Programmable protein expression using a genetically encoded m6A sensor

Abstract

The N⁶-methyladenosine (m⁶A) modification is found in thousands of cellular mRNAs and is a critical regulator of gene expression and cellular physiology. m⁶A dysregulation contributes to several human diseases, and the m⁶A methyltransferase machinery has emerged as a promising therapeutic target. However, current methods for studying m⁶A require RNA isolation and do not provide a real-time readout of mRNA methylation in living cells. Here we present a genetically encoded m⁶A sensor (GEMS) technology, which couples a fluorescent signal with cellular mRNA methylation. GEMS detects changes in m⁶A caused by pharmacological inhibition of the m⁶A methyltransferase, giving it potential utility for drug discovery efforts. Additionally, GEMS can be programmed to achieve m⁶A-dependent delivery of custom protein payloads in cells. Thus, GEMS is a versatile platform for m⁶A sensing that provides both a simple readout for m⁶A methylation and a system for m⁶A-coupled protein expression.

Figure 1:. Overview of the GEMS system.

Schematic illustrating the GEMS system for sensing m⁶A methylation. The m⁶A reporter mRNA is shown with a region of the m⁶A sensor sequence expanded. When this sequence is unmethylated, APO1-YTH does not bind to the sensor sequence and no editing takes place. As a result, EGFP-DHFR is produced and rapidly degraded (left). Methylation of either adenosine (red) in the sensor sequence results in recruitment of APO1-YTH and C-to-U editing of either or both convertible stop codons. Translation leads to EGFP production and cellular fluorescence (right).

Figure 2:. GEMS depends on m⁶A recognition.

a. HEK293T cells were transfected with the GEMS reporter mRNA alone or the reporter mRNA together with APO1-YTH and imaged 24 h later. Cells expressing the reporter mRNA with APO1-YTH exhibit robust EGFP fluorescence, whereas cells expressing the reporter mRNA alone are dark. Scale bar: 100 μm. b. RT-PCR and Sanger sequencing of the m⁶A sensor sequence from cells in (a) shows C-to-U editing (C-to-T in cDNA) of the convertible stop codon cytidines (marked with an asterisk) only in the presence of APO1-YTH. Quantification of %C2U at the indicated cytidines is shown below the sequencing traces. ***p < 0.001 by two-sided Student’s t-testby two-sided Student’s are presented as mean values ± SEM. t-test, n = 3 biological replicates. All data are presented as mean values ± SEM. c. Western blot from cells in (a) indicates EGFP production from the GEMS reporter mRNA only in cells expressing APO1-YTH. d. HEK293T cells were transfected with the GEMS system containing APO1-YTH or m⁶A binding-deficient APO1-YTHmut and imaged after 24 h. Cells expressing APO1-YTH contain EGFP fluorescence whereas cells expressing APO1-YTHmut are dark. Scale bar: 100 μm. e. Western blot from cells in (d) indicates loss of EGFP production from the GEMS reporter mRNA when cells co-express APO1-YTHmut. f. RT-PCR and Sanger sequencing from cells in (d) shows C-to-U editing of the m⁶A sensor sequence only in cells co-expressing APO1-YTH. Quantification of %C2U at the indicated cytidines is shown below the sequencing traces. ***p < 0.001 by two-sided Student’s t-test, n = 3 biological replicates. All data are presented as mean values ± SEM. g. Relative m⁶A quantification using an RT-qPCR m⁶A detection approach shows similar levels of methylation of the m⁶A sensor sequence (GEMS GAC) and endogenous ACTB site A1222 on which the sensor sequence is based. m⁶A is not detected at non-consensus adenosines in the m⁶A sensor sequence (UAC and CAG). Dotted line at 0.5 represents the minimum cutoff value indicating the presence of m⁶A. Schematic below shows the presence of the indicated adenosines within the m⁶A sensor sequence. ***p < 0.001 by two-sided Student’s t-test, n = 3 biological replicates. All data are presented as mean values ± SEM.

Figure 3:. GEMS is METTL3-dependent.

a. The GEMS system was transfected into HEK293T cells containing an auxin-inducible degron tag fused to endogenous METTL3. Addition of auxin leads to reduced EGFP fluorescence. Scale bar: 100 μm. b. Western blot confirms loss of METTL3 and EGFP following auxin treatment of cells in (a). c. Quantification of EGFP/EGFP-DHFR ratio from western blot samples in (b) shows a decrease in EGFP/EGFP-DHFR in response to auxin treatment. ***p < .001 by two-sided Student’s t-test, n = 5 biological replicates. All data are presented as mean values ± SEM. d. Sanger sequencing traces show C-to-U editing of the m⁶A sensor sequence in RNA extracts from cells in (a). Editing is reduced in auxin-treated cells. ***p < 0.001 by two-sided Student’s t-test, n = 3 biological replicates e. RT-qPCR quantification of m⁶A sensor sequence methylation decreased m⁶A after auxin-mediated METTL3 depletion. ***p < 0.001 by two-sided Student’s t-test, n = 3 biological replicates. All data are presented as mean values ± SEM. f. The GEMS system was transfected into HEK293T cells in the presence or absence of METTL3 overexpression. EGFP fluorescence is increased in METTL3-overexpressing cells. Scale bar: 100 μm. g. Western blot analysis shows an increase in EGFP protein expression in cells overexpressing METTL3. h. Quantification of EGFP/EGFP-DHFR ratio from western blot data indicates increased EGFP/EGFP-DHFR in METTL3-overexpressing cells. *p = 0.02 by two-sided Student’s t-test, n = 3 biological replicates. All data are presented as mean values ± SEM. i. Sanger sequencing analysis of RNA extracts from cells in (f) shows increased C-to-U editing of the m⁶A sensor sequence in response to METTL3 overexpression. Stop codon 1: **p = 0.007, Stop codon 2: **p = 0.009 by two-sided Student’s t-test, n = 3 biological replicates. All data are presented as mean values ± SEM. j. Correlation between the change in m⁶A level in the reporter mRNA and EGFP protein production (EGFP/EGFP-DHFR) in response to METTL3 overexpression. Line represents best-fit and shaded area represents 95% confidence interval around the fit. Pearson’s r = 0.971, ***p < 0.001 by two-sided Student’s t-test, n = 4 biological replicates.

Figure 4:. GEMS detects differences in methylation across cell types.

a. The GEMS system containing an internal m⁶A-independent DsRed reporter was transfected into HEK293T, HeLa, and Huh-7 cells followed by fluorescence microscopy 24 h later. m⁶A-coupled EGFP fluorescence is reduced in Huh-7 cells compared to HEK293T and HeLa cells. Scale bar: 100 μm. b. Western blot analysis of cells in (a) shows decreased EGFP expression in Huh-7 cells compared to HEK293T and HeLa cells. c. Quantification of EGFP/EGFP-DHFR ratio relative to DsRed expression in HEK293T, HeLa, and Huh-7 cells. n.s.= no significant difference, ***p < 0.001 by two-sided Student’s t-test, n = 3 biological replicates. All data are presented as mean values ± SEM. d. Sanger sequencing shows C-to-U editing of the m⁶A sensor sequence in RNA samples from cells in (a). Huh-7 cells have reduced C-to-U editing compared to HEK293T and HeLa cells. Quantification of %C-to-U is shown on the right. ***p < 0.001 by two-sided Student’s t-test, n = 3 biological replicates. All data are presented as mean values ± SEM. e. Mass spectrometry was used to quantify m⁶A in purified mRNA from HEK29T, HeLa, and Huh-7 cells. n = 2 biological replicates per cell line. All data are presented as mean values.

Figure 5:. GEMS senses changes in m⁶A caused by small molecule inhibition of METTL3.

a. EGFP fluorescence from the GEMS system is reduced in HEK293T cells treated with the METTL3 inhibitor STM2457. GEMS-expressing cells were treated with 30 μM STM2457 for 24 h. Scale bar: 100 μm. b. Quantitative microscopy was performed on HEK293T cells expressing the GEMS system and treated with 30 μM STM2457. Treatment with STM2457 shows a significant reduction in EGFP fluorescence intensity. ***p < 0.001 by two-sided Student’s t-test, Single dot represents individual cells (DMSO: n = 777 cells; STM2457: n = 796 cells). Boxes span the interquartile range (25th to 75th percentile); horizontal lines indicate the median (50th percentile); whiskers extend to minima and maxima. EGFP signal in each cell was normalized to DsRed. c. Western blot shows decreased EGFP protein in STM2457-treated cells. d. Densitometry analysis indicates reduced EGFP/EGFP-DHFR ratio in cells treated with 30 μM STM2457 for 24 h. ***p <0.001 by two-sided Student’s t-test, n = 3 biological replicates. All data are presented as mean values ± SEM. e. Sanger sequencing of the m⁶A sensor sequence in RNA samples from cells in (a) shows reduced C-to-U editing in cells treated with STM2457. Quantification of %C-to-U is shown on the right. ***p <0.001 by two-sided Student’s t-test, n = 3 biological replicates. All data are presented as mean values ± SEM. f. RT-qPCR-based m⁶A detection was used to quantify relative m⁶A levels of endogenous ACTB A1222 and the m⁶A sensor sequence. STM2457 treatment leads to similar reductions in m⁶A in endogenous ACTB and the m⁶A sensor sequence. ***p < 0.001 by two-sided Student’s t-test, n = 3 biological replicates. All data are presented as mean values ± SEM.

Figure 6:. m⁶A-coupled effector protein delivery counteracts the effects of m⁶A hypermethylation in cancer cells.

a. Schematic showing m⁶A-coupled expression of a tumor suppressor protein to counteract the effects of m⁶A hypermethylation in cancer cells. b. GEMS was used to deliver either EGFP (GEMS-EGFP) or SOCS2 (GEMS-SOCS2) into Huh-7 cells. Western blot indicates robust expression of SOCS2 in cells expressing GEMS-SOCS2. c. RT-qPCR shows SOCS2 coding sequence expression in Huh-7 cells transfected with GEMS-SOCS2. ***p < 0.001 by two-sided Student’s t-test, n = 3 biological replicates. All data are presented as mean values ± SEM. d. Sanger sequencing of the m⁶A sensor sequence from cells in (b) indicates similar C-to-U editing rates of the GEMS-EGFP and GEMS-SOCS2 mRNAs. n.s. = no significant difference by two-sided Student’s t-test, n = 3 biological replicates. All data are presented as mean values ± SEM. e. Quantification of EGFP/EGFP-DHFR ratio and SOCS2/SOCS2-DHFR ratio from western blot data from cells expressing GEMS-EGFP or GEMS-SOCS2 indicates similar ratios. n.s. = no significant difference by two-sided Student’s t-test, n = 3 biological replicates. All data are presented as mean values ± SEM. f. Western blot analysis of downstream SOCS2 targets shows a decrease in STAT5 and JAK2 phosphorylation in Huh-7 cells expressing GEMS-SOCS2. g. RT-qPCR shows reduced expression of SOCS2 target mRNAs IGF1 and CyclinD1 in Huh-7 cells expressing GEMS-SOCS2. ***p < 0.001 by two-sided Student’s t-test, n = 3 biological replicates. All data are presented as mean values ± SEM. h. Cell growth assays show reduced growth of Huh-7 cells transfected with GEMS-SOCS2 compared to non-transfected cells (Control). Growth curves for both were normalized to cells expressing GEMS-EGFP. ***p < 0.001 by two-sided Student’s t-test, n = 3 biological replicates. All data are presented as mean values ± SEM. i. Huh-7 cell migration is diminished following expression of GEMS-SOCS2 compared to GEMS-EGFP. ***p < 0.001 by two-sided Student’s t-test, n = 3 biological replicates. All data are presented as mean values ± SEM. Scale bar: 100 μm. j. Western blot shows elevated p53 levels in Huh-7 cells expressing GEMS-p53 compared to GEMS-EGFP. Top panel shows brightfield images of cells migration, bottom panel shows quantification of the total number of cells migrated. ***p < 0.001 by two-sided Student’s t-test, n = 3 biological replicates. k. Cell growth is reduced in Huh-7 cells transfected with GEMS-p53 compared to non-transfected cells (Control). Growth curves for both were normalized to cells expressing GEMS-EGFP. ***p < 0.001 by two-sided Student’s t-test, n = 3 biological replicates. All data are presented as mean values ± SEM.