panels are color coded in the same way.

(B) SPR characterizating the binding affinity and kinetics of bc339. Sensorgrams represent β-catenin concentrations from 1 μM to 122 pM in 2× serial dilutions. Data are fitted to a 1:1 binding model.

(C) Schematic of DS aptamers with flexible linkers (NL8F, T1-linker-bc339). Five adenine residues are inserted upstream of bc339 to stabilize the transcribed U6 terminator.

(D) IP-WB on HEK293T cells expressing the indicated aptamers. N=3.

(E) and (F) IP-WB on HEK293T cells expressing the indicated aptamers and treated with 50 μM Ac5S for 20 hr (E) or Wnt3A-conditioned medium for 4 hr (F). Irrelevant lanes are cropped. N=3.

(G) Schematic of DS aptamers (3JB8F, T1-linker-bc339) with folded linkers.

(H) Predicted tertiary structure of the DS aptamer 3JB8F+12. Nucleotides are color coded in the same way as in (G). Complete sequence and secondary structure prediction are shown in Figure S1C.

(I) to (L) IP-WB on HEK293T cells expressing the indicated aptamers, under −Wnt (I, J) and +Wnt (K, L) conditions. (J) and (L) Quantification of (I) and (K). N=4.

(M) and (N) Confocal images of PLA on HEK293T cells expressing the indicated aptamers. (N) Quantification of (M). Total PLA puncta is counted and normalized to the number of nuclei in each image. N=8.

(O) Subcellular localization of 3JB8F+12. Nucleus and cytoplasm of cells stably expressing 3JB8F+12 was isolated, and RNA in each fraction was quantified by RT-qPCR. RNA abundances are normalized to that of 3JB8F+12 in whole cell lysate (purple bar in upper panel). NEAT1 is a nuclear RNA control. GAPDH is a cytoplasmic RNA control. Cyt. – Cytoplasm. Nuc. – Nucleus. W.C. – Whole Cell. N=4.

T1 · bc339 – Individual T1 and bc339 aptamers (control). NL8F50/70/100 – DS aptamers (T1-linker-bc339) with flexible linkers of 50/70/100 nt. 3JB8F – DS aptamers (T1-linker-bc339) with folded linkers. +4/+12, serial additions (bp) into the folded linker. Quantitated data are normalized to the control groups (T1 · bc339). Aptamers were expressed from plasmids. Student’s t-test in (E), (F), (N). One-way ANOVA test in other panels. Data are represented as mean ± SD. ns, p ≥ 0.05; *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001.

See also Figure S1 and S4.

NL8F70 and NL8F100 increased O-GlcNAcylation of endogenous β-catenin by 2.3- and 1.9-fold, respectively (Figure 4D). NL8F100 moderately stabilized β-catenin (Figure S4A). Meanwhile, global O-GlcNAcylation, as well as the abundance of OGT and OGA, remained unchanged (Figure S4A). These results reveal good selectivity of the DS aptamers. Inhibiting OGT prevented aptamer-induced O-GlcNAcylation (Figure 4E). Efficacy of NL8F70 was confirmed in mass-shift assay (Figure S4B). However, when Wnt signaling was activated, efficacy of NL8F70 became less clear (Figures 4F, 4K, 4L).

We also constructed DS aptamers with folded linkers. By replacing the GFP aptamer (AP3) in the 3JB1F constructs with bc339, we designed the 3JB8F constructs (Figures 4G, 4H, S1C). Among them, 3JB8F+12 increased O-GlcNAcylation of β-catenin by 4-fold in the absence of Wnt (Figures 4I, 4J), and by 2.5-fold in the presence of Wnt (Figures 4K, 4L). It also stabilized β-catenin (Figures S4C, S4D). Global O-GlcNAcylation was not changed by 3JB8F+12 under either condition (Figures S4E, S4F). Like other DS aptamers, 3JB8F+12 requires the activity of OGT to induce O-GlcNAcylation under both −Wnt and +Wnt conditions (Figures S4G, S4H).

In PLA and Co-IP, we found that 3JB8F+12 substantially promoted the interaction between OGT and β-catenin (Figures 4M, 4N, S4I). These results support our hypothesis that DS aptamers regulate O-GlcNAcylation by targeting a substrate protein to OGT. We studied the subcellular localization of DS aptamers by nucleus/cytoplasm fractionation and RT-qPCR (Figure 4O). Compared to the RNA species enriched in nucleus (NEAT1) and cytoplasm (GAPDH), 3JB8F+12 localizes in both. Therefore, DS aptamers could regulate O-GlcNAcylation on both nuclear and cytoplasmic proteins.

We constructed another DS aptamer (3JB8R+12) by swapping T1 and bc339 on 3JB8F+12 (Figures S1D, S4J). Though it consists of identical modules, 3JB8R+12 was unable to regulate O-GlcNAcylation of β-catenin under either status of Wnt signaling (Figures S4K and S4L). This observation further supports our hypothesis that not only the distance, but also the orientations of the individual aptamers on the folded linker are critical to the efficacy of DS aptamers.

In conclusion, the T1-linker-bc339 DS aptamers specifically increase O-GlcNAcylation on endogenous β-catenin. DS aptamers could induce O-GlcNAcylation in both nucleus and cytoplasm.

O-GlcNAc Regulates β-catenin’s Interactions With Epigenetic Modifiers and Increases Its Transcriptional Activity

We then studied how O-GlcNAc regulates the functions of β-catenin using DS aptamers. As a transcription factor that does not have a DNA binding domain, the functions of β-catenin heavily rely on its interactions with other proteins, including EZH2, KAT2A and EP300^41–43.

Enhancer of zeste homolog 2 (EZH2) is the catalytic subunit of the Polycomb Repressive Complex 2 (PRC2), which plays important roles in gene repression and during development^44,45. It catalyzes the tri-methylation of Lys27 on Histone H3 (H3K27me3), which marks silent promoters⁴⁶. In PLA and Co-IP, we found that the interaction between β-catenin and EZH2 was increased by 3JB8F+12, regardless of the Wnt signaling status (Figures 5A, 5B, 5I, 5J, S4M, S4N). This increase was prevented by inhibition of OGT (Figures 5C, 5D, 5K, 5L). Furthermore, 3JB8R+12, the DS aptamer that does not change O-GlcNAc levels on β-catenin, had no effect on this interaction (Figures 5E, 5F, 5M, 5N). Proximity between β-catenin and H3K27me3 was also increased by 3JB8F+12 (Figures 5G, 5H, 5O, 5P). We conclude that O-GlcNAcylation of β-catenin promotes its interaction with EZH2.

Figure 5. O-GlcNAc Promotes β-catenin’s Interactions With EZH2.

(A) to (H) Confocal images of PLA on HEK293T cells expressing the control aptamers (T1 · bc339) or a DS aptamer (3JB8F+12 or 3JB8R+12), under −Wnt or +Wnt conditions. PLA was performed with antibodies targeting β-catenin and EZH2 or H3K27me3. In (C) and (D), cells were treated with 50 μM Ac5S for 20 hr.

(I) to (P) Quantification of (A) to (H). PLA puncta in nucleus are counted and normalized to the number of nuclei in each image. Student’s t-test, N=10 (I to N) or 8 (O, P). Data are represented as mean ± SD. ns, p ≥ 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001.

T1 · bc339 – Individual T1 and bc339 aptamers (control). 3JB8F+12/R+12 – DS aptamers (T1-linker-bc339) with folded linkers.

See also Figures S5 and S6.

KAT2A (GCN5) is a histone acetyltransferase that acetylates the Lys9 of histone H3 (H3K9ac). H3K9ac correlates with actively expressed genes⁴⁷. In PLA, we found that 3JB8F+12 increased the interaction between β-catenin and KAT2A in spite of the Wnt signaling status (Figures S5A, S5B, S5I, S5J). Inhibition of OGT prevented this change (Figures S5C, S5D, S5K, S5L). 3JB8R+12 had no effect on this interaction (Figures S5E, S5F, S5M, S5N). Moreover, 3JB8F+12 also increased the proximity between β-catenin and H3K9ac (Figures S5G, S5H, S5O, S5P). In conclusion, O-GlcNAc on β-catenin enhances its interaction with KAT2A.

3JB8F+12 regulates the interactions between β-catenin and other histone modifiers. It significantly inhibited the interaction between β-catenin and EP300, an acetyltransferase and transcription coactivator (Figures S6A, S6B, S6I, S6J)⁴⁸. It also impaired the interaction between β-catenin and KAT5 (TIP60), a lysine acetyltransferase that is involved in the regulation of transcription, in DNA repair and in apoptosis (Figures S6C, S6D, S6K, S6L)⁴⁹. Additionally, this DS aptamer inhibits the interaction between endogenous β-catenin and β-TrCP (Figures S6E, S6F, S6M, S6N), agreeing with our observation on GFP-β-catenin and β-TrCP (Figures 3A, 3B). Negative control experiments revealed good specificity of PLA experiments (Figures S6G, S6H, S6O, S6P).

Using aptamers, we studied how O-GlcNAc regulates β-catenin’s transcriptional activity in Wnt signaling. In TOPFlash assays, 3JB8F+12 elevated the activity of β-catenin in response to Wnt (Figure S7A), which required the activity of OGT (Figure S7B). These results agree with the observations when the O-GlcNAc modifying enzymes was inhibited with genetic or chemical approaches. Knockdown of OGT considerably impaired the activity of β-catenin, which was restored by rescuing this enzyme (Figure S7C). Reducing global O-GlcNAcylation with the OGT inhibitor (Ac5S) impaired the activity of β-catenin, but elevating global O-GlcNAcylation with the OGA inhibitor (TMG) further activated it (Figure S7D). In a control experiment, neither single or double individual aptamers (T1, bc339, T1 · bc339) affected the activity of β-catenin (Figure S7E). In conclusion, O-GlcNAc on β-catenin promotes its activity in response to Wnt. However, the luciferase reporter in TOPFlash assay is encoded on a plasmid and driven by an artificial promoter that contains seven TCF/LEF binding sites. Its expression is highly sensitive to β-catenin but has a different epigenetic context comparing to chromosomal genes. Thus, this assay measures the activity of β-catenin at high sensitivity, but the native regulations on chromosomal genes are not represented.

O-GlcNAc on β-catenin Recruits EZH2 to Promoters and Shifts the Transcriptome

Next, we wondered if the enhanced interaction with β-catenin regulates the chromatin binding of EZH2. We performed Cleavage Under Targets and Release Using Nuclease (CUT&RUN) on HEK293T cells expressing the individual aptamers (T1 · bc339) or the DS aptamer (3JB8F+12)^50,51. CUT&RUN sequencing revealed that 3JB8F+12 increased the binding of EZH2 on 923 and 3473 sites in the absence and presence of Wnt, respectively (Figures 6A to 6C; Table S4). The differential binding sites predominantly localized on the promoter regions of the genome (Figures S7F, S7G). All differential binding events of EZH2 were prevented by knocking-down β-catenin, or by inhibiting OGT (Figures S7H, S7I, S7K), indicating they required increased O-GlcNAcylation and β-catenin. In addition, β-catenin is critical to the chromatin localization of EZH2, as its knockdown drastically impaired the promoter binding of EZH2 without affecting its abundance (Figures S7J, S7K). Given that O-GlcNAc on β-catenin enhances its interaction with EZH2 (Figures 5, S4M, S4N), we conclude that the strengthened interaction recruits EZH2 to promoters.

Figure 6. O-GlcNAc on β-catenin Recruits EZH2 to Promoters and Shifts the Transcriptome.

(A) and (B) Heatmap of the differential binding sites of EZH2. CUT&RUN sequencing was performed to HEK293T cells expressing the indicated aptamers and exposed to regular (−Wnt) or Wnt3A-conditioned (+Wnt) medium.

(C) The binding peaks of EZH2 on three promoters under indicated conditions.

(D) to (I) Volcano plot of RNA-seq data. HEK293T cells expressing the control (T1 · bc339) or DS (3JB8F+12) aptamers were treated with regular (−Wnt) or Wnt3A-conditioned (+Wnt) medium. In (F) and (G), EZH2 was knocked-down in cells with a shRNA. In (H) and (I), cells were treated with Ac5S.

T1 · bc339 – Individual T1 and bc339 aptamers (control). 3JB8F+12 – DS aptamer (T1-linker-bc339) with a folded linker.

See also Figure S7; Tables S4, S5.

We further asked if increased O-GlcNAcylation of β-catenin impacts the transcriptome. RNA sequencing (RNA-Seq) was performed on HEK293T cells expressing either the individual (T1 · bc339) or the DS (3JB8F+12) aptamers, in the absence and presence of Wnt. Surprisingly, the DS aptamer altered the transcriptome in opposite ways depending on the status of Wnt signaling. In the absence of Wnt, 3JB8F+12 promoted the expression of 83 genes. In contrast, when Wnt signaling was activated, 3JB8F+12 repressed the expression of 85 genes (Figures 6D, 6E; Table S5). These differential expressions required both EZH2 and elevated O-GlcNAc, as they were prevented by knocking-down EZH2 or inhibiting OGT (Figures 6F to 6I, S7K). In addition, when EZH2 was knocked-down and Wnt was present, O-GlcNAcylation of β-catenin slightly enhanced the transcription of eight other genes (Figures 6G, 6I). This observation suggests that O-GlcNAc on β-catenin also has mild, EZH2-independent activating effects on transcription, which could be overwritten by its EZH2-dependent repressive effects. Considering that EZH2 is associated with PRC2 that is a transcription repressor, the increased promoter binding of EZH2 may function in the repression of genes caused by O-GlcNAcylation of β-catenin when Wnt signaling was activated. However, the mechanism underlying the activation of genes without Wnt awaits further investigation. As EZH2 was reported to activate the androgen receptor gene independently of PRC2 and its methyltransferase activity⁵², it is possible that EZH2 activated the genes we observed in the same way. It is also possible that O-GlcNAc on β-catenin recruited some transcription activators to the chromatin when Wnt was absent, which we have not identified in this research.

In conclusion, O-GlcNAc on β-catenin strengthens its interaction with EZH2 and recruits EZH2 to the promoters. Globally, it regulates the transcriptome in two ways: it activates gene expression in the absence of Wnt but represses gene expression when Wnt signaling is activated. Our study reveals a synergistic mechanism that β-catenin regulate gene expression by recruiting histone modifiers.

Inducible Regulation of O-GlcNAcylation by Coupling DS Aptamers to Riboswitches or a Tet-On System

DS aptamers allow us to regulate O-GlcNAcylation on a single protein and to study protein-specific effects of this PTM. However, there is a concern about this method: will the normal functions of OGT and substrate proteins be interfered, if DS aptamers remain bound to them? This concern is partly solved by the fact that the individual and dual-specificity aptamers are short-lived (Figures 1J, 2B). Nevertheless, an inducible mechanism that forces the aptamers to dissociate from their targets would be ideal.

We designed inducible DS aptamers by incorporating riboswitch elements. Riboswitches are naturally derived or artificially designed RNA that change their conformations upon binding of ligands^53,54. They usually contain three modules: a sensor that is a ligand-binding aptamer; an actuator that carries out downstream effects; and a transmitter sequence that couples the two⁵⁵. In our design, two sets of riboswitches were incorporated into the DS aptamer NL1F50. In the newly generated inducible DS aptamer LRS1F50C (Figures 7A, 7B), the fluorogenic “Mango” aptamers serve as sensors^56,57. TO1-biotin (TO1), the ligand of Mango, is a non-toxic, membrane permeable molecule that binds Mango with low nanomolar affinity⁵⁸. The T1 and AP3 aptamers are the actuators, since it is their binding that the riboswitches aim to regulate. The transmitter elements were designed so that they bind the core sequences of Mango in the absence of TO1, but switch to bind T1 and AP3 in the presence of TO1 (STAR Methods). As a result, addition of TO1 triggers the folding of Mango, which disrupts the folding of T1 and AP3 and “turns off” the function of the DS aptamer (Figures 7A, 7B). In cells, LRS1F50C increased O-GlcNAcylation on GFP-β-catenin by 3-fold without affecting that of endogenous β-catenin, and it was inactivated by TO1 (Figures 7C to 7E).

Figure 7. Inducible Regulation of O-GlcNAcylation by Coupling DS Aptamers to Riboswitches or a Tet-On System.

(A) and (B) Schematics of LRS1F50C and its conformation transitions. The transmitter elements are boxed in red, their alternative binding sequences are boxed in blue and purple. (A) The “ON” state of LRS1F50C. (B) The “OFF” state of LRS1F50C.

(C) to (E) IP-WB on cells expressing GFP-β-catenin and the indicated aptamers, and treated with 750 μM TO1 (+TO1) or equal volume of DMF (−TO1). (D) and (E) Quantification of (C). N=3.

(F) Schematic of controlling 3JB8F+12 with a Tet-On system.

(G) and (H) IP-WB on cells transfected with the indicated plasmids, and treated with Dox (+Dox) or equal volume of DMSO (−Dox). (H) Quantification of (G). N=3.

Two-way ANOVA test. Data are represented as mean ± SD. ns, p ≥ 0.05; *, p < 0.05; ***, p < 0.001; ****, p < 0.0001.

We also used a Tet-On inducible expression system to control the function of DS aptamers. In this system, 3JB8F+12 is constitutively expressed, driven by a wild-type U6 promoter (U6WT). An antidote RNA of T1 (atdT1) is expressed under an inducible promoter (U6TO), which combines a truncated U6 promoter with two Tet operator elements. Doxycycline (Dox) activates the expression of atdT1, which disrupts the folding of 3JB8F+12 (Figure 7F). In cells, this system induced O-GlcNAcylation on endogenous β-catenin in the absence of Dox, and addition of Dox rendered it inactive (Figures 7G, 7H). Since aptamers are short-lived in cells (Figures 1J, 2B), inducible expression could allow the dynamic control of DS aptamers at high time-resolution.

Taken together, using riboswitches or the Tet-On system, we showed the proof-of-principle that DS aptamers can be used for inducible regulation of protein-specific O-GlcNAcylation. Further optimization on the riboswitches and the Tet-On system are necessary to improve the specificity and dynamic ranges of these tools.

DISCUSSION

We developed DS aptamers that induce protein-specific O-GlcNAcylation, without changing the activity of OGT or OGA, or tagging these enzymes that interferes with their normal functions. Using these tools, we found that O-GlcNAc regulates β-catenin by tuning its interactions with other proteins like β-TrCP and EZH2. Riboswitches or inducible expression systems can be used for the dynamic control of DS aptamers. The modular feature endows DS aptamers broad potential applications, such as regulating other PTMs, facilitating the formation of protein complexes, and even inducing proximity between proteins and/or organelles.

Unmapped O-GlcNAc Sites on β-catenin

Our research reveals unmapped O-GlcNAc sites on β-catenin (Figures S3A to S3J). We also find that O-GlcNAc on β-catenin regulates its interactions with KAT2A and EP300 (Figures S5, S6). Considering that β-catenin binds EP300 and KAT5 via its C-terminus^41,59, and that O-GlcNAc interplays with phosphate on Ser675 (Figure 3C), some of the unmapped sites probably reside in this region. O-GlcNAc stabilizes the GFP-β-catenin^4A mutant (Figures S3B, S3F). Since O-GlcNAc stabilizes β-catenin by inhibiting its interaction with β-TrCP (this study), which recognizes phosphorylation on Ser33 and Ser37¹⁵, some of the unmapped sites could be near this region. Mapping these sites will be important for investigating the site-specific effects of O-GlcNAc in the future.

Critical Roles of the Linker in DS Aptamers

The linker domain, which critically affects the efficacy of DS aptamers, can adopt either flexible or folded structures. Our optimization of the folded linkers reveals that in a DS aptamer, both the distance between the two individual aptamers and their orientations are important factors. Folded linkers tend to increase the efficacy of DS aptamers (Figures 2C, 2D, 2J; 4D, 4F, 4I to 4L), probably because 1) they induce higher degrees of proximity than flexible linkers; and 2) the folding of these linkers is highly spontaneous (ΔG ≪ 0), which facilitates the folding of the aptamer motifs. Meanwhile, the rigidity of folded linkers seems to reduce the degrees of freedom of the bound proteins, so that some variations of these DS aptamers do not increase O-GlcNAcylation. Therefore, when using these DS aptamers to target a new protein, an optimization that covers 0 – 11 bp variations of the RNA helix (0 – 360° rotation) will be needed. Alternative choices are folded linkers with more flexibility, like the 4-way RNA junction we used in this study (Figure 2I). In this linker, the H1 and H2 helices “swings”, which provides the bound proteins higher degrees of freedom.

The rigidity of folded linkers could provide better site-selectivity when inducing O-GlcNAcylation. Though NL1F30 is no more effective than 3JB1F+6 on inducing O-GlcNAcylation on the GFP-β-catenin mutant (Figures S3A, S3E), it is much more effective in stabilizing the protein (Figures S3B, S3F), indicating that the two DS aptamers have different site-selectivity. On 3JB1F+6, the rigid RNA helix restricts the bound proteins so that only certain O-GlcNAc sites on β-catenin are accessible to OGT, thus O-GlcNAcylation is preferably induced on these residues. Therefore, the substrate protein could be “rotated” gradually by introducing serial additions or deletions to the RNA helix where it resides, and O-GlcNAc on different sites can be induced.

Limitations of the Study

We used IP-WB and mass-shift assay to detect O-GlcNAc on β-catenin. Antibodies of β-catenin were used to precipitate the protein in IP, or to detect its shifted bands in WB. However, most available antibodies recognize β-catenin on its N- or C- terminus, where O-GlcNAc sites are identified or indicated in this study. We have found that O-GlcNAc and PEG at or near the epitope inhibits antibody recognition, thus the aptamer-induced O-GlcNAcylation in this study may be underestimated.

We show proof-of-principle of using riboswitches and Tet-On system to control DS aptamers, further optimization is needed on their designs. Also, the T1 and bc339 aptamers can be optimized for higher affinity and shorter lengths. Our data suggests unidentified O-GlcNAc sites on β-catenin but we did not map them, thus the site-selectivity of DS aptamers are speculative. In future studies, site-mapping will be needed to confirm the physiological relevance of the aptamer-induced O-GlcNAcylation in this study. The mechanism of EZH2-dependent transcription activation and repression also awaits investigation.

STAR METHODS

RESOURSE AVAILABILITY

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Dr. Gerald W. Hart (gwhart@jhmi.edu).

Materials Availability

Plasmids generated in this study have been deposited to Addgene. Cell lines generated in this study are available from the lead contact with a complete Materials Transfer Agreement.

Data and Code Availability

• CUT&RUN-seq and RNA-seq data have been deposited at GEO and are publicly available as of the date of publication. Accession numbers are listed in the key resources table. All other data have been deposited at Figshare and are publicly available as of the date of publication. The DOI is listed in the key resources table.

• This paper does not report original code.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Mouse monoclonal anti-O-GlcNAc IgM (strain CTD110.6)	Comer et al.81 Produced in the lab	N/A
Rabbit polyclonal anti-OGT (AL28)	Iyer et al.82 Produced in the lab	N/A
Rabbit polyclonal anti-OGT (AL24)	Kreppel et al.80 Produced in the lab	N/A
Chicken polyclonal anti-OGA IgY (345)	Butkinaree et al.83 Produced in the lab	N/A
Mouse monoclonal anti-β-Catenin (clone 14)	BD Biosciences	Cat#: 610153; RRID: AB_397555
Rabbit polyclonal anti-β-Catenin	Cell Signaling Technology	Cat#: 9562; RRID: AB_331149
Rabbit polyclonal anti-β-Catenin	ProteinTech	Cat#: 17565-1-AP; RRID: AB_2088102
Rabbit monoclonal anti-Non-phospho (Active) β-Catenin (Ser33/37/Thr41) (D13A1)	Cell Signaling Technology	Cat#: 8814; RRID: AB_11127203
Rabbit monoclonal anti-Non-phospho (Active) β-Catenin (Ser45) (D2U8Y)	Cell Signaling Technology	Cat#: 19807; RRID: AB_2650576
Recombinant Rabbit polyclonal anti-phospho-β-Catenin (Ser33/37) (4HCLC)	Invitrogen	Cat#: 711404; RRID: AB_2633009
Rabbit polyclonal anti-phospho-β-Catenin (Ser45)	Cell Signaling Technology	Cat#: 9564; RRID: AB_331150
Rabbit monoclonal anti-phospho-β-Catenin (Ser552) (D8E11)	Cell Signaling Technology	Cat#: 5651; RRID: AB_10831053
Rabbit monoclonal anti-phospho-β-Catenin (Ser675) (D2F1)	Cell Signaling Technology	Cat#: 4176, RRID: AB_1903923
Recombinant Rabbit monoclonal anti-GFP	Invitrogen	Cat#: G10362; RRID: AB_2536526
Mouse monoclonal anti-GFP (3E6)	Invitrogen	Cat#: A-11120; RRID: AB_221568
GFP-Trap Magnetic Agarose	ProteinTech	Cat#: gtma-10; RRID: AB_2631358
Rabbit polyclonal anti-β-TrCP	Invitrogen	Cat#: PA5-106884; RRID: AB_2854548
Rabbit monoclonal anti-EZH2 (D2C9)	Cell Signaling Technology	Cat#: 5246; RRID: AB_10694683
Mouse monoclonal anti-EP300 (clone NM11)	Active Motif	Cat#: 61401; RRID: AB_2716754
Rabbit monoclonal anti-KAT2A/GCN5L2 (C26A10)	Cell Signaling Technology	Cat#: 3305; RRID: AB_2128281
Rabbit polyclonal anti-KAT5	ProteinTech	Cat#: 10827-1-AP; RRID: AB_2128431
Rabbit monoclonal anti-H3K27me3 (C36B11)	Cell Signaling Technology	Cat#: 9733; RRID: AB_2616029
Rabbit monoclonal anti-H3K9ac (C5B11)	Cell Signaling Technology	Cat#: 9649; RRID: AB_823528
Rabbit polyclonal anti-β-Tubulin	Cell Signaling Technology	Cat#: 2146; RRID: AB_2210545
Normal Rabbit IgG	Millipore	Cat#: 12-370; RRID: AB_145841
Normal Rabbit IgG	Cell Signaling Technology	Cat#: 2729; RRID: AB_1031062
Mouse IgG1 Isotype Control (MOPC-21)	Invitrogen	Cat#: MA1-10406; RRID: AB_2536774
Goad anti-Mouse IgM (μ-chain specific) polyclonal secondary antibody, HRP-conjugated	Sigma-Aldrich	Cat#: A8786; RRID: AB_258413
Goat anti-Rabbit IgG (Heavy Chain), Superclonal recombinant secondary antibody, HRP-conjugated	Invitrogen	Cat#: A27036; RRID: AB_2536099
Goat anti-Mouse IgG (H+L), Superclonal recombinant secondary antibody, HRP-conjugated	Invitrogen	Cat#: A28177; RRID: AB_2536163
Bacterial and virus strains
Stellar Competent Cells: E. coli HST08 strain	TaKaRa	Cat#: 636766
NiCo21 (DE3) Competent E. coli	New England BioLabs	Cat#: C2529H
Biological samples
Chemicals, peptides, and recombinant proteins
TransIT-X2	Mirus Bio	Cat#: MIR 6000
Dulbecco’s Modified Eagle Medium (DMEM)	Gibco	Cat#: 10567014
Fetal bovine serum (FBS)	Gibco	Cat#: 16140071
Penicillin-Streptomycin	Gibco	Cat#: 15140163
Hygromycin B	Gibco	Cat#: 10687010
Puromycin dihydrochloride	Sigma-Aldrich	Cat#: P8833
TrypLE Express Enzyme	Gibco	Cat#: 12604013
2-Propanol	Sigma-Aldrich	Cat#: I9516-500mL; CAS: 67-63-0
Q5 High-Fidelity DNA Polymerase	New England BioLabs	Cat#: M0491
SuperScript IV Reverse Transcriptase	Invitrogen	Cat#: 18090010
TOPO TA Cloning Kit for Sequencing	Invitrogen	Cat#: 450030
Amine Coupling Kit	GE Healthcare/Cytiva	Cat#: BR100050
Antarctic Phosphatase	New England BioLabs	Cat#: M0289
[γ−32P]-ATP	American Radiolabeled Chemicals	Cat#: ARP 0102F-250 μCi
T4 Polynucleotide Kinase	New England Biolabs	Cat#: M0201
Acid-Phenol: Chloroform, pH 4.5	Invitrogen	Cat#: AM9720
Phenol: Chloroform, pH 8.05	Invitrogen	Cat#: 15593031
Yeast tRNA	Invitrogen	Cat#: AM7119
Bovine Serum Albumin	New England BioLabs	Cat#: B9001
IPTG	Invitrogen	Cat#: 15529019
Protease Inhibitor Cocktail	Roche	Cat#: 11873580001
Lysozyme	Thermo Scientific	Cat#: 89833
PreScission Protease	GenScript	Cat#: Z02799
PMSF	Roche	Cat#: 10837091001
DNA Polymerase I, Large (Klenow) Fragment	New England BioLabs	Cat#: M0210
Halt Phosphatase Inhibitor Cocktail	Thermo Scientific	Cat#: 78426
RNaseOUT Recombinant Ribonuclease Inhibitor	Invitrogen	Cat#: 10777019
Ribonucleoside Vanadyl Complex	New England BioLabs	Cat#: S1402
TURBO DNase	Invitrogen	Cat#: AM2238
GalTY289L	Gift from K. Moremen	N/A
UDP-GalNAz	Chemily	Cat#: SN02012
Amino-dPEG12-Tris (dPEG12-Tris(m-dPEG11)3)3	Quanta BioDesign	Cat#: 10482
DBCO-NHS ester	Click Chemistry Tools	Cat#: A133
Iodoacetamide	Sigma-Aldrich	Cat#: I1149
Dichloromethane	Sigma-Aldrich	Cat#: D65100
TaqMan Fast Advanced Master Mix	Applied Biosystems	Cat#: 4444557
TRIzol Reagent	Invitrogen	Cat#: 15596026
16% Formaldehyde (w/v), Methanol-free	Thermo Scientific	Cat#: 28908
TO1-3PEG-Biotin	Applied Biological Materials (abm)	Cat#: G955
CKII peptide (PGGSTPVSSANMM)	Synthesized by JHMI Synthesis & Sequencing Facility	N/A
Critical commercial assays
Pierce BCA Protein Assay Kit	Thermo Scientific	Cat#: 23227
UDP-Glo Glycosyltransferase Assay kit	Promega	Cat#: V6971
In-Fusion HD Cloning Plus kit	TaKaRa Bio	Cat#: 638909
Purelink Fast Low Endotoxin Midi Plasmid Purification kit	Invitrogen	Cat#: A35892
Pierce Protein A/G Magnetic Beads	Thermo Scientific	Cat#: 88802
Ni-NTA Agarose	Qiagen	Cat#: 30210
Pierce Glutathione Agarose Beads	Thermo Scientific	Cat#: 16100
NuPAGE 4 to 12%, Bis-Tris, 1.0 mm, Midi Protein Gel, 20-well	Invitrogen	Cat#: WG1402BOX
Amersham Hybond P 0.2 PVDF membrane	GE Healthcare/Cytiva	Cat#: 10600021
SuperSignal West Femto ECL Substrate	Thermo Scientific	Cat#: 34095
SuperSignal West Pico PLUS ECL Substrate	Thermo Scientific	Cat#: 34577
PARIS Kit	Invitrogen	Cat#: AM1921
Duolink PLA Reagents	Sigma-Aldrich	Cat#: DUO92001, DUO92005, DUO92013, DUO82049
AlexaFluor 488 phalloidin	Invitrogen	Cat#: R37110
AlexaFluor 555 phalloidin	Invitrogen	Cat#: R37112
ProLong Glass Antifade Mountant with NucBlue Stain	Invitrogen	Cat#: P36985
CUTANA ChIC/CUT&RUN Kit	EpiCypher	Cat#: 14-1048
Qubit 1X dsDNA HS Assay Kit	Invitrogen	Cat#: Q33230
NEBNext Ultra II DNA Library Prep with Sample Purification Beads	New England BioLabs	Cat#: E7103L
NEBNext Multiplex Oligos for Illumina (96 Unique Dual Index Primer Pairs)	New England BioLabs	Cat#: E6440S
Nano-Glo Dual-Luciferase Reporter Assay System	Promega	Cat#: N1610
Deposited data
Human Reference Genome GRCh38.p13	Ensembl	http://www.ensembl.org/Homo_sapiens/Info/Index
Escherichia coli Reference Genome ASM584v2 (str. K12 substr. MG1655)	Ensembl Bacteria	https://bacteria.ensembl.org/Escherichia_coli_str_k_12_substr_mg1655_gca_000005845/Info/Index
CUT&RUN Sequencing Data (Raw and Analyzed)	This paper	GEO: GSE190172
CUT&RUN Sequencing Data, shβ-catenin samples (Raw and Analyzed)	This paper	GEO: GSE214775
CUT&RUN Sequencing Data, Ac5S treated samples (Raw and Analyzed)	This paper	GEO: GSE214774
RNA Sequencing Data (Raw and Analyzed)	This paper	GEO: GSE189378
RNA Sequencing Data, Ac5S treated samples (Raw and Analyzed)	This paper	GEO: GSE189622
RNA Sequencing Data, shEZH2 samples (Raw and Analyzed)	This paper	GEO: GSE214777
Raw and analyzed data of this paper	This paper	Figshare: https://doi.org/10.6084/m9.figshare.17129945.v4
Experimental models: Cell lines
HEK293T	ATCC	Cat#: CRL-3216, RRID: CVCL_0063
L Wnt-3A	Willert et al.79 ATCC	Cat#: CRL-2647; RRID: CVCL_0635
T1	This paper	N/A
NL1F50	This paper	N/A
3JB8F+12	This paper	N/A
shβ-catenin	This paper	N/A
shEZH2	This paper	N/A
Experimental models: Organisms/strains
Oligonucleotides
DNA oligos for SELEX, see Table S6	This paper	N/A
Primers and Probes for TaqMan Assays, see Table S6	This paper	N/A
Aptamer Sequences, see Table S7	This paper	N/A
DNA oligo: NH2-T24: /5AmMC6/TTTTTTTTTTTTTTTTTTTTTTTT	Chang et al.73 Synthesized by IDT	N/A
Recombinant DNA
pET24a-ncOGT-FL	This paper	RRID: Addgene_190821
pET24a-ncOGT-TPR	This paper	RRID: Addgene_190822
pET28a-β-catenin	Gift from R. Moon, unpublished	RRID: Addgene_17198
pGEX-6P-1-CpOGA	Rao et al.35	N/A
pEGFP-β-catenin_WT	Murase et al.76	RRID: Addgene_16071
pEGFP-β-cateninS23A, T40A, T41A, T112A	This paper	RRID: Addgene_190824
pEGFP-C1-ERα	Stenoien et al.77	RRID: Addgene_28230
mEmerald c-SRC	Gift from M. Davidson, unpublished	RRID: Addgene_54118
GFP-AMPKα2	Hudson et al.78	N/A
pSilencer2.1-U6-T1	This paper	RRID: Addgene_190825
pSilencer2.1-U6-AP3	This paper	RRID: Addgene_190826
pSilencer2.1-U6-17-6 (Non-aptamer RNA)	This paper	RRID: Addgene_190827
pSilencer2.1-U6-T1-U6-AP3 (T1 ∙ AP3)	This paper	RRID: Addgene_190828
pSilencer2.1-U6-NL1F30	This paper	RRID: Addgene_190829
pSilencer2.1-U6-NL1F35	This paper	RRID: Addgene_190830
pSilencer2.1-U6-NL1F50	This paper	RRID: Addgene_190831
pSilencer2.1-U6-3JB1F	This paper	RRID: Addgene_190832
pSilencer2.1-U6-3JB1F+2	This paper	RRID: Addgene_190833
pSilencer2.1-U6-3JB1F+4	This paper	RRID: Addgene_190834
pSilencer2.1-U6-3JB1F+6	This paper	RRID: Addgene_190835
pSilencer2.1-U6-3JB1F+8	This paper	RRID: Addgene_190836
pSilencer2.1-U6-3JB1F+10	This paper	RRID: Addgene_190837
pSilencer2.1-U6-3JB1F+12	This paper	RRID: Addgene_190838
pSilencer2.1-U6-3JB1R	This paper	RRID: Addgene_190839
pSilencer2.1-U6-3JB1R+2	This paper	RRID: Addgene_190840
pSilencer2.1-U6-3JB1R+4	This paper	RRID: Addgene_190841
pSilencer2.1-U6-3JB1R+6	This paper	RRID: Addgene_190842
pSilencer2.1-U6-3JB1R+8	This paper	RRID: Addgene_190843
pSilencer2.1-U6-3JB1R+10	This paper	RRID: Addgene_190844
pSilencer2.1-U6-3JB1R+12	This paper	RRID: Addgene_190845
pSilencer2.1-U6-4JC1FF	This paper	RRID: Addgene_190846
pSilencer2.1-U6-4JC1RR	This paper	RRID: Addgene_190847
pSilencer2.1-U6-bc339	This paper	RRID: Addgene_190848
pSilencer2.1-U6-T1-U6-bc339 (T1 ∙ bc339)	This paper	RRID: Addgene_190849
pSilencer2.1-U6-NL8F50	This paper	RRID: Addgene_190850
pSilencer2.1-U6-NL8F70	This paper	RRID: Addgene_190851
pSilencer2.1-U6-NL8F100	This paper	RRID: Addgene_190852
pSilencer2.1-U6-3JB8F+4	This paper	RRID: Addgene_190853
pSilencer2.1-U6-3JB8F+12	This paper	RRID: Addgene_190854
pSilencer2.1-U6-3JB8R+12	This paper	RRID: Addgene_190855
pSilencer2.1-U6-LRS1F50C	This paper	RRID: Addgene_190856
TOPFlash	Veeman et al.85	RRID: Addgene_12456
FOPFlash	Veeman et al.85	RRID: Addgene_12457
pNL1.1.TK[Nluc/TK]	Promega	Cat#: N1501
TRC2-pLKO.5-scrambled	Sigma	Cat#: SHC216
TRC2-pLKO.5-shOGT	Sigma	Cat#: TRCN0000286200
TRC2-pLKO-shβ-catenin	Sigma	Cat#: TRCN0000314921
pLKO.1-shEZH2	Sigma	Cat#: TRCN0000040077
pMD2.G	Gift from D. Trono, unpublished	RRID: Addgene_12259
psPAX2	Gift from D. Trono, unpublished	RRID: Addgene_12260
pcDNA3.1-OGT_opt	This paper	RRID: Addgene_190858
pcDNA3.1-empty	This paper	RRID: Addgene_190859
Software and algorithms
cutadapt	Martin65	https://cutadapt.readthedocs.io/en/stable/
Trim Galore!	Krueger et al.88	https://github.com/FelixKrueger/TrimGalore
FASTAptamer	Alam et al.66	https://github.com/FASTAptamer/FASTAptamer
RNAfold webserver	Lorenz et al.67	http://rna.tbi.univie.ac.at/cgi-bin/RNAWebSuite/RNAfold.cgi
RNAStructure, version 6.3	Reuter and Mathews68	https://rna.urmc.rochester.edu/RNAstructure.html
RNAComposer	Popenda et al.69 Antczak et al.70	http://rnacomposer.cs.put.poznan.pl/
QRNAS	Stasiewicz et al.71	http://genesilico.pl/software/standalone/qrnas
UCSF ChimeraX	Pettersen et al.72	https://www.cgl.ucsf.edu/chimerax/
Biacore Evaluation Software	GE Healthcare/Cytiva	29310602
ggplot2	Wickham74	https://ggplot2.tidyverse.org
Fiji	Schindelin et al.75	https://imagej.net/software/fiji/
BioVoxxel Toolbox	Brocher86	https://imagej.net/plugins/biovoxxel-toolbox
HISAT2	Kim et al.89	http://daehwankimlab.github.io/hisat2/
featureCounts	Liao et al.90	http://subread.sourceforge.net/
annotateMyIDs	Dunning91	https://github.com/markdunning/galaxy-annotateMyIDs
DESeq2	Love et al.92	https://bioconductor.org/packages/release/bioc/html/DESeq2.html
EnhancedVolcano	Blighe et al.93	https://github.com/kevinblighe/EnhancedVolcano
Bowtie2	Langmead and Salzberg96	http://bowtie-bio.sourceforge.net/bowtie2/index.shtml
samtools	Li et al.97	http://www.htslib.org/
bedtools	Quinlan and Hall98	https://bedtools.readthedocs.io/en/latest/
SEACR	Meers et al.99	https://github.com/FredHutch/SEACR
The Integrative Genomics Viewer (IGV)	Robinson et al.105	http://software.broadinstitute.org/software/igv/home
Diffbind	Stark and Brown100 Ross-Innes et al.101	https://bioconductor.org/packages/release/bioc/html/DiffBind.html
ChIPseeker	Yu et al.102	https://bioconductor.org/packages/release/bioc/html/ChIPseeker.html
profileplry	Carroll and Barrows103	https://www.bioconductor.org/packages/release/bioc/html/profileplyr.html
GraphPad Prism	GraphPad Software	https://www.graphpad.com/
Other
Biacore T100 SPR system	GE Healthcare/Cytiva
Sensor Chip PEG, Series S	GE Healthcare/Cytiva	Cat#: 29239810
Slide-A-Lyzer G2 Dialysis Cassettes, 20K MWCO	Thermo Scientific	Cat#: 87735
Vivaspin protein concentrator, 30K MWCO	GE Healthcare/Cytiva	Cat#: 28932235
SpectraMax i3x	Molecular Devices
iBright FL 1500 Imaging System	Invitrogen	Cat#: A44115
8-well Chamber Slide w/ removable wells	Thermo Scientific	Cat#: 154941
HybEZ II Hybridization System	Advanced Cell Diagnostics	Cat#: 321710
Amersham Imager 600 RGB	GE Healthcare/Cytiva	Cat#: 29-0834-67

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Cell Lines

The HEK293T (CRL-3216) and L Wnt-3A (CRL-2647) cell lines were obtained from ATCC. The T1, NL1F50 and 3JB8F+12 cell lines were generated by transfecting the respective plasmids into HEK293T cells using TransIT-X2 (Mirus Bio, MIR 6000), and selected with 500 μg/mL hygromycin. Cell lines expressing shβ-catenin and shEZH2 were generated by lentiviral transduction of HEK293T cells, and selected with 1 μg/mL puromycin. Experimental details are described below.

METHOD DETAILS

SELEX

Library Construction

Systematic evolution of ligands by exponential enrichment (SELEX) was based on previous publications^60–62. DNA oligos used in SELEX are listed in Table S6. The initial single-stranded DNA (ssDNA) libraries (L1-N60-ncs for OGT aptamers, L2-N50-ncs for β-catenin aptamers) were synthesized by Integrated DNA Technologies (Coralville, IA, USA), using a customized recipe (A: 29%, C: 29%, G: 19%, T: 23%) for the random regions. ssDNA libraries were purified by Polyacrylamide gel electrophoresis (PAGE, Invitrogen EC6875BOX) and eluted from gel with Probe Elution Buffer (0.5 M Sodium Acetate, 1 mM EDTA, 0.2% SDS). Purified ssDNA was converted to double-stranded DNA (dsDNA) by annealing to L1–5’-T7-cs (for OGT aptamers) or L2–5’-T7-cs (for β-catenin aptamers) and Klenow extension (NEB M0210). RNA libraries were synthesized from PAGE-purified dsDNA libraries by in vitro transcription with T7 RNA polymerase (NEB M0251). RNA libraries were PAGE-purified, precipitated with isopropanol (Sigma-Aldrich I9516–500mL), washed with 70% ethanol and dissolved in nuclease-free water.

Partition and Amplification

To improve target specificity and to minimize PCR bias, “Toggle SELEX” and “RAPID-SELEX” was applied in the partition and amplification steps^63,64. Detailed schemes of partition and amplification are described in Table S2. Partition of library was performed in Aptamer Binding Buffer (ABB, 20 mM Tris-HCl, 150 mM NaCl, 5 mM MgCl₂, 0.05% Tween-20, pH 7.5) for OGT aptamers, or Intracellular Buffer (IB, 20 mM HEPES, 110 mM KCl, 10 mM NaCl, 0.4 mM MgCl₂, 5 mM KH₂PO₄, 0.05% Tween-20, pH 7.2) for β-catenin aptamers. Each RNA library and its desired target protein were incubated at 37 °C for 1 hr with rotation. The target proteins were either pre-immobilized on His-tag binding magnetic beads (MB, Invitrogen 10103D), or captured by nitrocellulose membrane disks (NC, Millipore HATF02500) after incubation. Unbound RNA sequences were washed away with 1 mL of ABB or IB. Elution of bound RNA was either by Proteinase K (Thermo Scientific EO0492) digestion at 37 °C for 30 min followed by denaturation in 100μL formamide (Thermo Scientific 17899) at 95°C for 2 minutes, or with 300 mM imidazole. Eluted RNA was converted to cDNA using SuperScript IV Reverse Transcriptase (Invitrogen 18090010) and RT primers (L1–3’-const-ncs for OGT aptamers; L2–3’-const-ncs for β-catenin aptamers), then PCR amplified using Taq Polymerase (NEB M0273L) for 16 cycles. Counter selection (CS) was included in indicated rounds with 1 μg Hexa-His peptide (Abbiotec 350220) immobilized onto MB. In certain rounds of selection, eluted RNA was proceeded to the next selection cycles without amplification, in order to reduce amplification bias and time consumption⁶⁴. Eluted RNA in the final cycles was converted to cDNA by RT-PCR (Invitrogen 18090010).

Library Sequencing and Data Analysis

For OGT aptamers, the final DNA library was cloned into the pCR4-TOPO plasmid using the TOPO TA cloning kit (Invitrogen 450030) and transformed into Chemically Competent E. coli cells (TaKaRa 636766). 84 colonies were subjected to plasmid purification and Sanger-sequencing at the JHMI Synthesis & Sequencing Facility. Frequency of all resulted sequences were calculated, and the most represented sequence (T1) was chosen for validation.

For β-catenin aptamers, enriched DNA libraries from rounds 5 and 8 were submitted for Next-Generation Sequencing at the JHMI Transcriptomics and Deep Sequencing Core Facility, on an Illumina HiSeq 2500 sequencer.

Analysis of the sequencing data started from trimming of the sequencing adaptors and the constant flanking regions using cutadapt⁶⁵. Enrichment analysis was carried out using FASTAptamer⁶⁶. Frequencies of all sequences in each library were counted and clustered based on their similarity. Enrichment of each cluster was assessed across different SELEX rounds. The most enriched sequence (bc339) was chosen for validation.

RNA Structure Prediction

Secondary structures of RNA were predicted using the RNAfold webserver⁶⁷ and drawn with StructureEditor in the RNAStructure 6.3 package⁶⁸. Tertiary structures were predicted using RNAComposer^69,70, refined by QRNAS⁷¹, and visualized with UCSF ChimeraX⁷².

Surface Plasmon Resonance (SPR)

For SPR, Multi Cycle Kinetics was performed on a Biacore T100 (GE Healthcare/Cytiva) instrument at 25 °C, following the protocol by Chang et al.⁷³. In assay preparation, an amine-modified T24 DNA linker (NH2-T24) was immobilized to ~100 RU in both reference and test flow cells on a sensorchip PEG (GE Healthcare/Cytiva 29239810), using the Amine Coupling Kit (GE Healthcare/Cytiva BR100050). RNA of T1 or bc339 with a A28 tail on the 3’-end were diluted to 100 nM in ABB (T1) or IB + 200 mM NaCl (bc339). Diluted RNA solutions were denatured at 65 °C for 5 min and re-folded at room temperature for 5 min. At the beginning of each cycle, the A28-tagged RNA solutions were injected into the test flow cell and captured by the immobilized T24 DNA linker. Protein analytes were diluted in ABB (OGT) or IB (β-catenin) and injected sequentially into both reference and test flow cells, at a rate of 30 μL/min. Concentrations tested are listed in the figure legends. Each concentration was duplicated. At the end of each cycle, surface was regenerated with 25 mM of NaOH for 30 sec at 30 μL/min, which removed the captured RNA and protein. Data was analyzed using the Biacore Evaluation Software (GE Healthcare/Cytiva), where the overlaid sensorgrams were fitted into a 1:1 binding model. Curves were exported and replotted with ggplot2⁷⁴ for better illustration.

Radiolabeled Dot-Blot Assay

Radio-Labeling of RNA

For end preparation, the terminal phosphate groups of RNA were removed by Antarctic Phosphatase (NEB M0289). Then the 5’-end of 200 pmol RNA was labeled with 10 μCi of [γ−32P]-ATP (American Radiolabeled Chemicals, ARP 0102F-250 μCi) using T4 Polynucleotide Kinase (NEB M0201). Labeled RNA was extracted with Acid-Phenol: Chloroform, pH4.5 (Invitrogen AM9720), precipitated with isopropanol, washed with 70% ethanol, and dissolved in nuclease-free water.

Dot-Blot Assay

Radiolabeled RNA was diluted in ABB, denatured at 65 °C for 5 min and re-folded at room temperature for 5 min. Then 100 nM of RNA was incubated with multiple concentrations (see figure legends) of purified ncOGT at 37 °C for 30 min. The volume of each assay was 50 μL. 1 μM yeast tRNA (Invitrogen AM7119) and 50 μg/mL Bovine Serum Albumin (BSA, NEB B9001S) were included as non-specific competitors. After incubation, the reactions were filtered sequentially through a nitrocellulose membrane (Amersham 10600042) and a Nylon membrane (Invitrogen AM10102), on a Bio-Dot Microfiltration Apparatus (Bio-Rad 170–6545). Membranes were washed with 100 μL ABB to remove unbound RNA. Washed membranes were air-dried and exposed to a film (Amersham 28906836) for 2 hr. The film was developed and scanned, and the image was analyzed with Fiji⁷⁵. Quantified data was fitted into a Michaelis-Menten model using GraphPad Prism.

Protein Expression and Purification

Bacterial Expression of Recombinant Proteins

For OGT-FL, TPR and β-catenin: NiCo21 (DE3) Competent E.coli (NEB C2529H) transformed with a plasmid (pET24a-ncOGT-FL, pET24a-TPR or pET28a-β-catenin) was inoculated into 50 mL LB medium for overnight culture at 37 °C, 250 rpm. On Day 2, the whole culture was transferred into 500 mL LB medium, and the growth was observed by measuring its OD₆₀₀ every 30 min. When the OD₆₀₀ reached 1.2, the whole culture was cooled on ice for 15 min. Protein expression was induced by adding 200 μM (OGT-FL and TPR) or 500 μM (β-catenin) IPTG (Invitrogen 15529019). Then the culture was continued at 16 °C, 180 rpm for 24 hr. Cells were harvested by centrifugation at 4°C, 5000 rcf for 15 min and washed with cold PBS.

For CpOGA: NiCo21 (DE3) cells transformed with pGEX-6P-1-CpOGA³⁵ was cultured and induced in the same way as above, except that protein expression was induced with 200 μM IPTG at 22°C, 250 rpm for 24 hr.

Protein Purification

For OGT-FL, TPR and β-catenin: All steps in protein purification were performed at 4 °C. Protease Inhibitor Cocktail (Roche 11873580001) was added into the Stock, Lysis, Wash and Elution Buffers below. Bacterial pellet was re-suspended in 10 mL Stock Buffer (50 mM Tris, 300 mM NaCl, pH 8.0). Cells were lysed by adding 10 mL Lysis Buffer (Stock Buffer + 10 mM Imidazole, 1% Triton X-100, and 10 mg/ml Lysozyme (Thermo Scientific 89833)) and incubate on ice for 30 min. Lysate was homogenized by sonication on ice for 5 cycles of 12 sec ON, 30 sec OFF at power 4 (Fisher Scientific F550), centrifugation at 13,000 rcf for 15 min and collection of the supernatant. The supernatant was loaded onto a chromatography column packed with 2 mL Ni-NTA agarose slurry (Qiagen 30210) and passed the column twice under gravity flow. Then the column was washed with 45 mL Wash Buffer (Stock Buffer + 30 mM Imidazole), and protein was eluted with 4 mL Elution Buffer (Stock Buffer + 250 mM Imidazole). Elute was dialyzed three times against 4 L of Dialysis Buffer (20 mM Tris pH 7.5, 50 mM NaCl and 0.2 mM PMSF (Roche 10837091001) in a dialysis cassette (20K MWCO, Thermo Scientific 87735). The first and third dialysis were for 3 hr, and the second was for overnight. Dialyzed protein solution was concentrated with a Vivaspin protein concentrator (30K MWCO, Cytiva 28932235). Protein concentration was measured by BCA assay (Thermo Scientific 23227). Purified proteins were stored at −80 °C for up to a year.

For CpOGA: Bacterial pellet was re-suspended and lysed in the same way as above, except that the Stock Buffer and Lysis Buffer were replaced with Buffer A (50 mM HEPES, 250 mM NaCl, pH7.5) and Lysis Buffer A (Buffer A + 10 mg/mL lysozyme). The cleared lysate was loaded to an equilibrated column packed with 2 mL Glutathione Agarose Beads (Thermo Scientific 16100). The loaded column was capped and rotated for 2.5 hr at 4 °C, to capture the GST-tagged CpOGA from lysate. Then lysate was allowed to pass through the column by gravity, and the beads was washed with 20 mL Buffer A. The CpOGA protein was eluted by cleaving the GST-tag. The cleavage was performed by adding 100 U PreScission Protease (GenScript Z02799–100U) in 5 mL Cleavage Buffer (50 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 1 mM DTT, pH 7.0 at 25 °C) to the column, and rotating for overnight at 4 °C. The elute was collected and dialyzed against 4 L of 1x TBS for overnight at 4 °C in a dialysis cassette (20K MWCO, Thermo Scientific 87735). The dialyzed protein solution was concentrated, quantified and stored as above.

OGT Activity Assay (UDP-Glo)

Activity of OGT was measured using the UDP-Glo Glycosyltransferase Assay kit (Promega V6971), following the manufacturer’s protocol. Assays were performed on a white 96-well half-area plate (Corning CLS3992–25EA). Each reaction contains 0.625 ng OGT, 50 μM CKII peptide substrate (PGGSTPVSSANMM, JHMI Synthesis & Sequencing Facility), 10 μM yeast tRNA and various concentrations (see figure legends) of RNA aptamer in 12.5 μL ABB. Reactions were incubated at room temperature for 15 min, and 12.5 μL UDP Detection Reagent was added to each reaction. After incubation at room temperature for 1 hr, luminescence was measured on a Plate Reader (SpectraMax i3x, Molecular Devices).

Cell Culture and Transfection

Cells were cultured in DMEM with 1 g/L glucose, pyruvate and GlutaMAX (Gibco, 10567022), supplemented with 10% heat-inactivated Fetal bovine serum (Gibco 16140071), 200 U/mL penicillin and 200 μg/mL streptomycin (Gibco 15140163). Culture medium was changed daily. Cells were cultured in a 37 °C incubator (Heracell VIOS 160i) with 5% CO₂ and 95% humidity. For subculturing, cells were rinsed with PBS, monodispersed with TrypLE (Gicbo 12604013) and counted on a cell counter (Nexcelom Auto 1000). Cells were discarded when their passage numbers reached 15.

For plasmid transfection in 10 cm dishes, 2×10⁶ HEK293T cells were seeded per dish, and transfection was done after 24 hr (at 40% – 60% confluency). Desired amounts of plasmids were dissolved in 100 μL Opti-MEM medium (Gibco 51985034), and TransIT-X2 Transfection Reagent (Mirus Bio, MIR 6000) was added at a ratio of 4 μL of Reagent per 1 μg of DNA. The mixture was gently mixed, incubated at room temperature for 20 min and distributed evenly into the culture dish. Medium was changed to fresh DMEM at 6 hr post-transfection. Plasmid usage in transfection: To target OGT to GFP-β-catenin, 0.1 μg of GFP-β-catenin⁷⁶ and 1 μg of each aptamer-encoding plasmid were co-transfected per 10 cm dish. To target OGT to other GFP-tagged proteins, 0.5 μg of GFP-Src (gift from M. Davidson, unpublished), or 1 μg of GFP-ERα/GFP-AMPKα2^77,78 and 2 μg of each aptamer-encoding plasmid were co-transfected per 10 cm dish. To target OGT to endogenous β-catenin, 1.5 μg of each plasmid was transfected per 10 cm dish. For cells cultured in different containers, the number of cells, amounts of DNA and transfection reagent were adjusted according to the bottom area of the container.

For cell treatments, Ac₄-5SGlcNAc (Ac5S) was added into medium at 50 μM, for 20 hr; Thiamet-G (TMG) was used at 2 μM for 6 hr; TO1-biotin (TO1) was used at 750 nM for 8 hr. Wnt stimulation was by 50% (v/v) Wnt3A-conditioned medium for 4 hr.

Generation of Stable Cell Lines

For cells stably expressing T1, NL1F50 and 3JB8F+12: the desired plasmids were transfected into HEK293T cells as above and selected with 500 μg/mL hygromycin B (Gibco 10687010) in culture medium. The selection started at 24 hr post-transfection and lasted for ~3 weeks, till most of cells died and the survived cells grew confluent. Then the cells were sub-cultured and stored in liquid nitrogen.

Cells expressing shβ-catenin and shEZH2 were generated by lentiviral transduction.

Production of Lentivirus

6×10⁵ HEK293T cells were seeded on a 6 cm dish. After 24 hr, 1 μg of shβ-catenin (Sigma TRCN0000314921) or 1 μg of shEZH2 (Sigma TRCN0000040077) plasmids was co-transfected with 0.75 μg of psPAX2 (Addgene 12260) and 0.25 μg of pMD2.G (Addgene 12259). Culture medium was replenished at 6 hr post-transfection. Culture continued without medium change, and the first batch of medium was harvested at 48 hr post-transfection. Medium was replenished and the second bath of medium was harvested at 72 hr post-transfection. The two batches were combined, spun at 1250 rpm for 5 min to remove cell pellet, and stored at −80 °C in 1 mL aliquots.

Lentiviral Transduction and Selection

6×10⁵ HEK293T cells were seeded on a 6 cm dish. After 24 hr, culture medium was changed to fresh DMEM with 8 μg/mL DEAE-Dextran, and 1 mL of thawed lentivirus solutions was added to the cell, and mixed by gently rocking. Culture medium was changed at 24 hr post-transfection. Antibiotic selection started at 48 hr post-transfection with 1 μg/mL puromycin (Sigma-Aldrich P8833) in medium, and continued till most cells died and the survived cells grew confluent. Then the cells were sub-cultured and stored in liquid nitrogen.

Production of Wnt3A-Conditioned Medium

L Wnt-3A cells⁷⁹ was sub-cultured at 1:10 in 15 cm dishes. Medium glucose level was monitored twice daily with OneTouch Ultra 2 meter and OneTouch Ultra Blue test strips. When glucose was below 20 mg/dL, it was replenished with 100 g/L glucose solution (Sigma-Aldrich G8270) that was filtered through 0.2 μm PES filter (Whatman 6896–2502), to about 120 mg/dL. The first batch of medium was harvested on the fourth day of culture, and fresh DMEM was added to the cells. Culture continued for another three days, and the second batch of medium was harvested. The two batches were combined, filtered through a 0.2 μm PES bottle-top filter (Fisher Scientific FB12566510), and stored at 4 °C for up to 6 months. This conditioned medium was diluted to 50% (v/v) with regular DMEM for cell treatment.

Molecular Cloning

For molecular cloning, we used the seamless cloning method using the In-Fusion HD Cloning Plus kit (TaKaRa Bio 638909), following manufacturer’s protocol.

Sequences encoding individual and DS aptamers were cloned into the pSilencer2.1-U6 hygro plasmid. The insert fragments were generated in three ways: Klenow extension, Gene Synthesis and PCR amplification. Klenow extension was used for fragments smaller than 100 bp without repetitive sequences. Two DNA oligos representing the 5’-end and 3’- end of the insert with 15–20 bp overhang were synthesized by Integrated DNA Technologies (Coralville, IA, USA). 1 nmol of each oligo was diluted in 20 μL Klenow Annealing Buffer (10 mM Tris-HCl, 10 mM MgCl₂, pH 8.0), denatured at 95 °C for 2 min and annealed to each other at room temperature for 5 min. Extension reactions were carried out using Klenow Fragment (NEB M0210). DNA was purified on 4% agarose gel and extracted with a gel extraction kit (Takara Bio 740609). Inserts with repetitive sequences or longer than 100 bp, and codon-optimized sequences were synthesized by GeneArt (ThermoFisher) and GeneScript (New Jersy, USA). All other inserts were amplified by PCR. 15 bp overhangs with desired linearized vectors were included on both ends of all inserts. Vectors were linearized by inverse PCR using Q5 High-Fidelity DNA Polymerase (NEB M0491). Ligation reactions were conducted with the In-Fusion HD Cloning Plus kit (TaKaRa Bio 638909). After ligation, chemically competent cells (TaKaRa 636766) were transformed with the ligation reaction, and single colonies were picked for sanger-sequencing. Colonies with correct sequences were stored as glycerol stock (20% glycerol) at −80 °C. For transfection of cultured mammalian cells, plasmids were purified with the PureLink Low-Endotoxin Midiprep Kit (Invitrogen A35892).

Point mutations were introduced by inverse PCR with overlapping primers containing the desired mutations. Products of inverse PCR were purified on agarose gels, and self-ligated using the In-Fusion HD Cloning Plus kit (TaKaRa Bio 638909). The ligation reactions were transformed into competent cells, and single colonies were picked and sequenced as above.

Immunoprecipitation (IP), Co-IP and RNA-IP

For immunoprecipitation (IP) of β-catenin, cells were cultured in 10 cm dishes, and harvested with 500 μL RIPA buffer (20 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% (w/v) sodium deoxycholate, 0.1% (w/v) SDS, 1% (v/v) NP-40, 50 mM NaF, 5 mM Na₄P₂O₇, 1mM EDTA and 1mM EGTA). 1 mM PMSF, 1 mM DTT, 1 μM PUGNAc, 2 μM TMG, 50 mM GlcNAc, 1x Protease Inhibitor Cocktail (Roche 11873580001) and 1x Phosphatase Inhibitor Cocktail (Thermo Scientific 78426) were freshly added. Cell lysate was sonicated at power 4 for 10 sec (Fisher Scientific F550), centrifuged at 17,000 rcf for 15 min. Supernatant was collected and protein concentration was measured by BCA assay (Thermo Scientific 23227). IP reactions were assembled with 0.5 mg of cell lysate and 0.5 μg of anti-β-catenin antibody (BD Biosciences 610153) or IgG control (Invitrogen MA1–10406), and the volume was adjusted to 0.5 mL with IP buffer (20 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% (v/v) TritonX-100, 50 mM NaF, 5 mM Na₄P₂O₇, 1mM EDTA and 1mM EGTA) with inhibitors freshly added. IP reactions were rotated at 4 °C for overnight. On the next day, 7.5 μL Protein A/G magnetic bead slurry (Thermo Scientific 88802) was added, and reactions were rotated at room temperature for 1 hr. Beads were collected on a magnetic stand and washed 5 times with 1 mL IP buffer. Proteins were eluted with 10 μL 2× Laemmli Buffer (Bio-Rad 161–0737) and denatured at 95 °C for 3 min. Denatured samples were loaded on a 4–12% SDS-PAGE gel (Invitrogen WG1402BOX) and proceeded to Western Blot protocols.

For IP and co-immunoprecipitation (Co-IP) of GFP, the same method for IP was used, except cells were lysed with Co-IP Lysis Buffer (150 mM NaCl, 30 mM Tris pH7.5, 1 mM EDTA, 1% TritonX-100, 10% glycerol) with the same inhibitors. IP reactions were assembled with 1 mg of cell lysate and 5 μL of GFP-trap magnetic agarose beads (ProteinTech gtma-20, for GFP-β-catenin), or 2 μg of anti-GFP antibody (Invitrogen G10362, for other GFP-tagged proteins) or IgG control (CST 2729). On the next day, 20 μL Protein A/G magnetic bead slurry (Thermo Scientific 88802) was added to capture the antibody-antigen complex.

For RNA-IP, we adopted the IP protocol with some modifications. 1 mM MgCl₂ was added into the RIPA and IP buffers for “+ Mg²⁺” samples. Cells were harvested in RIPA buffer with inhibitors, 100 U/mL RNaseOUT (Invitrogen 10777019) and 2 mM Ribonucleoside Vanadyl Complex (NEB S1402S) freshly added. IP reactions were assembled with 1 mg of cell lysate and 1 μg of anti-OGT antibody (AL24, purified in the lab)⁸⁰, and the volumes were adjusted to 1 mL with IP buffer (inhibitors freshly added). For IgG control, 1 μg normal rabbit IgG (Millipore 12–370) was used instead of the antibody. On the next day, 15 μL Protein A/G magnetic bead slurry (Thermo Scientific 88802) was added to capture the antibody-OGT-RNA complex. After washing, residual DNA was digested by 20 units of Turbo DNase (Invitrogen AM2238) in 200 μL Turbo DNase Buffer at 37 °C for 30 min. Then proteins were digested by 100 μg Proteinase K (Invitrogen 25530049) at 50 °C for 30 min. RNA was extracted with Acid-Phenol: Chloroform, pH4.5 (Invitrogen AM9720), precipitated with isopropanol, washed with 70% ethanol, and dissolved in nuclease-free water.

CpOGA Treatment

CpOGA treatment was performed on total cell lysate before IP. Cells were harvested with Co-IP Lysis Buffer with 1 mM DTT, 1 mM PMSF and 1x Protease Inhibitor Cocktail (Roche 11873580001). PUGNAc and GlcNAc were not included in the lysis buffer so that CpOGA would not be inhibited. 50 μg of purified CpOGA or equal volume of PBS was added to 0.5 mg of lysate, and incubated at 37 °C for 1 hr. Then the reaction was cooled on ice and subjected to IP.

RT-qPCR

Purified RNA was first converted into cDNA using the SuperScript IV Reverse Transcriptase (Invitrogen 18090010) and RT primers (Table S6). A “RT -” control reaction without the enzyme was included for all samples. Synthesized cDNA was quantified by qPCR on a Stratagene Mx3000P (Agilent Technologies) using TaqMan Fast Advanced Master Mix (Applied Biosystems 4444557), and desired TaqMan assays (Table S6). Amount of residual genomic/plasmid DNA was subtracted from each sample using the readout from its “RT -” control.

Western Blot

Eluted IP reactions or 20 μg of total cell lysates (in Laemmli Buffer and denatured at 95 °C) were loaded on a 4–12% SDS-PAGE gel (Invitrogen WG1402BOX) and run at 160 V for 90 min. Gel was transferred to a 0.2 μm pore-sized PVDF membrane (GE Healthcare/Cytiva 10600021). Wet transfer was performed in a Criterion Blotter (Bio-Rad 1704070) at 0.45 A for 120 min in Bis-Tris Transfer Buffer (25 mM Bicine, 25 mM Bis-Tris, 1 mM EDTA and 10% (v/v) Methanol). After transfer, membrane was blocked at room temperature for 30 min with rocking, in Blocking Buffer (5% (w/v) BSA (Rockland BSA-1000) dissolved in TBS-T). Blocked membrane was incubated with primary antibody (diluted in Blocking Buffer unless indicated below), at 4 °C for overnight with rocking. On the next day, the membrane was washed for 3 times of 5 min in TBS-T with rocking. Washed membrane was incubated with HRP-conjugated secondary antibody diluted in Blocking Buffer at room temperature for 1 hr with rocking. Then the membrane was washed for 3 times of 5 min, and imaged with ECL reagent (Thermo Scientific 34095) on an iBright FL1500 Imaging system (Invitrogen A44115) or an Amersham Imager 600 RGB (GE Healthcare/Cytiva 29–0834-67) under Chemiluminescence mode.

The images were analyzed using Fiji⁷⁵. Background signal was subtracted using a rolling-ball method (radius = 20 or 30), then the lanes were plotted, and band intensities were quantified. To quantify protein-specific O-GlcNAcylation, the band intensities from a O-GlcNAc blot were normalized to those of their corresponding protein-specific blot. Statistical analysis was performed with the GraphPad Prism software.

Dilutions of primary antibodies (in 5% (w/v) BSA unless indicated): anti-O-GlcNAc (CTD110.6)⁸¹ – 1:2000; anti-β-catenin (CST 9562) – 0.2 μg/mL; anti-β-catenin (ProteinTech 17565–1-AP, used in mass-shift assay) – 1:2000; anti-Non-phospho (Active) β-Catenin (Ser33/37/Thr41) – 1:1000; anti-Non-phospho (Active) β-Catenin (Ser45) – 1:1000; anti-phospho-β-Catenin (Ser33/37) – 2 μg/mL; anti-phospho-β-Catenin (Ser45) – 1:1000; anti-phospho-β-Catenin (Ser552) – 1:1000; anti-phospho-β-Catenin (Ser675) – 1:1000; anti-OGT (AL28)⁸² – 1:1000 in 5% (w/v) non-fat milk (Nestle NES22928) in TBS-T; anti-OGA (345)⁸³ – 1:1000 in 5% (w/v) non-fat milk (Nestle NES22928) in TBS-T; anti-EZH2 – 1:1000; anti-GFP (3E6) – 2 μg/mL; anti-β-tubulin – 1:1000.

Dilutions of secondary antibodies (in 5% (w/v) BSA): anti-Mouse IgM (Sigma-Aldrich A8786) – 1:2000; anti-Rabbit IgG (Invitrogen A27036) – 1:10000; anti-Mouse IgG (Invitrogen A28177) – 1:10000.

Mass-Shift Assay

Mass-shift assay was performed following the protocol by Hardiville et al.⁸⁴ with the following modifications. Cells were harvested in RIPA buffer, sonicated and cleared as mentioned above. In each GalT labeling reaction, 50 μg of total cell lysate was labeled using the GalT^Y289L enzyme (a kind gift from Dr. Kelly Moremen at CCRC) and UDP-GalNAz (Chemily SN02012). A “-GalT” control group without GalT^Y289L enzyme was included in all experiments. All 50 μg of the labeled sample was subjected to the SPAAC Click-chemistry reaction, where it was incubated with 10 mM of DBCO-PEG (8.5 kDa) or equal volume of DMSO at 37 °C for 16 hr. PEG-labeled proteins were precipitated with chloroform-methanol and subjected to Western blot. The shifted and unshifted bands of β-catenin was detected with a different antibody (ProteinTech 17565–1-AP). DBCO-PEG (8.5 kDa) was synthesized according to Hardiville et al.⁸⁴, except that Amino-dPEG₁₂-Tris (dPEG₁₂-Tris(m-dPEG₁₁)₃)₃ (Quanta BioDesign 10482) was used.

Pulse-Chase Experiments on RNA Half-life

HEK293T stable cell lines that express desired aptamers were cultured in 6-well plate with 1.5 mL medium. Each treatment was replicated in three wells. On day 1, 2.5×10⁵ cells were seeded in each well. Medium was changed daily. On day 3, actinomycin D was added into each well at 5 μg/mL, on a series of timepoints (see figure legends). When the treatment periods are over, cells were harvested by discarding the medium and adding 750 μL of TRIzol reagent (Invitrogen 15596026) into each well. The lysate was mixed and incubated at room temperature for 5 min and transferred into a Phasemaker tube (Invitrogen A33248). The aqueous phase was separated by adding 150 μL chloroform, thoroughly vertexing and incubating at room temperature for 5 min, and spinning at 12,000 rcf for 15 min at 4 °C. The upper aqueous layer was carefully collected without toughing the other layers. RNA was precipitated with isopropanol, washed with 70% ethanol, dissolved in nuclease-free water, and proceeded to RT-qPCR.

Nucleus/Cytoplasm Fractionation

Nucleus/Cytoplasm Fractionation was performed using the PARIS Kit (Invitrogen AM1921), following the kit manual. 100 U/mL RNaseOUT (Invitrogen 10777019) and 2 mM Ribonucleoside Vanadyl Complex (NEB S1402S) were freshly added to buffers. 2.5×10⁵ cells of the HEK293T stable cell line were cultured for 2 days on 6-well plates, 8 wells in total. Cells were monodispersed with TrypLE (Gibco 12604013) and rinsed with PBS. All steps between cell lysis and addition of 2X Lysis/Binding Solution were performed at 4 °C. 1.0×10⁶ cells were collected from each well. For the 4 replicates of Nucleus/Cytoplasm fractionation, cells were lysed by resuspending in 300 μL ice cold Cell Fractionation Buffer and incubating on ice for 10 min. The nuclear and cytoplasmic fractions were separated by spinning at 500 rcf for 5 min. The upper cytoplasmic fraction was carefully collected and transferred into a RNase free tube with equal volume of 2X Lysis/Binding Solution. The pelleted nuclear fraction was washed by resuspending in 300 μL cold Cell Fractionation Buffer and spinning at 500 rcf for 1 min. The nuclear pellet was collected and lysed with 300 μL ice cold Cell Disruption Buffer, followed by 300 μL of 2X Lysis/Binding Solution. For the 4 replicates of total cell lysate (control), collected cells were lysed in 300 μL ice cold Cell Disruption Buffer + 300 μL of 2X Lysis/Binding Solution. RNA was purified with the spin columns included in the kit and subjected to RT-qPCR.

TOPFlash/FOPFlash Luciferase Reporter Assay

Amount and time of transfection for each plasmid is listed in Table S3. On day 1, 2.5×10⁵ cells were seeded in each well on 6-well plates. After 24 hr, plasmids encoding the Firefly Luciferase (TOPFlash or FOPFlash)⁸⁵ and the NanoLuc Luciferase (pNL1.1.TK[Nluc/TK], Promega N1501) were transfected into cells. For OGT knockdown and rescue assays, plasmids encoding shOGT and codon-optimized ncOGT were also transfected at this time. On day 3, cells were detached with TrypLE, and 1.5×10⁴ of desired cells were seeded on a 96-well plate in 80 μL medium. On Day 4, cells were transfected with plasmids encoding desired aptamers. On Day 5, culture medium was changed to fresh DMEM (in −Wnt wells) or 50% Wnt3A-conditioned medium (in +Wnt wells) at 20 hr post-transfection. At 24 hr post-transfection, Dual Luciferase Reporter Assay was performed using the Nano-Glo Dual-Luciferase Reporter Assay System (Promega N1521), following the kit manual. Luminescence generated by Firefly Luciferase (TOPFlash or FOPFlash) and NanoLuc Luciferase was measured on a SpectraMax i3x plate reader (Molecular Devices). In data analysis, signal intensity of the Firefly Luciferase was normalized to that of NanoLuc Luciferase in the same well, as a control for variations in transfection efficiency. Then the normalized readings of the TOPFlash wells were normalized again to that of the FOPFlash wells with the same treatments. The second normalization was purposed to control the background (leaky) expression of Firefly Luciferase. Statistical analysis was performed with the GraphPad Prism software.

Proximity Ligation Assay (PLA)

For PLA, cells were cultured on 8-well Chamber Slides (Thermo Scientific 154941). 1.0×10⁴ cells were seeded in each well with 250 μL medium. After 24 hr, 19.7 ng/well of aptamer-encoding plasmids were transfected into cells (equivalent to 1.5 μg in 10 cm dish). Medium was changed 6 hr after transfection. At 24 hr post-transfection, the plastic chambers were removed, and cells were fixed with 4% formaldehyde (Thermo Scientific 28908, diluted in PBS) at room temperature for 15 min, rinsed twice with PBS, and permeabilized in freshly made Permeabilization Buffer (0.5% Triton X-100 in PBS) at room temperature for 20 min. Cells were washed twice in PBS, barriers were created around each well with a Hydrophobic Barrier Pen (Vector Laboratories H-4000), and proceeded to PLA. PLA was performed using the Duolink Probes, Buffers and Detection Reagents (DUO92001, DUO92005, DUO92013 and DUO82049), following the kit manual. Incubation steps were performed in a HybEZ Hybridization Oven with a humidity control tray (ACD 321711). After the final washes, cells were stained with AlexaFluor488-conjugated phalloidin (Invitrogen R37110) or AlexaFluor555-conjugated phalloidin (Invitrogen R37112) for 30 min at room temperature and washed in water for 1 min. Slides were air dried for 15 min at 60 °C in dark. Dried slides were mounted with Prolong Glass Mountant with NucBlue (Hoechst 33342) (Invitrogen P36985) and cover glasses. Mounted slides were left at room temperature for 24 hr before imaging.

Primary antibody usage: anti-β-catenin (BD Biosciences 610153) – 1:250; anti-GFP (3E6) – 0.8 μg/mL; anti-OGT (AL28) – 1:1000; anti-β-TrCP – 1:200; anti-EZH2 – 1:200; anti-EP300 – 0.5 μg/mL; anti-KAT2A/GCN5L2 – 1:100; anti-KAT5 – 1:100; anti-H3K27me3 – 1:800; anti-H3K9ac – 1:400.

Mounted slides were imaged on an Olympus FV1200 confocal microscope using an Olympus UPLFLN40XO lens (40x, NA = 1.3, Oil Immersion). Image acquisition was controlled by the Olympus FluoView (Ver. 4.2a) software. Samples were imaged under 405 nm (for NucBlue/Hoechst, HV = 650, Gain = 1, Offset = 5, Laser Power = 9.0%), 488 nm (for Phalloidin, HV = 630, Gain = 1, Offset = 5, Laser Power = 7.0%) or 559 nm (for Phalloidin, HV = 630, Gain = 1, Offset = 5, Laser Power = 5.0%) and 635 nm (for PLA, HV = 550, Gain = 1, Offset = 5, Laser Power = 6.0%). The pinhole was set to 1 Airy Unit (80 μm). 12-bit images of the three channels were acquired sequentially.

Analysis of the imaging data was done with Fiji⁷⁵. Total PLA puncta in an image were labeled and counted with the “Find maxima” function (Prominence > 500). For nuclear PLA puncta, total puncta in the PLA channel were first labeled in the same way, then the nuclear areas on the Hoechst channel were labeled by adjusting threshold (> 500). Nuclear PLA puncta were counted with the “Speckle Inspector” function in the BioVoxxel Toolbox⁸⁶, using the threshold-adjusted Hoechst channel as the “Primary objects”, and the maxima-labeled PLA channel as the “Secondary objects”. Number of nuclei was manually counted. the ratio of total PLA puncta / nuclei, or nuclear PLA / nuclei was calculated for each image. Statistical analysis was performed with GraphPad Prism.

For easy visualization in the figure panels, image contrast was enhanced by resetting the upper limit of display range in Fiji as follow: Hoechst and Phalloidin channels = 2047, PLA channel = 1023 (OGT, EZH2, KAT2A and KAT5 figures) or 2047 (β-TrCP and EP300 figures). Adjustments were applied only to the presented figures, not to the original images nor during image analysis.

RNA Sequencing and Data Analysis

HEK293T cells (wildtype or knockdown of EZH2) were cultured in 10 cm dishes and transfected with desired plasmids. Ac5S and Wnt treatments were performed as described in the Cell Culture and Transfection section. At 24 hr post-transfection, medium was discarded, and cells were lysed using 1 mL of TRIzol Reagent (Invitrogen 15596026) per culture dish. Lysates were mixed by pipetting, transferred into a Phasemaker tube (Invitrogen A33248) and incubated at room temperature for 5 min. The aqueous phase was separated by adding 200 μL chloroform, thoroughly vertexing and spinning at 12,000 rcf for 15 min at 4 °C. The upper aqueous layer was carefully collected without touching the other layers. RNA was precipitated with isopropanol, washed with 70% ethanol, dissolved in nuclease-free water, and submitted to GENEWIZ (South Plainfield, NJ, USA) for sequencing. Sequencing was done on an Illumina HiSeq 4000 sequencer in 2×150 bp mode.

Analysis of the RNA-Seq data followed the protocols by Doyle et al.⁸⁷. Sequencing adaptors were trimmed from the raw reads using Trim Galore^65,88 with parameters (--paired --clip_R1 10 --clip_R2 10 --three_prime_clip_R1 1 --three_prime_clip_R2 1). Trimmed reads were mapped to the human reference genome (GRCh38.p13) using HISAT2⁸⁹ with parameters (--dta). Mapped reads were counted to each genes by featureCounts⁹⁰ with parameters (-C -F GTF -g gene_id -p --countReadPairs -t transcript). The count matrix was annotated using annotateMyIDs⁹¹. Differential expression of genes was analyzed with DESeq2⁹² and plotted using EnhancedVolcano⁹³.

CUT&RUN Sequencing and Data Analysis

Cleavage Under Targets and Release Using Nuclease (CUT&RUN) was performed using the CUTANA ChIC/CUT&RUN kit (EpiCypher 14–1048), following the kit manual. DTT (1 mM), PUGNAc (1 μM), TMG (2 μM) and Protease Inhibitors Cocktail (1x) were freshly added into all buffers. Cell culture and transfection were carried out on 6-well culture plates. At 24 hr post-transfection, cells were monodispersed by scrapping and pipetting. 5.0×10⁵ cells were used for each reaction. Nuclei of cells were isolated by resuspending 5.0×10⁵ cells in 100 μL of cold Nuclear Extraction Buffer (20 mM HEPES pH 7.9, 10 mM KCl, 0.10% (v/v) Triton X-100, 20% (v/v) glycerol, inhibitors and 0.5 mM Spermidine freshly added) and incubating on ice for 10 min. Nuclei were collected by centrifugation at 600 rcf for 3 min at 4 °C, then re-suspended in 100 μL of cold Nuclear Extraction Buffer. Following steps were performed as instructed in the kit manual. 500 ng of EZH2 antibody (CST 5246) was used in each 50 μL of binding reaction. A negative control reaction using 500 ng of rabbit IgG (EpiCypher 13–0042k) was included. After chromatin digestion, 0.5 ng of E.coli genomic DNA (kit included) was spiked-in to each reaction with the Stop Buffer. After chromatin release, DNA fragments were purified by Phenol: Chloroform (Invitrogen 15593031) extraction and isopropanol precipitation. Concentrations of DNA were measured using the Qubit 1X dsDNA HS Assay Kit (Invitrogen Q33230).

Preparation of Sequencing Library was performed using the NEBNext Ultra II DNA Library Prep Kit (NEB E7103L), and NEBNext Multiplex Oligos for Illumina (NEB E6440S), following the protocol by Liu⁹⁴. 6 ng DNA from each CUT&RUN reaction was used as the starting material. A negative control (Input) with 6 ng of sonication-sheard genomic DNA from HEK293T cells was included. Prepared DNA library was dissolved in nuclease-free water, and submitted to GENEWIZ (South Plainfield, NJ, USA) for sequencing. Sequencing was performed on an Illumina HiSeq 4000 sequencer in 2×150 bp mode.

Analysis of the sequencing data followed the protocol by Zheng et al.⁹⁵. In brief, sequencing adaptors were trimmed from the raw reads using Trim Galore^65,88 with parameters (--paired --clip_R1 10 --clip_R2 10 --three_prime_clip_R1 1 --three_prime_clip_R2 1). Trimmed reads were mapped to the human reference genome (GRCh38.p13) using Bowtie2⁹⁶ with parameters (-I 10 -X 700 --no-mixed --no-discordant --dovetail --very-sensitive-local). For spike-in calibration, the trimmed reads were additionally mapped to the E.coli reference genome (ASM584v2) with parameters (-I 10 -X 700 --no-mixed --no-discordant --dovetail --very-sensitive --end-to-end), and scale factors were calculated according to the protocol⁹⁵. Mapped BAM files were filtered by samtools view⁹⁷ with paramters (-q 20 -F 0×4) and converted to BED format using bedtools bamtobed⁹⁸. The BED files were spike-in calibrated using the calculated scale factors, and converted to genomic coverage bedgraph files using the bedtools genomecov function⁹⁸. Peak calling was performed with SEACR⁹⁹ under the stringent mode with parameters (-n non -m stringent), using the bedgraph file of IgG as control. Differential binding analysis was performed using Diffbind^100,101, using the mapped BAM reads (to human and E.coli reference genomes) from Bowtie2 and the called peaks from SEACR as input with parameters (dba.blacklist: blacklist = DBA_BLACKLIST_GRCH38, greylist = FALSE; dba.normalize: method = DBA_DESEQ2, normalize = DBA_NORM_RLE, spikein = TRUE; dba.analyze: method = DBA_DESEQ2). The differential binding sites were annotated with ChIPSeeker¹⁰². The heatmap was generated using the dba.plotprofile function in Diffbind, and the profileplyr package¹⁰³. For visualization, normalized bedgraph files were converted to bigwig files using UCSC bedGraphToBigWig¹⁰⁴. The genomic browser views were generated by inspecting the bigwig files in the Integrative Genomics Viewer (IGV)¹⁰⁵.

Design of Riboswitch Elements

To generate the inducible aptamer LRS1F50C, two sets of riboswitch mechanisms each including a mango aptamer (Mango-II^A22U or Mango-III^A10U)^56,57 and a transmitter element were incorporated into the DS aptamer NL1F50 (Figures 7A and 7B). A transmitter element was designed to imperfectly pair with both the core sequence of a Mango aptamer, and with part of T1/AP3. This design allows the transmitter to disrupt the folding of either Mango or T1/AP3 by binding to one of them. This alternative binding is switched by addition/removal of the effector molecule (TO1, abm G955), which facilitates the correct folding of Mango aptamers. In addition, the linker region serves as an insulator, so that the two riboswitch mechanisms do not interfere with each other and can be designed individually.

Design of a transmitter element’s sequence was guided by the change of Gibbs Free Energy. For spontaneous processes, the change of Gibbs Free Energy must be negative (ΔΔG < 0). In the system consists of an RNA riboswitch (including a sensor aptamer, an actuator aptamer, and a transmitter), a ligand (TO1) and a target protein (OGT or a GFP-tagged protein), the ΔΔG of a conformational change is influenced by 1) the folding of the RNA; and 2) the binding of aptamers to the ligand and the target protein.

Taking the Mango-III^A10U riboswitch that regulates T1 as an example, its conformational change triggered by addition of TO1 (from the “ON” state shown in Figure 7A, to the “OFF” state shown in Figure 7B) should be a spontaneous process:

$$\Delta \Delta G(+TO1)=\Delta \Delta G_{\mathit{\text{RNA folding}}}(ONtoOFF)+\Delta \Delta G_{\mathit{\text{binding}}}(+TO1)<0$$ Eq.1

Meanwhile, removal of TO1 triggers the reverse conformational change:

$$\Delta \Delta G(-TO1)=\Delta \Delta G_{\mathit{\text{RNA folding}}}(OFFtoON)+\Delta \Delta G_{\mathit{\text{binding}}}(-TO1)<0$$ Eq.2

Note that

$$\Delta \Delta G_{\mathit{\text{RNA folding}}}(OFFtoON)=-\Delta \Delta G_{\mathit{\text{RNA folding}}}(ONtoOFF)$$ Eq.3

Eq.1 and Eq.2 are the criteria of a good riboswitch design.

In (1), after addition of TO1:

$$\Delta \Delta G_{\mathit{\text{binding}}}(+T01)=\Delta G_{\mathit{\text{binding}}}({\mathit{\text{Mango III}}}^{A10U}toTO1)-\Delta G_{\mathit{\text{binding}}}(T1toOGT)$$ Eq.4

In Eq.2, after removal of TO1:

$$\Delta \Delta G_{\mathit{\text{binding}}}(-T01)=\Delta G_{\mathit{\text{binding}}}(T1toOGT)$$ Eq.5

By combining Eq.1 to Eq.5, we get

$$\begin{gathered} \Delta G_{\mathit{\text{binding}}}(T1toOGT)<\Delta \Delta G_{\mathit{\text{RNA folding}}}(ONtoOFF) \\ <\Delta G_{\mathit{\text{binding}}}(T1toOGT)-\Delta G_{\mathit{\text{binding}}}({\mathit{\text{Mango III}}}^{A10U}toTO1) \end{gathered}$$ Eq.6

In a good riboswitch design, its ΔΔG_{RNA folding} (ON to OFF) must satisfy Eq.6.

In Eq.6,

$$\Delta \Delta G_{\mathit{\text{RNA folding}}}(ONtoOFF)=\Delta G_{\mathit{\text{RNA folding}}}(OFF)-\Delta G_{\mathit{\text{RNA folding}}}(ON)$$ Eq.7

In these formulars,

1) the ΔG_{RNA folding} items are sensitive to the transmitter sequences but insensitive to the presence of TO1, and they were calculated using the RNAfold webserver⁶⁷;

2) the ΔG_binding items are insensitive to the transmitter sequences, and they were calculated from the equation:

$$\Delta G_{\mathit{\text{binding}}}=-RT\text{\hspace{0.17em}ln\hspace{0.17em}}{K}_a=-RT\text{\hspace{0.17em}ln\hspace{0.17em}}{K}_d^{-1}=RT\text{\hspace{0.17em}ln\hspace{0.17em}}{K}_d$$ Eq.8

In Eq.8, R is the gas constant (8.314 J·K⁻¹·mol⁻¹) and T = 298.15 K. The dissociation constants (K_d) were measured by SPR (T1 to OGT: 60.49 nM, measured at 298.15 K) or reported previously (AP3 to GFP: 5.1 nM, measured at 295.15 K; Mango-II^A22U/Mango-III^A10U to TO1: 0.9/1.7 nM, temperature unknown)^26,56,57 were used in the calculations.

Sequences of the transmitter elements in both riboswitch mechanisms were manually designed to satisfy formulars Eq.6 to Eq.8.

QUANTIFICATION AND STATISTICAL ANALYSIS

Statistical analysis was performed in the GraphPad Prism 9 software. Unpaired two-tailed Student’s t-test was applied to comparisons between two groups. For single-variable comparisons between multiple groups, one-way ANOVA test followed by Dunnett’s test comparing to the control groups (T1 · AP3 or T1 · bc339) was applied. For two-variable comparisons, two-way ANOVA test followed by Turkey’s test was applied. Error bars in all plots represents Standard Deviation (SD). For experiments performed with cells, N represents the number of biological replicates. For PLA, N represents the number of different views analyzed. Details of sample number, data representation and statistical comparison method of each plot can be found in the figures and legends. ns, p ≥ 0.05; *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001.

Supplementary Material

1

Figure S1. Predicted Secondary Structures of Several Dual-Specificity Aptamers Used in This Research. Related to All Figures

(A) to (D) Sequences and predicted secondary structures of 3JB1F+6, 4JC1RR, 3JB8F+12 and 3JB8R+12, respectively.

2

Figure S2. Dual-Specificity Aptamers Increase O-GlcNAcylation on GFP-Tagged Proteins. Related to Figure 2

(A) IP-WB of HEK293T cells expressing GFP-β-catenin and the indicated aptamers. N=5.

(B) Optimization of co-transfection of GFP-β-catenin. IP-WB on cells co-transfected with various amounts of plasmids encoding GFP-β-catenin, and 1 μg plasmids encoding the indicated aptamers.

(C) Quantification of (B). Data are normalized to the control group (T1 · AP3) with 0.1 μg GFP-β-catenin plasmid. N=3.

(D) IP-WB on cells expressing GFP and the indicated aptamers. The lane with protein ladder is included as a positive control. N=3.

(E) WB of total cell lysates from Figure 2C.

(F) Mass-shift assays on cells expressing GFP-β-catenin and the indicated aptamers. Two blots on β-catenin with short and long exposure time are shown. Intensities of the shifted bands are normalized to that of β-tubulin. N=3.

(G) IP-WB of cells expressing GFP-β-catenin and the indicated aptamers, and stimulated with Wnt3A-conditioned medium for 4 hr. Irrelevant lanes are cropped. N=3.

(H) Schematic of 3JB1R.

(I) Optimization of the linker domain of 3JB1R. IP-WB on cells expressing GFP-β-catenin and the indicated aptamers. N=3.

(J) and (K) IP-WB (K), and Co-IP (K) on cells expressing GFP-Src and the indicated aptamers. N=3.

(L) IP-WB of cells expressing GFP-AMPKα2 and the indicated aptamers. N=3.

T1 · AP3 – Individual T1 and AP3 aptamers (control). NL1F35/50 – DS aptamers (T1-linker-AP3) with flexible linkers of 35/50 nt. 3JB1F/3JB1R/4JC1RR – DS aptamers (T1-linker-AP3) with folded linkers. +2/+4/…/+12, serial additions (bp) into the folded linker. Quantitated data are normalized to the control groups (T1 · AP3). Student’s t-test in (A), (F), (G). Two-way ANOVA test in (C). One-way ANOVA test in other panels. Data are represented as mean ± SD. ns, p ≥ 0.05; *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001.

3

Figure S3. DS Aptamers Reveal Unmapped O-GlcNAc Sites on β-catenin, and O-GlcNAc Stabilizes GFP-β-catenin by Inhibiting Its Interaction With β-TrCP. Related to Figure 3

(A), (C), (E) and (G) IP-WB on cells expressing GFP-β-catenin and the indicated aptamers.

In (C) and (G), cells were treated with 50 μM Ac5S for 20 hr. N=3.

(B), (D), (F) and (H) WB on cells expressing GFP-β-catenin and the indicated aptamers. In (D) and (H), cells were treated with 50 μM Ac5S for 20 hr. In (F) and (H), both chemiluminescence and fluorescence WB were performed and they generated consistent results. Intensity of each band is normalized to that of β-tubulin. N=3.

(I) O-GlcNAc on the GFP-β-catenin^{S23A; T40A; T41A; T112A; or 4A} mutant was reduced by CpOGA. Lysate of HEK293T cells expressing GFP-β-catenin^4A was treated with CpOGA and subjected to IP-WB. N=3.

(J) Antibody recognition of O-GlcNAc on GFP-β-catenin^4A was competed by 1 M GlcNAc. IP-WB of cells expressing GFP-β-catenin^4A. 1 M GlcNAc was included in the antibody solution during WB for the membrane on the left. Lanes on the left and right are from two membranes but imaged together for equal exposure. Borders were added to the top blots. N=3.

(K) and (L) Quantifications of phosphorylation on endogenous β-catenin and abundance of proteins in Figures 3D (K) and 3E (L). Intensity of each band is normalized to that of β-tubulin. N=3.

T1 · AP3 – Individual T1 and AP3 aptamers (control). NL1F30 – DS aptamers (T1-linker-AP3) with flexible linkers of 30 nt. 3JB1F+6 – DS aptamers (T1-linker-bc339) with a folded linker. Quantitated data are normalized to the control groups (T1 · AP3) unless indicated. Student’s t-test. Data are represented as mean ± SD. ns, p ≥ 0.05; *, p < 0.05; **, p < 0.01; ***, p < 0.001.

4

Figure S4. DS Aptamers Increase O-GlcNAcylation on Endogenous β-catenin. Related to Figures 4, 5

(A) WB of whole cell lysate from Figure 4D and quantification. Intensity of each band in WB is normalized to that of β-tubulin. One-way ANOVA test, N=3.

(B) Mass-shift assays on HEK293T cells expressing the indicated aptamers. Two blots of β-catenin with short and long exposure time are shown. Intensity of the shifted bands is normalized to that of β-tubulin. N=3.

(C) and (D) WB on cells expressing the indicated aptamers, and treated with regular (C) or Wnt3A-conditioned (D) medium. Intensity of each β-catenin band is normalized to that of β-tubulin. N=3.

(E) and (F) WB on HEK293T cells expressing the indicated aptamers, and treated with regular (E) or Wnt3A-conditioned (F) medium. N=4.

(G) and (H) IP-WB of HEK293T cells expressing the indicated aptamers, treated with Ac5S and regular (G) or Wnt3A-conditioned (H) medium. N=3.

(I) Co-IP between β-catenin and OGT, on cells expressing the indicated aptamers. Intensity of each OGT band is normalized to that of β-catenin. N=3.

(J) Predicted tertiary structure of the DS aptamer 3JB8R+12. Nucleotides are color coded in the same way as its secondary structure prediction in Figure S1D.

(K) and (L) IP-WB on HEK293T cells expressing the indicated aptamers, and treated with regular (K) or Wnt3A-conditioned (L) medium. N=3.

(M) and (N) Co-IP between β-catenin and EZH2, on cells expressing the indicated aptamers, and treated with regular (M) or Wnt3A-conditioned (N) medium. Intensity of each EZH2 band is normalized to that of β-catenin. N=6.

T1 · bc339 – Individual T1 and bc339 aptamers (control). NL8F50/70/100 – DS aptamers (T1-linker-bc339) with flexible linkers of 50/70/100 nt. 3JB8F/3JB8R – DS aptamers (T1-linker-bc339) with folded linkers. +12, 12 bp additions into the folded linker. Quantitated data are normalized to the control groups (T1 · bc339). One-way ANOVA test in (A). Student’s t-test in other panels. Data are represented as mean ± SD. ns, p ≥ 0.05; *, p < 0.05, ***, p < 0.001.

5

Figure S5. O-GlcNAc Promotes β-catenin’s Interactions With KAT2A. Related to Figure 5

(A) to (H) Confocal images of Proximity Ligation Assay (PLA) on HEK293T cells expressing the indicated aptamers and stimulated with regular (−Wnt) or Wnt3A-conditioned (+Wnt) medium. PLA was performed with antibodies targeting β-catenin and KAT2A or H3K9ac. In (C) and (D), cells were treated with 50 μM Ac5S for 20 hr.

(I) to (P) Quantification of (A) to (H). PLA puncta in nucleus are counted and normalized to the number of nuclei in each image. Student’s t-test. N=10. Data are represented as mean ± SD. ns, p ≥ 0.05; ***, p < 0.001; ****, p < 0.0001.

T1 · bc339 – Individual T1 and bc339 aptamers (control). 3JB8F+12/R+12 – DS aptamers (T1-linker-bc339) with folded linkers.

6

Figure S6. O-GlcNAc Regulates β-catenin’s Interactions With EP300, KAT5 and β-TrCP. Related to Figure 5

(A) to (F) Confocal images of PLA on HEK293T cells expressing the indicated aptamers, and stimulated with regular (−Wnt) or Wnt3A-conditioned (+Wnt) medium. PLA was performed with antibodies targeting β-catenin and EP300, KAT5, or β-TrCP.

(G) and (H) Confocal images of the negative control experiments of PLA. Cells were treated with regular (−Wnt) or Wnt3A-conditioned (+Wnt) medium. PLA was performed with the antibody targeting β-catenin and normal rabbit IgG.

(I) to (P) Quantification of (A) to (H). Nuclear or total PLA puncta are counted and normalized to the number of nuclei in each image. Student’s t-test. N=10 (I to N) or 5 (O to P). Data are represented as mean ± SD. **, p < 0.01; ***, p < 0.001; ****, p < 0.0001.

T1 · bc339 – Individual T1 and bc339 aptamers (control). 3JB8F+12 – A DS aptamer (T1-linker-bc339) with a folded linker.

7

Figure S7. O-GlcNAc on β-catenin Increases Its Transcriptional Activity and Recruits EZH2 to Promoters. Related to Figure 6

(A) to (B) TOPFlash luciferase reporter assay on HEK293T cells expressing the indicated aptamers, and stimulated with regular (−Wnt) or Wnt3A-conditioned (+Wnt) medium. In (B), cells were treated with 50 μM Ac5S for 20 hr. N=3.

(C) TOPFlash assay on cells expressing a scrambled RNA (scrambled), a shRNA targeting OGT (shOGT), or the shRNA with a codon-optimized OGT (shOGT + OGT). N=6.

(D) TOPFlash assay on cells treated with vehicle (DMSO), an inhibitor of OGT (Ac5S) or an inhibitor of OGA (TMG). N=3.

(E) TOPFlash assay on cells expressing the indicated RNA. N=3.

(F) and (G) Distribution of the differential binding sites of EZH2 on the genome, under −Wnt (A) and +Wnt (B) conditions.

(H) and (I) The binding peaks of EZH2 on the three promoters shown in Figure 6C under the indicated conditions. N=3.

(J) Numbers of the binding peaks of EZH2 on the genome under the indicated conditions. Analysis is performed on the data presented in Figures 6C and S7H. N=3.

(K) WB on HEK293T cells that are WT, or stably expressing a shRNA targeting β-catenin or EZH2. Intensity of each band is normalized to that of β-tubulin. Irrelevant lanes are cropped. N=2.

T1 · bc339 – Individual T1 and bc339 aptamers (control). 3JB8F+12 – A DS aptamer (T1-linker-bc339) with a folded linker. Two-way ANOVA test in (A) to (E) and (J). Data are represented as mean ± SD. ns, p ≥ 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001.

9

Table S3. Transfections in TOPFlash Assay. Related to Figure S7

10

Table S4. Differential Binding Sites of EZH2 Identified in CUT&RUN. Related to Figure 6

11

Table S5. Differentially Expressed Genes Identified in RNA-Seq. Related to Figure 6

12

Table S6. DNA Oligos for SELEX and TaqMan Assays. Related to STAR Methods and Key Resources Table

13

Table S7. Aptamer Sequences. Related to Key Resources Table

Highlights.

• Dual-specificity RNA aptamers regulates O-GlcNAcylation on individual proteins

• O-GlcNAc stabilizes β-catenin by inhibiting its interaction with β-TrCP

• O-GlcNAc on β-catenin recruits EZH2 to promoters and alters transcription

• Dual-specificity aptamers can be controlled by riboswitches or a Tet-On system