Abstract
N6-methyladenosine (m6A) is a highly prevalent modification found in mammal mRNA molecules that plays a crucial role in the regulation of cellular function. m6A RNA immunoprecipitation sequencing (MeRIP-seq) has been frequently used in transcriptomics research to identify the location of m6A. MABE572 (Millipore) is the most widely utilized and efficient anti-m6A antibody for MeRIP-seq. However, due to the high dose and price of this antibody, which has also been taken off the market, we discovered that CST's anti-m6A antibody can be used instead of MABE572 to map the m6A transcriptome. In the present study, we performed different concentrations of the CST anti-m6A antibodies with the corresponding initiation RNA of HEK293T cells, 2.5 μg antibody with 1 μg total RNA, 1.25 μg antibody with 0.5 μg total RNA, and 1.25 μg antibody with 0.1 μg total RNA. By comparing the m6A peak calling, enriched motifs, alternative splicing events, and nuclear transcripts modified by m6A between the CST and Millipore libraries, it was found that the CST library presented similar data to Millipore, even at incredibly low doses. The volume and cost of antibodies are significantly reduced by this refined MeRIP-seq using CST antibody, making it convenient to map future large-scale sample m6A methylation.
Introduction
RNAs are the molecules that perform numerous important functions within cell to control cellular processes such as gene expression, gene post-transcriptional regulation, and gene silencing. More than 170 chemical modifications that post-transcriptionally embellish RNAs have been identified so far [1]. Various modifications can take place on distinct RNA molecules, including messenger RNA (mRNA), transfer RNA (tRNA), and ribosomal RNA (rRNA). N6-methyladenosine (m6A) is a widely prevalent and extensively researched mRNA modification in mammals [2,3]. This modification is reversible under the regulation of ‘writers’ and ‘erasers’ [4–6] and can be recognized by ‘readers’ [7–9]. The m6A modification plays a critical role in gene post-transcriptional regulation, such as RNA processing [10], splicing [11,12], stability [13], and translation [14]. Most m6A modification regulators are expressed in the nucleus, and some of them play an important role in regulating RNA splicing and chromatin state. For instance, HNRNPC, serving as a m6A reader, can bind the splicing silencer sequence on pre-mRNA to inhibit splicing events [15]. Another nuclear reader, YTHDC1, targeted at carRNAs modified by METTL3, has been reported to play a role in the decay of carRNAs and chromatin opening [16].
m6A RNA immunoprecipitation sequencing (MeRIP-seq) is a widely used technique in transcriptomics research that enables the identification and quantification of RNA modifications [17,18]. Specifically, it enriches RNA fragments modified by m6A through binding specific antibodies to the RRACH motif (R = G or A; H = A, C, or U), which can then be sequenced to identify the sites of m6A modification. MeRIP-seq makes it possible to map the m6A of various organisms, such as zebrafish embryonic development [19], drosophila neuronal development [20], and mouse and human tissues [21,22]. Mapping of m6A requires a large amount of total RNA. The required minimum amount of starting RNA is 50 ng with the anti-m6A antibody from Millipore (MABE572), revealing the m6A landscape of the transcriptome during maternal-to-zygotic transition (MZT). This antibody has been shown to be more sensitive than other commercial antibodies and is widely used for m6A profiling in both human and murine tissues. However, the MABE572 antibody requires large quantities, meaning high cost, thus limiting its application to thousands of clinical samples. Moreover, the antibody is basically in stop production, and it has not yet been proven whether the officially produced MABE572I has functional substitutability. There is a need for a low-cost, highly sensitive antibody to detect m6A modification.
In the present work, we have tried to evaluate a new commercial antibody from CST (Cell Signaling Technology) by designing different input RNAs that correspond to different antibody dosage levels. We analyze the quantity, abundance, and enriched RNA production of m6A detected peaks, as well as RNA splicing events and nuclear transcripts modified by m6A. Our findings suggest that a low concentration of CST antibody can obtain similar data compared with the Millipore library when injected with 1 μg RNA samples. Furthermore, by reducing the dose of CST antibodies and corresponding RNA initiation levels, we observed a decrease in the number of peaks. However, there was almost no difference in the amount of enriched RNA. This is particularly significant as it demonstrates that CST antibodies can be used efficiently and cost-effectively in RNA methylation studies, even at lower concentrations.
Materials and methods
Cell culture
HEK293T (Human embryonic kidney) cells were cultured in DMEM basic (ThermoFisher, gibco, C11995500BT) with 10% FBS (Gibco, Australia, 10099141), 1% Penicillin/Streptomycin/Amphotericin B Solution (Sterile) (Sangon Biotech, B540733-0010), and 10 μg/ml Ciprofloxacin (YEASEN, 60201ES05) at 37°C with 5% CO2. Cell lines have not been authenticated or tested for mycoplasma.
RNA isolation and purification
Total RNA was extracted from cells using RNAiso Plus (Takara, 9109) following the manufacturer’s instructions. The RNA samples were assessed with 28S/18S ratios > 2 to ensure good quality. To eliminate DNA contamination, Turbo DNase (Invitrogen, AM2239) treatment was implemented by incubating the RNA samples at 37°C for 30 min. Then, the RNA was purified by adding 1/10 volume of 3 M NaAc, 2 μl of glycogen (5 mg/ml, Invitrogen, AM9510), and an equal volume of isopropanol. The mixture was then incubated at −80°C overnight. The RNA was centrifuged for 15 min at full speed, and the RNA pellet was washed twice with 75% ethanol. Finally, the RNA pellet was resolved in RNase-free water. A Qubit RNA HS Assay Kit (Thermo Fisher Scientific, Q32855) was employed to measure the RNA concentration.
MeRIP
The procedure for MeRIP sequencing was performed according to the published low-input m6A-seq protocol with slight modifications [23]. Each group of 15, 1, 0.5, and 0.1 μg total RNA was fragmented into approximately 200 nt-long fragments using RNA Fragmentation Reagents (Thermo Fisher Scientific, AM8740) at 70°C for approximately 5 min in a preheated thermal cycler (Eastwin, ETC821). Then, 2 μl of stop solution was added and mixed (Thermo Fisher Scientific, AM8740). The fragmented RNA was purified by adding 1/10 volume of 3M NaAc, 2 μl of glycogen and, an equal volume of isopropanol, as mentioned above.
Approximately 10 ng of purified RNA was kept as input, and the remaining RNA was prepared for m6A-MeRIP. Approximately 30 μl protein G magnetic beads (Thermo Fisher Scientific, 10004D) and 30 μl protein A magnetic beads (Thermo Fisher Scientific, 10002D) were mixed and washed twice with IP buffer (10 mM pH 7.5 Tris-HCl, 150 mM NaCl, and 0.1% IGEPAL CA-630). Then, the mixed beads were resuspended in 200 μl IP buffer. The anti-m6A antibodies were added to the resuspended beads, and the mixture was rotated at 4°C overnight. Two anti-m6A antibodies were used in the present study. One is 5 µg Millipore anti-m6A antibody (Millipore, ABE572). The other is 2.5 µg/1.25 µg CST anti-m6A antibody (Cell Signaling Technology, #56593).
The remaining fragmented RNA was added into the pre-cleaning bead-antibody mixture at 4°C for approximately 2 h, followed by being washed twice with IP buffer, washed twice with low-salt IP buffer (10 mM pH 7.5 Tris-HCl, 50 mM NaCl, and 0.1% IGEPAL CA-630), and washed twice with high-salt IP buffer (10 mM pH 7.5 Tris-HCl, 500 mM NaCl, and 0.1% IGEPAL CA-630) for 10 min each at 4°C. After washing, the bound RNA was eluted by competition with 6.7 mM N6-methyladenosine in 200 μl IP buffer (Selleckchem, S3190). Immunoprecipitation RNA was purified with RNeasy MiniElute spin column (Qiagen, 74104), and 14 μl RNase-free H2O was used to elute the m6A RNA.
Library preparation
The input RNA (10 ng fragmented RNA) and total IP RNA were used as starting materials to construct libraries with SMARTer Stranded Total RNA-Seq Kit v2-Pico Input Mammalian (Takara-Clontech, 634488), according to the standard protocol. The input RNA underwent 14 PCR cycles, while the m6A RNA was subjected to 16 cycles. Sequencing was performed using Illumina NovaSeq 6000 PE150.
Peak intensity and the calculated overall methylation level
To calculate the methylation intensity of a corresponding region in MeRIP-seq, we use the following formula: (IP FPKM/Input FPKM), where Input FPKM is (counts of mapped fragments × 109) / (Length of peak × Total count of the mapped fragments) and IP FPKM is (Counts of mapped fragments × 109) / (Length of peak × Total count of the mapped m6A fragments). To identify the m6A methylation level of a gene, we use the following formula: , where mi is the methylation level, li is the length of the peak in the gene, n is the number of peaks located in the gene, and L is the longest transcript length of every gene.
MeRIP-seq data analysis
The m6A sequencing method involved the identification of m6A-enriched peaks from all groups using MACS software (2.2.7) [24]. The resulting narrow peak data from all libraries were aligned to the hg19 reference genome from UCSC using the ChIPseeker (1.26.2) package [25]. The m6A methylation level of genes was defined as the average value of the score value of corresponding m6A peaks. The package ‘Guitar’ [26] was utilized to analyze m6A RNA-treated genomic features. The annotation data were visualized by R software (version 4.0.3) and GraphPad Prism 8.0 software. The m6A peaks from Millipore and other groups were classified as unique and common based on whether they overlapped or not.
Quantification of long-read transcripts
To assess the expression of longread transcripts and genes with TPM (Transcripts Per Million) and FPKM (fragments per kilobase million), we used salmon (v0.60.0) to quantify the mapped reads of transcriptome using only GENCODE(GRCh37) annotation or GENCODE annotation augmented with long-read transcripts as reference.
Motif discovery
All peaks were chosen for m6A motif analysis with HOMER (v4.11) in each group [26]. We selected the first ‘RRACH’ motif from the obtained motifs of each group (Millipore, CST I, CST II, and CST III) and recorded the P value.
Unique and overlapping m6A peaks
For the intersections of m6A peaks among specific groups, we use BEDTools (version 2.29.2) with function ‘intersect’, with default parameter.
Identification and characterization of AS events
Seven types of alternative splicing (AS) events were identified using SUPPA2 (v2.2.1) in each stage, including skipping exons (SE), alternative 5′ or 3′ splice sites (A5/A3), retained introns (RI), mutually exclusive exons (MX), and alternative first or last exons (AF, AL). Then, the percent spliced in (PSI) values were calculated by SUPPA2 based on the TPM values of transcripts in each sample with psiPerEvent subcommand. AS genes were defined as genes associated with AS events. In each stage, we applied SUPPA2 to calculate the PSI of each AS event using RNA-seq data with the default parameters, and merged long-read transcripts were used as a reference.
Identification and characterization of paRNA and repeats RNA
Raw sequencing reads were trimmed by Trimmomatic to remove low-quality bases and adapters. The output was first subjected to Trim_galore (http://www.bioinformatics.babraham.ac.uk/projects/trim_galore/) for quality control and trimming adaptors. The quality threshold was set to 20, and the minimum length required for reads after trimming was 30 nt. All the m6A-seq and input raw data reads were first aligned against rRNA (hg38, downloaded from UCSC Genome Browser) using bowtie2 (version 2.4.1), with the unmapped reads kept for further analysis. The remaining reads were mapped to GENCODE human hg19 rmsk genome using STAR (version 2.7.8a) with default parameters. Then the results obtained from STAR were transformed into the bam format using BEDTools (version 2.29.2). m6A peaks were called using MACS2 (version 2.2.7) with the parameter ‘–nomodel’ separately, and significant peaks with q < 0.01 were considered. We constructed the UCSC Genome Browser hg19 annotation file by using ‘makeTxDbFromGFF’ function of the GenomicFeatures R package. The candidate peaks were assigned to the nearest genes by annotatePeak in ChIPseeker. The numerical count of reads in m6A peaks was calculated by featureCounts with parameters ‘-p -t exon -g gene_id’.
Results
Whole-transcriptome m6A profiling in Millipore and CST libraries
To investigate an optimized protocol for the MeRIP-seq analysis with CST anti-m6A antibody, we estimated the performance of different concentrations of the CST anti-m6A antibody in relation to the starting amount of RNA from HEK293T cells (CST I: 2.5 μg CST anti-m6A antibody with 1 μg total RNA, CST II: 1.25 μg CST anti-m6A antibody with 0.5 μg total RNA, CST III: 1.25 μg CST anti-m6A antibody with 0.1 μg total RNA) compared to the widely used Millipore anti-m6A antibody (Millipore: 5 μg Millipore anti-m6A antibody with 15 μg total RNA) (Figure 1A). The optimized MeRIP-seq used extensive high/low salt washing after incubation of the antibody-bead complex and RNA fragments, and then the SMARTer Stranded Total RNA-Seq Kit v2 (Pico Input Mammalian) was employed for library construction as previously reported [23] (Figure 1A). We first compared the number of m6A peaks and m6A-modified genes among the four groups. Millipore captured an obviously larger number of m6A peaks (30957) and genes (8314) compared with CST. In the other three groups, CST I identified the most m6A peaks (23188) and m6A genes (7108), while CST II detected slightly fewer m6A peaks (14842) and genes (6919), and CST III showed significant decreases in both (Figure 1B). In all groups, the m6A peaks were most abundant near the stop codon, 3′UTR, and internal coding sequence (CDS) (Supplementary Figure S1). This was similar to what other studies had found [27]. Then, we analyzed the distribution of differential m6A peaks in the transcriptome. The distribution of m6A peaks in the genome was quite similar between Millipore and CST I, while in CST II and CST III, the m6A peaks were enriched more in the 5′UTR, exon, and 3′UTR and less in the intron (Figure 1C). On average, there are 3-5 m6A peaks on each m6A-modified mRNA. Based on our analysis, most of the m6A genes exhibited one or two m6A peaks, and Millipore showed obvious advantages in the m6A gene number among all m6A peak groups (Figure 1D). The difference in the total number of genes between CST I and Millipore can be explained by whether or not m6A peak sites were found (Figure 1B,D). CST II exhibited a similar number of genes containing 1-3 m6A peaks as CST I. However, in the CST III group, the number of m6A genes identified was very small, whatever the group of m6A peaks (Figure 1D). As expected, samples clustered well according to m6A genes in all four groups. Additionally, both CST I and CST II were highly correlated with Millipore (Figure 1E). Transcriptome-wide m6A site mapping revealed the typical consensus sequence RRACH (R = G or A; H = A, C, or U). The top consensus motif in CST I and CST II, conforming to the m6A sequence feature, tended to be close to the motif enriched in Millipore (Figure 1F), while CST III was inconsistent. So we will no longer analyze CST III in the future.
Overview of the m6A from different anti-m6A libraries in HEK293T cells
(A) Schematic diagram of the procedure of m6A MeRIP and naming of each group with different m6A antibodies combined with different starting RNA amounts. (B) The total number of m6A peaks and m6A genes identified by each group. (C) The percentages of m6A peaks in different transcript segments: promoter, 5′UTR intron, exon, intron, 3′UTR, distal intergenic, and downstream regions. (D) The number of m6A genes that contain one or more m6A peaks (peak = 1, peak = 2, peak = 3, peak = 4, peak = 5, and peak ≥ 6) in each group. (E) Heat map of Pearson correlation coefficient between four antibody groups based on common m6A genes. (F) Top m6A motifs identified within m6A peaks in each group.
(A) Schematic diagram of the procedure of m6A MeRIP and naming of each group with different m6A antibodies combined with different starting RNA amounts. (B) The total number of m6A peaks and m6A genes identified by each group. (C) The percentages of m6A peaks in different transcript segments: promoter, 5′UTR intron, exon, intron, 3′UTR, distal intergenic, and downstream regions. (D) The number of m6A genes that contain one or more m6A peaks (peak = 1, peak = 2, peak = 3, peak = 4, peak = 5, and peak ≥ 6) in each group. (E) Heat map of Pearson correlation coefficient between four antibody groups based on common m6A genes. (F) Top m6A motifs identified within m6A peaks in each group.
These results indicated that the m6A peaks and genes identified in the four groups were highly trusted. Although Millipore has been shown to have great advantages with regular amounts of RNA, CST I and CST II appear to be reliable substitutes for MeRIP-seq with low input RNA.
m6A enrichment genes in CST II approach to Millipore
To compare the consistency between CST and millipore antibodies, we analyzed the obtained sequencing data from various aspects. First, we detected m6A peaks and m6A genes, those exhibited in CST I and CST II were highly overlapped when compared with Millipore (Figure 2A,D). Although CST I detected more m6A peaks (Figure 2A) than CST II, a similar fraction of unique m6A peaks in CST I and CST II was observed – approximately 20% of themselves (Supplementary Figure S2A). And CST II identified an even higher enrichment of total m6A peaks than CST I (Figure 2B). To further investigate the similarity and difference of m6A-modified genes in the three groups, a Venn diagram was shown (Figure 2C). We found that approximately 75% of m6A genes were common to all three groups. Compared with CST I, CST II showed greater m6A enrichment in both the common peaks and the peaks that overlapped with Millipore (Supplementary Figure S2B and S2C).
Enrichment of m6A peaks and genes in Millipore, CST I and CST II
(A) The total m6A peak number of Millipore, overlapped, and unique m6A peaks identified by CST I and CST II when compared to Millipore. (B) Relative enrichment of total m6A peaks in three groups. (C) Venn diagram showing the overlap of m6A genes in three groups. (D) Total m6A gene number of Millipore, overlapped and unique m6A genes identified by CST I and CST II when compared with Millipore. (E) The m6A methylation levels of total m6A genes in three groups. (F) The average RNA expression level of the unique m6A genes identified in CST I was significantly lower than that of m6A genes overlapped with CST II, ***P < 1 × 10−4. (G) Left: Density plot of overall m6A level of 100 randomly selected genes on random chromosomes in three groups. Right: Distribution of m6A genes on chromosomes in three groups.
(A) The total m6A peak number of Millipore, overlapped, and unique m6A peaks identified by CST I and CST II when compared to Millipore. (B) Relative enrichment of total m6A peaks in three groups. (C) Venn diagram showing the overlap of m6A genes in three groups. (D) Total m6A gene number of Millipore, overlapped and unique m6A genes identified by CST I and CST II when compared with Millipore. (E) The m6A methylation levels of total m6A genes in three groups. (F) The average RNA expression level of the unique m6A genes identified in CST I was significantly lower than that of m6A genes overlapped with CST II, ***P < 1 × 10−4. (G) Left: Density plot of overall m6A level of 100 randomly selected genes on random chromosomes in three groups. Right: Distribution of m6A genes on chromosomes in three groups.
In terms of genes overlapped with Millipore, the number of unique m6A-modified genes in CST II was slightly higher than that in CST I (Figure 2D and Supplementary Figure S2D). In addition, we evaluated the m6A levels of overall genes with m6A modification and the overlapped 5419 m6A genes in the three groups and found that CST II was more similar to Millipore (Figure 2E and Supplementary Figure S2E). We focused more on the correlation of common m6A genes among the three groups, and the results indicated that the majority of the m6A levels of every single m6A gene in CST II were closer to Millipore than in CST I (Supplementary Figure S2F). This may explain the reason why CST II targeted fewer m6A genes but identified a higher m6A level than CST I (Figure 2A,B). Compared witg CST I, the m6A levels of genes that overlapped with Millipore in CST II were much more similar to those of Millipore (Supplementary Figure S2G). Then, we grouped m6A genes by m6A levels. The top 30% of genes with the highest m6A methylation level were defined as high m6A level genes, the bottle 30% of genes were defined as low m6A level genes, and the rest 40% were defined as middle m6A level genes. Similar performance was also observed in m6A genes that overlapped with Millipore in both CST I and CST II, with CST II showing characteristics similar to those of Millipore (Supplementary Figure S2H). We next calculated the average expression level of the overlapped m6A genes between CST I and CST II with unique m6A genes in CST I. The results showed that the expression of unique m6A genes in CST I was relatively low, and these genes may not arouse close attention in reality (Figure 2F). Furthermore, we also calculated the number of m6A genes distributed on each chromosome in the three groups (Figure 2G, right). Briefly, Millipore identified the highest m6A genes on all chromosomes, while CST I detected more m6A genes than CST II, except on chromosome 11, 16, 17, 21, and 22. In addition, we randomly extracted 100 m6A genes from each of the three groups, and the density plot consistently showed a stronger correlation between the m6A levels of CST II and Millipore (Figure 2G, left).
In summary, CST I may have a significant advantage in the number of m6A peaks and genes, but CST II tends to be closer to Millipore at the m6A level and density of genes.
CST I captures greater consistent alternative splicing events than CST II
m6A is a crucial modification involved in the processing of pre-mRNA. The m6A reader YTHDC1 has been directly proven to regulate the splicing of mRNA [28]. We investigated the correlation between m6A methylation and alternative splicing (AS) events using three anti-m6A groups: CST I, CST II, and Millipore. All three groups displayed similar total mRNA AS events using SUPPA2 [29] (Supplementary Figure S3A). However, we found that m6A modified roughly half of the spliced genes. For example, in the CST I group, 7,624 out of 14,605 AS genes underwent m6A modification (Figure 3A), highlighting the importance of m6A in splicing events. Moreover, the m6A-modified spliced genes in both CST I and CST II had a lot in common with those in Millipore (Figure 3B). We observed a concordance probability in m6A gene AS events among the three groups, with no significant difference (Figure 3C). Notably, Millipore exhibited a significant advantage in AS events with m6A modification, reaching 23,140 (Figure 3D). Although the number of AS events in CST I (20,173) was relatively smaller than in Millipore, the distribution of the seven AS types and the m6A methylation levels of the spliced genes in CST I were consistent with those of Millipore (Figure 3E,F), and this concordance was greater than that of CST II. Conversely, the AS event count of m6A genes in CST II was much smaller than in Millipore (Figure 3D), and the distribution of total mRNA AS was similar among all three antibody groups (Supplementary Figure S3B). Additionally, to further explore the role of m6A modification level in AS events, we grouped m6A genes based on their m6A levels, as mentioned previously. Similarly, CST I exhibited greater concordance with Millipore in the three m6A level groups (Supplementary Figure S3C). These results suggest that genes modified by m6A in CST I are more similar to those of Millipore in terms of the number of splicing events, splicing probability, and distribution of splicing types. Therefore, the use of CST I antibodies can, to some extent, replace Millipore antibodies.
Characterization of alternative splicing of m6A genes in CST I, CST II and Millipore
(A) The fraction of non-m6A and m6A modification in alternative splicing genes in three antibody groups. (B) Venn diagram showing the overlap of alternative splicing genes with m6A modification. (C) The overall level of alternative splicing in three groups. (D) The number of alternative splicing events with m6A modification in three groups. (E) The percentage of alternative first exon (AF), skipping exon (SE), alternative 3′splice site (A3), alternative 5′splice site (A5), alternative last exon (AL), retained intron (RI), and mutually exclusive exons (MX) with m6A modification in three groups. (F) The level of each alternative splicing type in three groups.
(A) The fraction of non-m6A and m6A modification in alternative splicing genes in three antibody groups. (B) Venn diagram showing the overlap of alternative splicing genes with m6A modification. (C) The overall level of alternative splicing in three groups. (D) The number of alternative splicing events with m6A modification in three groups. (E) The percentage of alternative first exon (AF), skipping exon (SE), alternative 3′splice site (A3), alternative 5′splice site (A5), alternative last exon (AL), retained intron (RI), and mutually exclusive exons (MX) with m6A modification in three groups. (F) The level of each alternative splicing type in three groups.
CST II marks greater consistent m6A -modified nuclear transcripts than CST I
Chromatin-associated regulatory RNAs (carRNAs) are RNAs that associate with chromatin directly or indirectly. They have been found to be involved in gene and transcriptional regulation through multiple mechanisms and have important roles in different types of cancer [30]. Approximately 15–30% of carRNAs contain m6A in mESCs. m6A facilitates the degradation of carRNAs by regulating their stability and chromatin state [16,31]. In addition to mRNA, we also detected carRNAs in HEK293T cells, including promoter-associated RNA (paRNA) and RNA transcribed from transposable elements (repeats RNA). In the three groups, the overwhelming majority of the m6A-modified transcripts were mRNA, and promoter associated RNA (paRNA) as well as repeats only constituted a small percentage, approximately 8% (Figure 4A). Millipore still showed a superior ability to detect paRNA and Repeats, while CST II was only a little inferior to CST I (Figure 4B,C). The Venn diagram shows a great diversity of paRNA, but repeats were largely overlapped among the three groups. We further estimated the m6A level of paRNA and Repeats in each group. It turned out that the m6A levels of paRNA and Repeats in CST II were closer than those in CST I to those in Millipore (Figure 4D). Whatever the paRNA and Repeats in common across the three groups, and whatever the paRNA and repeats in CST I or CST II overlapped with Millipore, the results were also observed (Supplementary Figure S4A and S4B). Then, we analyzed the distribution of different repeat types with m6A modifications among the three groups. Both CST I and CST II enriched more LINE, CST I targeted more simple repeats than CST II (Figure 4E). But other than this, the percentage distribution of CST II resembled that of Millipore in general (Figure 4E). We further investigated the m6A level of each repeat type and found that the m6A levels of CST II were higher than CST I except for simple repeat (Figure 4F). These results indicate that CST II is more similar to Millipore in terms of the numbers and m6A levels of different carRNAs.
Characterization of paRNA and Repeats RNA m6A-modified transcripts in the nucleus of CST I and CST II
(A) The fraction of mRNA, paRNA and Repeats with m6A modification in three groups. (B) Venn diagram showing the overlap of m6A paRNA in three groups. (C) Venn diagram showing the overlap of m6A Repeats in three groups. (D) The overall methylation levels of paRNA and Repeats in three groups. (E) The percentages of different Repeat types with m6A modification: SINE, LINE, LTR, DNA repeat elements, simple repeat, Satellite, Retroposon and other types. (F) The methylation levels of different Repeat types with m6A modification: SINE, LINE, LTR, DNA repeat elements, simple repeat, Satellite, Retroposon and other types in three groups.
(A) The fraction of mRNA, paRNA and Repeats with m6A modification in three groups. (B) Venn diagram showing the overlap of m6A paRNA in three groups. (C) Venn diagram showing the overlap of m6A Repeats in three groups. (D) The overall methylation levels of paRNA and Repeats in three groups. (E) The percentages of different Repeat types with m6A modification: SINE, LINE, LTR, DNA repeat elements, simple repeat, Satellite, Retroposon and other types. (F) The methylation levels of different Repeat types with m6A modification: SINE, LINE, LTR, DNA repeat elements, simple repeat, Satellite, Retroposon and other types in three groups.
Discussion
MABE572, which was Millipore’s most commonly used anti-m6A antibody in MeRIP-seq to detect m6A modifications, has stopped being made. In the present study, we developed an anti-m6A antibody from CST for m6A mapping by comparing the transcripts, alternative splicing events, and the carRNAs in the nucleus that m6A marked, with the data from Millipore. Importantly, the m6A-marked peaks and genes of CST are the same as those of Millipore with less antibody concentration and less RNA. This saves both money and samples.
Each mRNA contains three m6A sites, but most m6A-marked mRNAs have only one m6A peak. Our results confirmed this observation. Whatever the three libraries of CST or Millipore, the number of genes with one m6A site was the highest. We found that even though CST II had fewer total m6A peaks than CST I, the number of m6A genes with a single m6A peak was higher in CST II than in CST I. This meant that the total number of m6A-modified genes was the same for both. We speculate that because CST is a high-titer antibody, it can lead to the superabundant binding of antibodies to protein A/G. The many effective binding sites for m6A of crowded antibodies may be embedded, making it very difficult for RNA containing m6A to approach and incorporate with the binding site of antibodies. In contrast, a low concentration of CST creates a suitable proportion of antibody-bead complexes to capture m6A RNA fragments under optimized conditions. This may explain why CST II was able to identify almost an equal amount of enriched RNA as CST I by reducing the use of CST antibodies and corresponding RNA initiation levels. When there is an excessive amount of CST antibody that can't effectively bind modification sites in CST III, the ratio of antibodies to input RNA can affect the effectiveness of IP.
Previous research has proposed that m6A-dependent RNA modification is important for alternative splicing regulation [32,33]. Zhao et al. showed the function of FTO, a m6A ‘eraser’, in the mediation of mRNA splicing and gene expression with transcriptome analyses of m6A-seq [12]. The exon junction complex (EJC) is recruited by spliceosome to inhibit METTL3-mediated m6A modification [34]. We also examined the efficiency of the CST antibody when it comes to alternative splicing with our MeRIP-seq. We observed that m6A-modified MX events were almost the most common (Figure 3E), but they were the least common in total mRNA (Figure 3B) out of the seven different types of alternative splicing. Mutually exclusive splicing is a specific type of alternative splicing in which only one of two or more reduplicated exons is spliced into the mature mRNA isoform. MXs on Dscam contribute to 38016 isoforms in Drosophila melanogaster [35]. MXs have been reported to be associated with protein modification functions [36]. The various isoforms produced maintain diversity in biological functions, and transcripts produced by different MXs exhibit tissue specificity. In both CST and Millipore libraries, there is an abundance of m6A on MX events, indicating that this is not a random occurrence. This suggests that m6A modification may influence protein functions by modifying mutually exclusive splicing, which may be related to the development of diseases and play a role in biological evolution.
This study has some limitations. We did not carry out MeRIP-seq in more types of cells or tissues. Subsequent studies of more abundant sample types and larger sample sizes are needed to demonstrate our findings. On the other hand, whether CST antibodies have a tissue preference, other commercial antibodies and tissues from different species should also be included in the study for comparison.
Conclusion
In conclusion, we utilized three distinct methodologies, each utilizing varying concentrations of the CST antibody with initial total RNA from HEK293T cell lines, to cater to the diverse information requirements captured by MeRIP-seq. In terms of comparing m6A peak calling, m6A-genes, and transcripts methylated in the nucleus with Millipore libraries, CST I exhibits superior performance in splicing events. Conversely, CST II proves to be more effective when assessing the total number of m6A-genes and carRNAs. Both libraries ensure data correspondence with Millipore, even at low concentrations of m6A-antibody and with minimal input RNA. Our study not only conserves cost resources but also facilitates the mapping of m6A methylation in large-scale future samples.
Data Availability
All data are available from the corresponding authors upon reasonable request.
Competing Interests
The authors declare that there are no competing interests associated with the manuscript.
Funding
This research was funded by the National Key Project of Research and Development Program [grant number 2021YFC2700602]; the National Natural Science Foundation of China grant numbers 92168104 and 82071720]; Jiangsu Province 333 Talent Grant [grant number 2022-3-2-155]; Suzhou Talent Training Program [grant numbers GSWS2019005 and GSWS2020057]; and Jiangsu Provincial Medical Key Discipline (Laboratory) Cultivation Unit [grant numberJSDW202214].
CRediT Author Contribution
Wenjuan Xia: Writing—original draft, Writing—review & editing. Ling Guo: Investigation, Methodology. Huapeng Su: Investigation, Methodology. Jincheng Li: Software, Visualization. Jiafeng Lu: Conceptualization. Hong Li: Conceptualization, Funding acquisition. Boxian Huang: Conceptualization, Funding acquisition.
Abbreviations
References
Author notes
These authors contributed equally to this work.