Dynamic regulation of N6,2′-O-dimethyladenosine (m6Am) in obesity

Ben-Haim, Moshe Shay; Pinto, Yishay; Moshitch-Moshkovitz, Sharon; Hershkovitz, Vera; Kol, Nitzan; Diamant-Levi, Tammy; Beeri, Michal Schnaider; Amariglio, Ninette; Cohen, Haim Y.; Rechavi, Gideon

doi:10.1038/s41467-021-27421-2

Download PDF

Article
Open access
Published: 10 December 2021

Dynamic regulation of N⁶,2′-O-dimethyladenosine (m⁶Am) in obesity

Moshe Shay Ben-Haim ORCID: orcid.org/0000-0003-4737-9102^1,2,3,4^na1,
Yishay Pinto⁵^na1,
Sharon Moshitch-Moshkovitz^1,2,3,
Vera Hershkovitz^1,3,
Nitzan Kol^1,3,
Tammy Diamant-Levi^1,3,
Michal Schnaider Beeri^6,7,
Ninette Amariglio^1,3,5,
Haim Y. Cohen⁵ &
…
Gideon Rechavi^1,2,3

Nature Communications volume 12, Article number: 7185 (2021) Cite this article

4226 Accesses
20 Citations
3 Altmetric
Metrics details

Subjects

Abstract

The prevalent m⁶Am mRNA cap modification was recently identified as a valid target for removal by the human obesity gene FTO along with the previously established m⁶A mRNA modification. However, the deposition and dynamics of m⁶Am in regulating obesity are unknown. Here, we investigate the liver m⁶A/m methylomes in mice fed on a high fat Western-diet and in ob/ob mice. We find that FTO levels are elevated in fat mice, and that genes which lost m⁶Am marking under obesity are overly downregulated, including the two fatty-acid-binding proteins FABP2, and FABP5. Furthermore, the cellular perturbation of FTO correspondingly affect protein levels of its targets. Notably, generally m⁶Am- but not m⁶A-methylated genes, are found to be highly enriched in metabolic processes. Finally, we deplete all m⁶A background via Mettl3 knockout, and unequivocally uncover the association of m⁶Am methylation with increased mRNA stability, translation efficiency, and higher protein expression. Together, these results strongly implicate a dynamic role for m⁶Am in obesity-related translation regulation.

NADP modulates RNA m6A methylation and adipogenesis via enhancing FTO activity

Article 27 July 2020

m6A epitranscriptomic regulation of tissue homeostasis during primate aging

Article 06 April 2023

The major cap-binding protein eIF4E regulates lipid homeostasis and diet-induced obesity

Article 18 February 2021

Introduction

Obesity and overweight are a major concern given their increasing prevalence and associated health problems. Recently, variants within the FTO gene—an mRNA demethylating enzyme—were identified as robustly associated with human obesity¹. While most attention was given to the m⁶A mRNA modification as the first identified demethylated target of FTO², new work has demonstrated that FTO can also efficiently remove the mRNA m⁶Am cap modification³. However, the deposition and possible dynamic involvement of m⁶Am in obesity regulation and metabolism have not yet been characterized.

Various approaches have been employed to attempt explaining the robust association of FTO with obesity. Some have implicated m⁶A methylation in metabolic processes such as adipogenesis through control of mRNA splicing⁴; or dopaminergic functioning⁵; while others have found effects of neighboring genes adjacent to FTO as possible metabolic contributors^6,7. Despite the important progress made, a clear picture for the mechanism underlying this association remains elusive. In particular, it is unclear if and in what way does the newly discovered substrate of FTO, m⁶Am, regulates obesity. In order to explore the possible involvement, dynamics, and deposition of the two types of mRNA modifications mediated by FTO, we systematically explored the m⁶A and m⁶Am methylomes and their functions in obesity-associated metabolic perturbations in high fat Western diet-fed mice and in ob/ob overfeeding mice. In addition, in order to reliably differentiate m⁶Am from m⁶A (which are detected by the same antibody), we constructed a validated map of m⁶Am genes in a context completely free of m⁶A background via knockout (KO) of its methyltransferase, Mettl3. Doing so, we were able to reliably and un-confoundingly resolve the controversy concerning the m⁶Am functions (see a detailed description of this controversy below).

Though the existence of chemical modifications of RNA molecules has been known for decades, only recently the development of designated antibody-based capturing coupled with next generation sequencing, allowed to map these modifications at high throughput and to begin exploring their physiologic functions systematically (see for example^8,9; for reviews^10,11). Mapping the human and mouse m⁶A methylomes deciphered the non-random deposition of this modification alluding to its conserved functions^12,13. m⁶A is the result of methylation at the N⁶ position of the adenosine (Fig. 1a) in mRNA, generally installed at a RRACH RNA consensus sequence and particularly enriched around stop codons and the 3′ UTR, as well as long internal exons^12,13. The m⁶Am modification, on the other hand, lays at the first transcribed adenosine, adjacent to the cap 7-methyl Guanosine (m⁷G). It is usually methylated at the 2′-O position (Am), and can be further methylated at the N⁶ position to form an N⁶, 2′-O-dimethyladenosine (m⁶Am) (Fig. 1b). mRNA cap-m⁶Am is typically installed in a context of a BCA consensus sequence within the DNA (where B denotes G or C or T, and A is the methylated adenosine¹⁴). The extent of m⁶Am was reported to be 2–15 times higher than Am in most tissues¹⁵, making it a rather abundant cap modification.

**Fig. 1: m⁶Am genes are associated with higher translation efficiency, mRNA stability, and increased protein levels in m⁶A depleted mES *Mettl3* KO cells.**

Both m⁶A and m⁶Am are recognized by anti-m⁶A antibodies used in m⁶A-seq methods^12,16. This makes it particularly difficult to reliably differentiate between the two modified nucleotides, and especially since m⁶A is also present in the 5′ UTR. Using a single nucleotide resolution approach allowed unequivocal detection of only 24 sites localized exactly on the annotated transcription start site position (TSS) in one study¹⁴ or 46 in another³. This small number of sites detected using this approach is partially due to the difficulty in determining the specific site where transcription begins given the vast variability in promoter regions and transcription initiation^17,18,19. Finding a way to identify multiple m⁶Am peaks reliably is critical in order to confidently and reliably decipher its functions.

Several recent studies have attempted to explore the m⁶Am locations and functions, though some of the findings arising from these studies point to opposing conclusions. One study has implicated m⁶Am in regulating mRNA stability³, whereas another study did not find such an effect²⁰. The two studies also suggested m⁶Am may be associated with increased translation efficiency based on correlations with external ribosome profiling data sources. Recently, several independent groups have identified the methyltransferase which catalyzes m⁶Am installation—Pcif1, and knocked it out in order to explore the m⁶Am functions^{21,22,23,24,25}. In doing so, Akichika et al. findings suggested that m⁶Am promotes translation²². However, Sendinc et al. findings suggest that m⁶Am shows rather reduced translation²³; whilst Boulias et al., and Raman Pandey et al., suggest rather it has an effect on mRNA stability^24,25, along with an identified reduced body-weight in KO mice²⁵. A possible reason for the discrepancy between the observed m⁶Am functions in Pcif1 KO studies is that these methods relay on identifying missing m⁶A peaks in KO samples as the criterion for m⁶Am identification. Yet, missing m⁶A peaks could essentially still include some original m⁶A sites that are absent from other reasons—such as indirect effects associated with the bulk knockout stress, or even general quality variability between samples. While this procedure does provide clues for potential m⁶Am candidate sites, here, we identify a completely unconfounded pool of m⁶Am sites based on sites recognized by m⁶A-seq in mouse embryonic stem cells (mESCs) in which Mettl3, the main m⁶A methyltransferase—is KO⁸. Since these cells are completely depleted from m⁶A in their mRNA⁸, we are able to deduce true m⁶Am peaks, and thereby uncover reliably its association with increased mRNA stability, translation efficiency, and protein expression. Furthermore, our examination of m⁶A and m⁶Am profiles in HFD and ob/ob mice, uncover the dynamic role of the m⁶Am modification in regulating obesity, showing clear metabolic involvement of altered methylated genes, including two fatty acid-binding proteins Fabp2, Fabp5. In addition, we identify that FTO is overexpressed in fat mice, and importantly downregulates its m⁶Am gene targets. These results strongly implicate a significant role for m⁶Am in dynamic translation regulation of obesity.

Results

Identifying mRNA m⁶Am in m⁶A-depleted mESCs strongly establishes an association of m⁶Am with higher protein levels, translation efficiency, and mRNA stability

We deduced the m⁶Am profile from m⁶A sites mapped in Mettl3 KO mESCs devoid entirely of mRNA m⁶A⁸ and identified m⁶Am peaks within 1,848 genes, of which 77% were detected within a window of up to 50 nucleotides from a known annotated TSS (Fig. 1c and Supplementary Data 1). In these reliable m⁶Am peaks we identified a robust correlation between the peak score (see “Methods”) and the percentage of adenosines at the annotated TSS (Fig. 1d, Pearson r = 0.82). Yet, given the considerable variability of TSS locations^17,18,19, it is likely that in this m⁶A depleted sample non-adenosine annotated TSS peaks represent m⁶Am signals derived from an alternative A—TSS in its vicinity. Similarly, the few remaining peaks that were located >50 nucleotides of a known TSS, could still represent in part unidentified tissue-specific alternative TSS locations (as well as some background noise). As expected, collapsing all the DNA sequences of the peaks preceding and following known TSSs have identified the canonical BCA genomic consensus in addition to a bias towards pyrimidines in position −1 and a preceding TATA box, which are two hallmarks of transcription initiation¹⁷, (Fig. 1e).

Importantly, the use of cells devoid of m⁶A in mRNA enables the unequivocal identification of m⁶Am peaks and thus to reliably explore their functions. Testing the mRNA levels of m⁶Am-decorated genes in Mettl3 KO mESCs⁸ revealed that these tend to be substantially more expressed than non-m⁶Am genes (Fig. 1f), even after removing genes that were decorated with m⁶A outside the 5′ UTR segment in wild type (WT) mESCs, (Supplementary Fig. 1a). A similar pattern of expression of m⁶Am modified genes was evident in WT mESCs, contrasting the inverse U-shaped expression pattern of m⁶A modified genes (Supplementary Fig. 1b). Notably, analysis of our previously published ribosome profiling data of the very same m⁶A-depleted samples⁸ revealed that m⁶Am genes tend to have higher ribosome footprint coverage even after controlling for their corresponding mRNA expression levels (Fig. 1g, p = 2.02 × 10⁻³⁵, ANCOVA), as well as after excluding genes decorated with m⁶A in WT mESCs, (p = 0.046, ANCOVA), indicating higher translation efficiency. In addition, m⁶Am genes had higher ribosome release scores (RRS) that are associated with higher translation rates²⁶ (Fig. 1h, p = 1.47 × 10⁻⁷, Mann Whitney), even after excluding m⁶A genes identified in WT mESCs (p = 0.012, Mann Whitney). Most importantly, high throughput proteomic profiling analyses showed that these effects were translated to significantly higher protein levels of m⁶Am- compared to non-m⁶Am-modified genes (Fig. 1i, p = 4.32 × 10⁻³³, Mann Whitney) even after adjustment for their respective mRNA expression levels (Supplementary Fig. 1c, p = 6.19 × 10⁻¹¹, ANCOVA); or upon excluding m⁶A-modified genes identified in WT mESCs (p = 7.4 × 10⁻⁴, ANCOVA). An identical pattern of higher protein levels of m⁶Am-modified genes was observed in WT mESCs controls (Supplementary Fig. 1d, p = 1.39 × 10⁻⁶, ANCOVA), contrasting the opposite pattern of reduced protein levels of m⁶A-modified genes (Supplementary Fig. 1e, p = 6.17 × 10⁻⁸, ANCOVA). Finally, analysis of the mRNA half-life rates of these m⁶Am-modified transcripts established that they exhibit higher mRNA stability than non-m⁶Am-decorated genes (Fig. 1j, p = 5.77 × 10⁻³⁶, Mann Whitney); even after excluding m⁶A-methylated genes that were identified in WT mESCs (p = 6.9 × 10⁻¹¹, Mann Whitney).

Our approach examines well-annotated mouse m⁶Am sites, free from m⁶A background, and in doing so enables the clear correlation of m⁶Am with higher protein levels, translation efficiency, and mRNA stability. Yet, since the comparisons are made between m⁶Am-marked genes versus different genes that are not methylated, the findings described here, and in previous studies, are still correlative. In order to examine if m⁶Am has a role in promoting these observed effects, it is necessary to capture the effects triggered by dynamic changes in m⁶Am deposition within the same genes, which lose or gain m⁶Am methylation through experimental perturbations. For this purpose and in order to explore the possible metabolic effects and dynamics of m⁶Am in-vivo, we explored its profile and functions following high fat diet in a mouse model. However, given the fact that Mettl3 KO is embryonic lethal⁸ and normal mice mRNAs are also m⁶A methylated, we devised a more stringent approach for m⁶Am peak definition. Based on analysis of m⁶Am peak distribution in Mettl3 KO cells, we considered only read ends falling within 25 nucleotides of a reported TSS site, with over 5 reads in IP, over two-fold change (IP/Input) and an FDR ≤5%. Applying these more stringent criteria to the same Mettl3 KO mESCs mapping data⁸ resulted in the identification of robust peaks in 849 genes. The analysis of this highly stringent peak dataset recapitulated all the aforementioned results regarding association with mRNA stability, translation efficiency and mRNA and protein levels. Finally, to assess the false positive rate of our approach we reversed the IP with the Input samples. This control analysis resulted in only 3.5% (30) peaks, which indicates a very low level of false positive peaks.

m⁶Am- rather than m⁶A-decorated genes are enriched in metabolic processes annotation

In order to explore the m⁶A and m⁶Am deposition dynamics and functions in an obesity-related perturbation, we investigated the liver m⁶Am and m⁶A methylomes in response to high fat Western-diet (HFD) stress perturbation. Four weeks old mice were divided into two groups of 5 mice each, feeding on either standard chow diet or HFD for 17 weeks ad libitum. Mice fed on HFD gained significantly more weight (weighing a mean of 43 grams at the end of the experiment) than mice that were fed regular chow diet (mean of 30 grams at the end of the experiment, p = 0.008 (Fig. 2a). Consequently, the liver of HFD mice weighed more (mean of 2.19 grams) than that of standard chow diet mice (mean of 1.76 grams, p = 0.05, Supplementary Fig. 2a). In addition, an NMR analysis of the weight distribution indicated that the difference in weight was mainly in fat mass rather than in muscle mass (Fig. 2b). As expected, blood glucose levels of HFD mice after 6 h of fasting were higher (135 µg/dL) than control mice (109 µg/dL, p = 0.01, Supplementary Fig. 2b).

**Fig. 2: High fat diet mice physiologic parameters and enrichment of m⁶Am-modified genes in GO terms associated with metabolic processes.**

A high throughput analysis of m⁶Am and m⁶A peaks in lean vs. fat mice liver, detected m⁶Am peaks (identified in at least 2 of 3 mice in each group) in 1,176 genes in lean mice (calculated based on read ends falling within the first 25nt of reported TSS as described above) and m⁶Am peaks in 607 genes in fat mice (Supplementary Data 2). The majority of peaks (70%) were either lean or fat specific, while only 30% were common to both groups. An indicative example of an exclusively lean-specific m⁶Am peak within a reported alternative A-TSS of the Fabp2 gene is provided in Fig. 2c. We also detected m⁶A peaks in 2,727 genes in lean mice (starting > 50 nt of reported TSS, beyond 5’ UTR regions) and in 2,233 genes in fat mice (Supplementary Data 3). Notably, a significant percentage of m⁶A peaks (52%) were common to both lean and fat mice with the remaining 48% specific to either lean or fat mice. Importantly, collapsing the identified m⁶Am peaks preceding and following the annotated TSS within the DNA, identified the canonical m⁶Am BCA genomic consensus (Fig. 2d). In contrast, analyzing the mRNA sequences preceding and following non-5′UTR m⁶A peaks identified the canonical (RRACH) m⁶A consensus (Supplementary Fig. 3a). Analyzing the mRNA expression profiles demonstrated higher expression levels of m⁶Am modified genes, and revealed an inverse U-shaped expression profile of m⁶A modified genes (Supplementary Fig. 3b). In order to assess the dominant processes associated with m⁶Am or m⁶A methylated genes, we performed a gene ontology analysis. Although this analysis is very broad, we remarkably discovered that generally m⁶Am-decorated genes across lean or obese mice are predominantly enriched in multiple obesity and metabolic-related processes, while m⁶A genes did not seem to clearly display significant enrichment for metabolic processes (Fig. 2e; Supplementary Fig. 3c). Similar metabolic pathways were identified when we analyzed m⁶Am-modifed genes discovered in published Pcif1 KO studies (Supplementary Table 1)^22,23,24,25. Similarly, the lists of genes with differentially methylated peaks in lean vs. fat mice showed mainly m⁶Am but not m⁶A enrichment in metabolic and obesity-related processes (see Supplementary Data 2, 3). Hence, in subsequent analyses, we focused primarily on the effects observed in differentially m⁶Am-modified gene transcripts.

m⁶Am regulates mRNA and protein levels in response to HFD

In order to test the effects and functions of diet-related m⁶Am dynamics, we looked at genes with differential methylation which have gained or lost m⁶Am in HFD fed mice. We first looked at mRNA expression levels of differentially expressed genes in lean and fat mice by RNA-seq. This differentially expressed list of genes included only genes that either gained or lost m⁶Am, and that were also >1.5-fold differentially expressed, thereby focused on a total of 101 genes. We found that among genes that gained m⁶Am there were a significantly higher fraction of upregulated transcripts than expected, while genes that lost m⁶Am in fat mice had a significantly higher fraction of downregulated genes than expected (Fig. 3a,b, p = 0.004, χ2). In order to test if this effect is observed also at the protein level, we performed a high throughput proteomic profiling. Again, we found that m⁶Am-decorated genes are associated with higher protein levels than non-m⁶Am-modified genes (Fig. 3c). This association was in accordance with diet-related differential m⁶Am methylation patterns as well. Genes that gained m⁶Am in HFD fed mice displayed higher fraction of upregulated genes than expected, while genes that lost m⁶Am displayed a higher fraction of downregulated genes than expected (Fig. 3d, p = 0.014, χ2), suggesting a role for m⁶Am decoration in the regulation of gene expression. While the effect on the protein level presented here is likely augmented by the effect observed at the RNA level, removing all the above >1.5-fold differentially RNA expressed genes, preserves this pattern of results (Supplementary Fig. 4a). Similarly, we observed that m⁶Am-modified genes consistently display higher protein levels than non-m⁶Am-modified genes, even after controlling for their RNA expression level (Supplementary Fig. 4b). For example, two obesity-related fatty acid-binding proteins, FABP2 and FABP5, which lost their m⁶Am under HFD were both significantly downregulated at the protein levels (Fig. 3e and Supplementary Fig. 4c), while Fabp5 also displayed >10-fold downregulation at the mRNA level as well.

**Fig. 3: m⁶Am significantly regulates mRNA and protein levels upon high fat diet.**

FTO is upregulated in HFD and its cellular perturbation affects protein levels of m⁶Am targets

Next, we wished to test whether FTO has a preference towards demethylating m⁶Am rather than m⁶A as was recently reported³. We, therefore, overexpressed, as well as knocked-down FTO in human hepatocarcinoma cell line, HepG2 cells, and compared the number of m⁶A and m⁶Am peaks relative to control cells. Overall, we detected considerably fewer peaks in the overexpressing cells (522 m⁶Am and 3,803 m⁶A out of 4,325 peaks) than in the control cells (1,142 m⁶Am, and 6,316 m⁶A out of 7,458 peaks) and in the knockdown cells, respectively (1,672 m⁶Am, and 8,998 m⁶A out of 10,670 peaks, see Supplementary Data 4). The ratio of m⁶A/m⁶Am in control cells and knockdown cells was lower (5.5 fold, and 5.3 fold respectively), while in FTO overexpressing cells this ratio was significantly higher (7.3 fold, p = 5.35 × 10⁻⁹, χ2) (Fig. 4a), indicating that FTO affected both modifications with a slight preference towards demethylating m⁶Am. To further explore if differential m⁶Am methylation under FTO perturbation has a role in regulating protein levels, we performed high-throughput proteomic profiling. We observed that FTO targets which lost their m⁶Am peak under FTO overexpression were overly downregulated (Fig. 4b). Conversely, newly identified m⁶Am peaks under FTO knockdown showed rather increased prevalence of genes with protein levels higher than expected (Fig. 4c). In order to further assess if Fto has a specific role in HFD, we analyzed its expression in HFD fed mice. Western blot analysis of FTO in the liver of HFD fed mice showed that it was significantly overexpressed compared to the lean chow diet control mice (Fig. 4d, e), suggesting that the enzyme may have a dynamic regulatory role in demethylating m⁶Am (and m⁶A) under high fat diet.

**Fig. 4: FTO is highly expressed in HFD mice and its cellular perturbation affects protein levels of its m⁶Am-modified targets.**

Analysis of ob/ob mice suggests a similar regulation pattern of m⁶Am and demonstrates higher expression levels of FTO

In order to substantiate the regulatory effects of m⁶Am in yet another in vivo perturbation of obesity, we performed a corresponding high throughput methylation analysis in ob/ob mice. ob/ob mice genetically lack leptin receptor, and as a result, eat excessively without reaching satiety. Both ob/ob mice and control WT littermates were feeding on the same diet but the ob/ob weighed almost double than the controls (mean of 53 vs. 26 grams, Fig. 5a). In high throughput m⁶Am methylation profiling, we similarly detected the typical BCA genomic consensus and an upstream TATA box (Fig. 5b), (for the non-5′ UTR m⁶A peaks RNA consensus see Supplementary Fig. 5a). Most notably, both Fabp2 and Fabp5 were uniquely m⁶Am methylated in WT lean mice, but not in fat ob/ob mice, a pattern that was also detected in the HFD fed mice. Again, we observed that the majority of m⁶Am peaks (63%) were either lean or fat specific, whereas most m⁶A peaks (62%) were peaks common to both lean and fat mice (for the full m⁶Am and m⁶A methylation maps see Supplementary Data 5, 6). Similarly, we identified higher expression levels of FTO in ob/ob mice than in WT controls (Fig. 5c, d). In order to examine if m⁶Am has comparable regulation properties in ob/ob mice, we examined the differential contribution of m⁶Am to mRNA levels by RNA-seq. Again, we found that genes which lost m⁶Am methylation in ob/ob mice, had a higher fraction of downregulated genes than expected, while among genes that gained m⁶Am methylation, a higher-than-expected fraction was upregulated (Fig. 5e). To complement this analysis, we performed high throughput proteomic profiling and observed a similar regulation pattern at the protein level (Fig. 5f) further supporting the regulatory contribution of m⁶Am to translation. In accordance, overall m⁶Am modified genes displayed consistently higher protein levels than non-m⁶Am modified genes, even after adjusting for their RNA expression level (Supplementary Fig. 5b).

**Fig. 5: m⁶Am regulates mRNA and protein expression in ob/ob mice.**

Discussion

In this study, we explored both the m⁶Am and m⁶A liver methylation profile dynamics and functions in mice under HFD and in ob/ob mice. We present data linking the m⁶Am modification with dynamic functional regulation of obesity. Gene ontology (GO) analysis revealed that m⁶Am- but not m⁶A-methylated transcripts were highly enriched in GO terms of metabolic processes. Additionally, differential removal of the m⁶Am modification in fat mice was associated with downregulation of protein and mRNA transcript levels, and correspondingly, upregulation of protein and mRNA expression upon m⁶Am addition. We identified differentially m⁶Am methylated key metabolic genes in lean and obese mice including the two m⁶Am lean specific methylated fatty acid-binding proteins Fabp2 and Fabp5. Most notably, the silencing of Fabp2 is known for its contribution to weight gain in male mice²⁷. Similarly, Fabp5 is a pivotal metabolic gene whose silencing is implicated in preserving healthier metabolic parameters including insulin sensitivity²⁸. Fabp5 was similarly identified as an m⁶Am-methylated gene transcript²⁴. We further observed that FTO is noticeably upregulated in fat mice and its cellular overexpression and knockdown correspondingly affect the protein levels of its m⁶Am-methylated targets. Most notably, the same pattern of results was recapitulated in two complementing in-vivo perturbations of obesity—HFD and in ob/ob overfeeding mice. Lastly, by analyzing the methylation profile of mESCs depleted entirely of m⁶A mRNA background via Mettl3 KO, we deduced m⁶Am sites in mRNA and their role in translation regulation in a controlled un-confounded way, and generated a list of confirmed mammalian m⁶Am-methylated genes. This analysis of unequivocal m⁶Am sites identified that capped m⁶Am genes have higher protein and mRNA levels, increased mRNA stability, ribosome footprints, and ribosome release scores (RRS). These results strongly establish the role of m⁶Am in translation regulation, and inform the unresolved debate on its functions^{21,22,23,24,25}.

Importantly, the m⁶A/m eraser, FTO, is considered one of the genes that are most robustly associated with human obesity based on single nucleotide polymorphism (SNP) in one of its introns¹. While Fto overexpression in mice was shown to actively promote obesity²⁹, affect hepatic leptin-stat3-glucose signaling³⁰, and conversely protect from obesity in an Fto KO mouse model³¹, the nature of this association is still not clearly defined. It is also noteworthy that the expression of Fto was shown to be dynamically regulated in various obesity models and tissues. For example, it was observed to be downregulated in adipose tissue in HFD mice³² and in mouse liver following insulin treatment³³; or upregulated in the liver of rats with Nonalcoholic Fatty Liver Disease³⁴ and in mouse liver of fasting mice³⁵; but notably, observed in this study as well to be significantly upregulated in the liver of both HFD and ob/ob overfeeding mice. These findings highlight the significance of a dynamic regulation of Fto expression in metabolic tissues and specifically in the liver. While Fto expression is one important layer of regulation for m⁶Am removal, it is possible that additional, yet unknown, erasers and writers are at play and further modulate m⁶Am deposition. These players may also participate in the cases where m⁶Am was detected only in fat mice. Note that in our data, we did not observe meaningful differences in the expression of the recently identified m⁶Am writer Pcif1, yet a small 1.39-fold increase in its mRNA expression was observed in HFD.

Despite the clear associations and expression dynamics of FTO in obesity, to date, little is known about the contributions and interplay between FTO’s two target modifications—m⁶A, and m⁶Am in affecting metabolism. The current work significantly promotes our understanding of the mRNA modification metabolic effects by analyzing both of the mRNA modification targets of FTO (m⁶A and m⁶Am) in two different in vivo obesity mouse models. Our findings support a more dominant metabolic contribution of m⁶Am rather than m⁶A in mediating metabolic effects under HFD and in ob/ob overfeeding mice. In addition, the identification of lean-specific m⁶Am-methylated key metabolic genes (including Fabp2, and Fabp5) further support m⁶Am function in the dynamic regulation of obesity. Importantly, the higher-than-expected fraction of downregulated genes that lost m⁶Am, suggests a significant involvement of m⁶Am in the dynamic regulation of protein and mRNA expression in obesity. In addition, the direct analysis of unequivocal m⁶Am targets in cells deprived of m⁶A background implies that the putative mechanisms of action of m⁶Am involve increased mRNA half-life, but also increased translation efficiency.

In this study, we found that m⁶Am-modified transcripts are enriched in GO terms of metabolic processes while m⁶A-modified transcripts do not exhibit a clear enrichment. This suggests that the m⁶Am modification may have a dominant role in obesity regulation. Yet, our findings do not rule out the possibility of additional roles of FTO that can also be mediated by m⁶A (e.g.,⁴^,⁵), since both modifications were reported to be valid substrates of FTO²^,³. While we found only a small subset of obesity metabolic genes that were differentially m⁶A methylated, it is possible that some of these genes have a yet unknown metabolic function.

RNA epitranscriptomics is a rapidly growing field that unveiled dramatic revelations of various mRNA modifications in affecting gene regulation and multiple cellular functions including alternative splicing (e.g.,¹²^,⁸), stability³⁶, and translation initiation⁹. The current work highlights the dynamic involvement of the m⁶Am modification in a disease state of obesity and in metabolic regulation. In particular, we also strictly support these observed functions of m⁶Am by validating them in pure experimental conditions devoid of the confounding effects of m⁶A background, a critical issue that is likely a significant source for the continuous disagreement on its postulated functions^{21,22,23,24,25}. Most notably, our findings demonstrate that m⁶Am is dynamically regulated in-vivo in response to high fat Western diet, and in response to excessive eating in ob/ob mice, thus enabling significant flexibility that is vital to changing metabolic environments³⁷. Taking together, our findings strongly support a significant new layer of dynamic functional regulation in obesity by the m⁶Am mRNA modification.

Methods

Cell lines, cultures, and treatments

Human hepatocellular carcinoma cell line, HepG2 (ATTC HB-8065), was maintained in DMEM (Gibco, Invitrogen) containing 4.5 g/l glucose and L-glutamine supplemented with 10% FBS and penicillin/streptomycin. Cells were routinely checked for mycoplasma contamination and tested negative. Where indicated, HepG2 cells were transfected with a human FTO gene plasmid that was received from² and is based on the mammalian vector pcDNA3 (Invitrogen) with N-terminal FLAG-tag. The FTO knockdown siRNA sequence was [AAAUAGCCGCUGCUUGUGAGA] (Sigma). Transient transfections were performed using the Lipofectamine reagent (Lipofectamine 2000 and Lipofectamine RNAimax, Invitrogen, USA) according to the manufacturer’s instruction. Mettl3 knockout mESCs and WT control mESCs were obtained from⁸.

Animal backgrounds and treatments

Animal rooms were maintained at 21–24 °C and 35–75% relative humidity, with 12/12 h (7 a.m. to 7 p.m.) dark–light cycle. Cages were routinely replaced every 10–14 days. We have complied with all relevant ethical regulations for animal testing and research. All experiments received ethical approval and were conducted in accordance with the Institutional Animal Care and Use Committee at Bar Ilan University.

High fat diet fed mice

Four weeks old male C57BL/6 mice were kept under specific pathogen-free and 12 h day/night conditions and fed a standard chow diet (24 kcal% Protein, 65 kcal% Carbohydrate,11 kcal% Fat, Altromin 1324 tpf) for the first 2 weeks, until 6 weeks of age. The mice were then randomly divided into two diet groups of 5 mice each. Mice were fed with a standard chow diet for an additional 17 weeks, and in the HFD group were fed high fat diet (17 kcal% Protein, 23 kcal% Carbohydrate,60 kcal% Fat, Altromin C-1090-60). All groups were fed ad libitum. Following this period, the mice were sacrificed and tissues harvested, snap-frozen in liquid nitrogen, and stored in −80 °C until further use. 3 different HFD biological replicate mice liver tissues and 3 controls were used for m⁶A/m⁶Am-seq analysis.

Ob/ob mice

Ob/ob and WT homozygotes/heterozygotes littermates mice were of C57BL/6 background. Liver tissues from 3 different ob/ob biological replicate mice (13 weeks-old; 2 males, 1 female) and 3 WT (13 weeks-old; 2 males, 1 female) mice controls were used for m⁶A/m⁶Am-seq analysis. Additional 3 ob/ob mice (1 male, 13 weeks-old; and 2 females 36 weeks-old) were used for the extended western blot.

Blood parameters

Blood glucose was measured from samples obtained by tail bleeding after 6 h fasting. Glucose levels were tested using Ascensia Elite glucose meter (Bayer, Leverkusen, Germany).

NMR analysis

Mice were scanned using the Minispec Analyzer (Bruker Minispec LF50), which measures lean and fat composition.

RNA purification

Total RNA from cells in culture and mouse tissues was purified using PerfectPure RNA Cultured Cell Kit (5′ Prime) and PerfectPure RNA Tissue Kit (5′ Prime), respectively, and DNase-treated. RNA integrity was evaluated on Bioanalyzer (Agilent 2100 Bioanalyzer), requiring a minimal RNA integrity number (RIN) of 8.5. Experiments were conducted in biological replicates.

m⁶A/m⁶Am-seq

Total RNA (>300 µg) of samples were chemically fragmented into ~100-nucleotide-long fragments by 5 min incubation at 94 °C in fragmentation buffer (10 mM ZnCl2, 10 mM Tris-HCl, pH 7). The fragmentation reaction was stopped with 0.05 M EDTA, followed by standard ethanol precipitation. Samples were resuspended in H₂O at ~1 mg ml/1 concentration and subjected to m⁶A-seq. Fragmented RNA was incubated for 2 h at 4 °C with 5 mg of affinity-purified anti-m⁶A polyclonal antibody (Synaptic Systems cat. no. 202 003) in IPP buffer (150 mM NaCl, 0.1% NP-40, 10 mM Tris-HCl, pH 7.4). The mixture was then immunoprecipitated with protein-A beads (Repligen) at 4 °C for an additional 2 h. After extensive washing, bound RNA was eluted from the beads with 0.5 mg ml/1 N⁶-methyladenosine (Sigma-Aldrich) in IPP buffer, and ethanol precipitated. RNA was resuspended in H₂O and used for library generation with the NEBNext Ultra RNA Library Prep Kit for Illumina (New England Biolabs). Sequencing was carried out on Illumina HiSeq2500 according to the manufacturer’s instructions, using 10 pM template per sample for cluster generation, and sequencing kit V2 (Illumina). These raw data are available at NCBI SRA series PRJNA701370 (public release date set to 2021-08-01).

Peak calling

Adaptors and low-quality bases were trimmed from raw sequencing reads using cutadapt³⁸. Reads were aligned to the relevant genome (human-hg19, mouse-mm10, using Tophat2 (version 2.0.12)³⁹. Non-unique reads mapping to more than five locations were discarded from downstream analysis. For m⁶A peaks, enrichment of immunoprecipitation over input experiments were identified using MACS2 (version 2.1.0.20140616)⁴⁰. MACS2-identified peaks were intersected with a database of exons of the relevant genome (Ensemble annotation). Only peaks identified (FC ≥ 2, FDR ≤ 0.05) were considered. Negative peaks as a control were identified by switching the immunoprecipitation and input samples.

Identifying m⁶Am sites

Since m⁶Am occurs only in transcription start sites (TSSs), we downloaded the genomic position of TSSs from Refseq (PMID: 24259432), UCSC known genes (PMID: 16500937), GENECODE Version M14 (Ensembl 89; PMID: 16925838), and EPD (PMID: 25378343). In each of the aligned samples, we counted the number of 5′ read ends that resides within 25 bp intervals downstream of the TSS. In order to avoid PCR duplicates we allowed a maximum of 5 duplicate reads ending in a single bp position. For each TSS we calculated the fold change as follows:

$$\frac{{{{{{{{\rm{Number}}}}}}\,{{{{{\rm{Of}}}}}}\,5^{\prime} \,{{{{{\rm{reads}}}}}}\,{{{{{\rm{in}}}}}}\,{{{{{\rm{TSS}}}}}}\,{{{{{\rm{(IP)}}}}}}}}}{{{{{{\rm{Number}}}}}}\,{{{{{\rm{Of}}}}}}\,5^{\prime} \,{{{{{\rm{reads}}}}}\,{{{{{\rm{in}}}}}}\,{{{{{\rm{TSS}}}}}}\,{{{{{\rm{(input)}}}}}}}}\,*\, \frac{{{{{{\rm{Library}}}}}}\,{{{{{\rm{size}}}}}}\,{{{{{\rm{(input)}}}}}}}{{{{{{\rm{Library}}}}}}\,{{{{{\rm{size}}}}}}\,{{{{{\rm{(IP)}}}}}}}$$

Cases with zero input coverage within 100 bp downstream of the TSS (to account for the average length of RNA fragments) were filtered out. For each TSS p-value was calculated using fisher’s exact test followed by FDR correction. Only TSSs with over 5 reads in IP, including a fold change greater than 2 and FDR ≤ 0.05 (using Benjamini–Hochberg multiple testing adjustment), and that did not have an identified m¹A peak in that location⁹ were kept for further analyses. In genes with more than one TSS we included the most significant TSS. For negative peak calculation as a control we performed the same approach but with using the IP as input and vice versa. Only peaks occurring in at least 2 of 3 mice were considered as m⁶Am methylated in the lean or fat mice groups.

m⁶Am genomic consensus motif

For each TSS suspected to be m⁶Am methylated we retrieved the nucleotide genomic sequence 40 bp upstream and 5 bp downstream. The nucleotide probability logo was created using webLogo3 (PMID: 15173120) on all genes with a significant TSS peak.

m⁶A MEME RNA consensus motif search

MACS-identified peaks with FDR#5% were sorted according to their fold change. The top 500 peaks falling within known genes were placed in de novo motif analysis. 101-nucleotide-long sequences derived from the sense strand and centered around the peak summit were used as input for MEME⁴¹.

RNA expression levels

Fragments per kilobase of transcript per million mapped reads (FPKM) values were calculated by the CUFFLINKS tool (version 2.2.1)⁴². And differential expression analysis was done using the DESeq Bioconductor package as part of the Cufflinks package⁴³. Features were considered as differentially expressed only when fold change >1.5.

Gene ontology (GO) enrichment

m⁶Am or CDS (>50 bp) m⁶A Methylated genes in ≥2 mice were uploaded to DAVID Bioinformatics Resources (http://david.abcc.ncifcrf.gov) and functional annotation clustering enrichment analysis was performed using the default parameters with all adequately expressed genes as background. The resulting top 10 most enriched clusters were reported using the most occurring relevant terms.

Proteomics

Sample preparation

Experiments were conducted in biological replicates. Cell pellets were lysed with 5% SDS in 50 mM Tris-HCl. Lysates were incubated at 96 °C for 5 min, followed by six cycles of 30 s of sonication (Bioruptor Pico, Diagenode, USA). Protein concentration was measured using the BCA assay (Thermo Scientific, USA) and a total of 100 μg protein was reduced with 5 mM dithiothreitol and alkylated with 10 mM iodoacetamide in the dark. Each sample was loaded onto S-Trap mini-columns (Protifi, USA) as described⁴⁴. In brief, after loading, samples were washed with 90:10% methanol/50 mM ammonium bicarbonate. Samples were then digested with trypsin (1:50 trypsin/protein) for 1.5 h at 47 °C. The digested peptides were eluted using 50 mM ammonium bicarbonate; trypsin was added to this fraction and incubated overnight at 37 °C. Two more elutions were made using 0.2% formic acid and 0.2% formic acid in 50% acetonitrile. The three elutions were pooled together and vacuum-centrifuged to dry. Samples were kept at −80 °C until analysis.

Liquid chromatography

ULC/MS grade solvents were used for all chromatographic steps. Each sample was fractionated using high pH reversed-phase followed by low pH reversed-phase separation. 200 µ digested protein was loaded using high Performance Liquid Chromatography (Acquity H Class Bio, Waters, Milford MA, USA). Mobile phase was: (A) 20 mM ammonium formate pH 10.0, (B) acetonitrile. Peptides were separated on an XBridge C18 column (3×100mm, Waters) using the following gradient: 3% B for 2 min, linear gradient to 40% B in 50 min, 5 min to 95% B, maintained at 95% B for 5 min and then back to initial conditions. Peptides were fractionated into 15 fractions. The fractions were then pooled: 1 with 8, 2 with 9, 3 with 10, 4 with 11, 5 with 12, 6 with 13, and 7 with 14–15. Each fraction was vacuum dried, then reconstituted in 25 µL 97:3 acetonitrile: water + 0.1% formic acid. Each pooled fraction was loaded using split-less nano-Ultra Performance Liquid Chromatography (10 kpsi nanoAcquity; Waters, Milford, MA, USA). The mobile phase was: A) H₂O + 0.1% formic acid and B) acetonitrile + 0.1% formic acid. Desalting of the samples was performed online using a reversed-phase Symmetry C18 trapping column (180 µm internal diameter, 20 mm length, 5 µm particle size; Waters). The peptides were then separated using a T3 HSS nano-column (75 µm internal diameter, 250 mm length, 1.8 µm particle size; Waters) at 0.35 µL/min. Peptides were eluted from the column into the mass spectrometer using the following gradient: 4 to 30%B in 150 min, 30 to 90%B in 5 min, maintained at 90% for 5 min, and then back to initial conditions.

Mass spectrometry

The nanoUPLC was coupled online through a nanoESI emitter (10 μm tip; New Objective; Woburn, MA, USA) to a quadrupole orbitrap mass spectrometer (Q Exactive HF, Thermo Scientific) using a FlexIon nanospray apparatus (Proxeon). Data were acquired in data-dependent acquisition (DDA) mode, using a Top20 method. MS1 resolution was set to 120,000 (at 200 m/z), mass range of 300–1650 m/z, AGC of 3e6, and maximum injection time was set to 20msec. MS2 resolution was set to 30,000, quadrupole isolation 1.7 m/z, AGC of 1e5, dynamic exclusion of 30 s, and maximum injection time of 60msec.

Data processing and analysis

Raw data were imported into the Expressionist^® software version 14 (Genedata, Switzerland) and processed as described in⁴⁵. The software was used for retention time alignment and peak detection of precursor peptides. A master peak list was generated from all MS/MS events and sent for database searching using Mascot v2.5.1 (Matrix Sciences). Data were searched against the mouse or human sequences UniprotKB (http://www.uniprot.org/) appended with common laboratory contaminant proteins. The fixed modification was set to carbamidomethylation of cysteines and variable modifications were set to oxidation of methionines and deamidation of N or Q. Search results were then filtered using the PeptideProphet algorithm to achieve maximum false discovery rate of 1% at the protein level. Peptide identifications were imported back to Expressionist to annotate identified peaks. Quantification of proteins from the peptide data was performed using an in-house script. Data was normalized base on the total ion current. Protein abundance was obtained by summing the three most intense, unique peptides per protein. Protein abundance data was then log2 transformed in subsequent statistical analysis.

Analyses of SILAC deep proteomics, mRNA stability half-life, and ribo-seq in Mettl3 KO and WT mESCs

We reanalyzed our previously published data⁸. The unprocessed raw data of these datasets were previously deposited in NCBI GEO series GSE61998, and the processed data is available here as a source data file.

Western blot analyses

Samples were separated on 10% (w/v) polyacrylamide Bis-Tris gels (Invitrogen) and transferred onto nitrocellulose membrane using iBlot gel transfer system (Invitrogen) set to P3 for 8 min with iBlot gel transfer stacks (Invitrogen). Membranes were blocked in 5% BSA, 0.05% Tween-20 in PBS for 1 h, and then incubated overnight at 4 °C with anti-FTO antibody (Millipore, 5-2H10; 1 µg/mL); anti-FABP5 antibody (R&D AF-1476; 0.5 µg/mL), anti-FABP2 antibody (BOSTER, PB9943; 0.5 µg/mL); anti-Mettl3 (Proteintech Group 15073-1-AP; 1:3000); anti-HSC-70 (Santa cruse SC-7298; 1:200), and a-TUBULIN (sigma T9026; 1:500). Proteins were visualized using the SuperSignal West Pico Luminol/Enhancer solution (Thermo scientific).

Statistical analysis

All statistical analyses (unless stated otherwise) were performed using the SPSS version 18 software for Statistical Computing.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability

Data supporting the findings of this work are available within the paper and its Supplementary Information files and data. Source data are provided with this paper. A reporting summary for this article is available as a Supplementary Information file. All raw m⁶A/m-seq sequence reads data generated in this study have been deposited in NCBI SRA database under accession code PRJNA701370. All raw mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with dataset identifiers PXD029217, PXD029268, and PXD029280. The processed proteomics profiling data have been deposited in Zenodo database under accession code 10.5281/zenodo.5339101. RNA Stability and ribo-seq assays were previously deposited in⁸ via NCBI GEO series GSE61998. The processed source data underlying Figs. 1c, d, f–j, 2a–c, 3a–e, 4a–d, 5a, c–f; as well as Supplementary Figs. S1a−e, S2a, b, S3b, S4a−c, and S5b are provided as a Source data file. Source data are provided with this paper.

References

Frayling, T. M. et al. A common variant in the FTO gene is associated with body mass index and predisposes to childhood and adult obesity. Science 316, 889–894 (2007).
Article ADS CAS PubMed PubMed Central Google Scholar
Jia, G. et al. N6-methyladenosine in nuclear RNA is a major substrate of the obesity-associated FTO. Nat. Chem. Biol. 7, 885–887 (2011).
Article CAS PubMed PubMed Central Google Scholar
Mauer, J. et al. Reversible methylation of m6Am in the 5’ cap controls mRNA stability. Nature 541, 371–375 (2017).
Article ADS CAS PubMed Google Scholar
Zhao, X. et al. FTO-dependent demethylation of N6-methyladenosine regulates mRNA splicing and is required for adipogenesis. Cell Res. 24, 1403–1419 (2014).
Article CAS PubMed PubMed Central Google Scholar
Hess, M. E. et al. The fat mass and obesity-associated gene (Fto) regulates activity of the dopaminergic midbrain circuitry. Nat. Neurosci. 16, 1042–1048 (2013).
Article CAS PubMed Google Scholar
Smemo, S. et al. Obesity-associated variants within FTO form long-range functional connections with IRX3. Nature 507, 371–375 (2014).
Article ADS CAS PubMed PubMed Central Google Scholar
Stratigopoulos, G. et al. Hypomorphism for RPGRIP1L, a ciliary gene vicinal to the FTO locus, causes increased adiposity in mice. Cell. Metab. 19, 767–779 (2014).
Article CAS PubMed PubMed Central Google Scholar
Geula, S. et al. Stem cells. m6A mRNA methylation facilitates resolution of naive pluripotency toward differentiation. Science 347, 1002–1006 (2015).
Article ADS CAS PubMed Google Scholar
Dominissini, D. et al. The dynamic N(1)-methyladenosine methylome in eukaryotic messenger RNA. Nature 530, 441–446 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Frye, M., Jaffrey, S. R., Pan, T., Rechavi, G. & Suzuki, T. RNA modifications: what have we learned and where are we headed? Nat. Rev. Genet. 17, 365–372 (2016).
Article CAS PubMed Google Scholar
Fu, Y., Dominissini, D., Rechavi, G. & He, C. Gene expression regulation mediated through reversible m(6)A RNA methylation. Nat. Rev. Genet. 15, 293–306 (2014).
Article CAS PubMed Google Scholar
Dominissini, D. et al. Topology of the human and mouse m6A RNA methylomes revealed by m6A-seq. Nature 485, 201–206 (2012).
Article ADS CAS PubMed Google Scholar
Meyer, K. D. et al. Comprehensive analysis of mRNA methylation reveals enrichment in 3′ UTRs and near stop codons. Cell 149, 1635–1646 (2012).
Article CAS PubMed PubMed Central Google Scholar
Linder, B. et al. Single-nucleotide-resolution mapping of m6A and m6Am throughout the transcriptome. Nat. Methods 12, 767–772 (2015).
Article CAS PubMed PubMed Central Google Scholar
Kruse, S. et al. A novel synthesis and detection method for cap-associated adenosine modifications in mouse mRNA. Sci. Rep. 1, 126 (2011).
Article PubMed PubMed Central Google Scholar
Meyer, K. D. & Jaffrey, S. R. Rethinking m6A Readers, Writers, and Erasers. Annu. Rev. Cell Dev. Biol. 33, 319–342 (2017).
Article CAS PubMed PubMed Central Google Scholar
Carninci, P. et al. Genome-wide analysis of mammalian promoter architecture and evolution. Nat. Genet. 38, 626–635 (2006).
Article CAS PubMed Google Scholar
Ni, T. et al. A paired-end sequencing strategy to map the complex landscape of transcription initiation. Nat. Methods 7, 521–527 (2010).
Article CAS PubMed PubMed Central Google Scholar
Plessy, C. et al. Linking promoters to functional transcripts in small samples with nanoCAGE and CAGEscan. Nat. Methods 7, 528–534 (2010).
Article CAS PubMed PubMed Central Google Scholar
Schwartz, S. et al. Perturbation of m6A writers reveals two distinct classes of mRNA methylation at internal and 5′ sites. Cell. Rep. 8, 284–296 (2014).
Article CAS PubMed PubMed Central Google Scholar
Sun, H., Zhang, M., Li, K., Bai, D. & Yi, C. Cap-specific, terminal N(6)-methylation by a mammalian m(6)Am methyltransferase. Cell Res. 29, 80–82 (2019).
Article CAS PubMed Google Scholar
Akichika, S. et al. Cap-specific terminal N (6)-methylation of RNA by an RNA polymerase II-associated methyltransferase. Science https://doi.org/10.1126/science.aav0080 (2019).
Sendinc, E. et al. PCIF1 catalyzes m6Am mRNA methylation to regulate gene expression. Mol. Cell 75, 620–630.e9 (2019).
Article CAS PubMed PubMed Central Google Scholar
Boulias, K. et al. Identification of the m(6)Am methyltransferase PCIF1 reveals the location and functions of m(6)Am in the transcriptome. Mol. Cell 75, 631–643.e8 (2019).
Article CAS PubMed PubMed Central Google Scholar
Pandey, R. R. et al. The mammalian cap-specific m(6)Am RNA methyltransferase PCIF1 regulates transcript levels in mouse tissues. Cell. Rep. 32, 108038 (2020).
Article CAS PubMed Google Scholar
Guttman, M., Russell, P., Ingolia, N. T., Weissman, J. S. & Lander, E. S. Ribosome profiling provides evidence that large noncoding RNAs do not encode proteins. Cell 154, 240–251 (2013).
Article CAS PubMed PubMed Central Google Scholar
Vassileva, G., Huwyler, L., Poirier, K., Agellon, L. B. & Toth, M. J. The intestinal fatty acid binding protein is not essential for dietary fat absorption in mice. FASEB J. 14, 2040–2046 (2000).
Article CAS PubMed Google Scholar
Maeda, K. et al. Role of the fatty acid binding protein mal1 in obesity and insulin resistance. Diabetes 52, 300–307 (2003).
Article CAS PubMed Google Scholar
Church, C. et al. Overexpression of Fto leads to increased food intake and results in obesity. Nat. Genet. 42, 1086–1092 (2010).
Article CAS PubMed PubMed Central Google Scholar
Bravard, A. et al. FTO contributes to hepatic metabolism regulation through regulation of leptin action and STAT3 signalling in liver. Cell. Commun. Signal. 12, 4-811X-12–4-811X-14 (2014).
Article Google Scholar
Fischer, J. et al. Inactivation of the Fto gene protects from obesity. Nature 458, 894–898 (2009).
Article ADS CAS PubMed Google Scholar
Ronkainen, J. et al. Fat mass- and obesity-associated gene Fto affects the dietary response in mouse white adipose tissue. Sci. Rep. 5, 9233 (2015).
Article CAS PubMed PubMed Central Google Scholar
Mizuno, T. M., Lew, P. S., Luo, Y. & Leckstrom, A. Negative regulation of hepatic fat mass and obesity-associated (Fto) gene expression by insulin. Life Sci. 170, 50–55 (2017).
Article CAS PubMed Google Scholar
Guo, J. et al. Fat mass and obesity-associated gene enhances oxidative stress and lipogenesis in nonalcoholic fatty liver disease. Dig. Dis. Sci. 58, 1004–1009 (2013).
Article CAS PubMed Google Scholar
Poritsanos, N. J., Lew, P. S. & Mizuno, T. M. Relationship between blood glucose levels and hepatic Fto mRNA expression in mice. Biochem. Biophys. Res. Commun. 400, 713–717 (2010).
Article CAS PubMed Google Scholar
Wang, X. et al. N6-methyladenosine-dependent regulation of messenger RNA stability. Nature 505, 117–120 (2014).
Article ADS PubMed Google Scholar
Goodpaster, B. H. & Sparks, L. M. Metabolic flexibility in health and disease. Cell. Metab. 25, 1027–1036 (2017).
Article CAS PubMed PubMed Central Google Scholar
Martin, M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet. J. 17, 10–12 (2011).
Article Google Scholar
Kim, D. et al. TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions, and gene fusions. Genome Biol. 14, R36-2013-14-4–r36 (2013).
Article Google Scholar
Zhang, Y. et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol. 9, R137-2008-9-9-r137 (2008).
Article Google Scholar
Bailey, T. L. & Elkan, C. Fitting a mixture model by expectation maximization to discover motifs in biopolymers. Proc. Int. Conf. Intell. Syst. Mol. Biol. 2, 28–36 (1994).
CAS PubMed Google Scholar
Trapnell, C. et al. Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat. Biotechnol. 28, 511–515 (2010).
Article CAS PubMed PubMed Central Google Scholar
Roberts, A., Pimentel, H., Trapnell, C. & Pachter, L. Identification of novel transcripts in annotated genomes using RNA-Seq. Bioinformatics 27, 2325–2329 (2011).
Article CAS PubMed Google Scholar
Elinger, D., Gabashvili, A. & Levin, Y. Suspension trapping (S-Trap) is compatible with typical protein extraction buffers and detergents for bottom-up proteomics. J. Proteome Res. 18, 1441–1445 (2019).
Article CAS PubMed Google Scholar
Shalit, T., Elinger, D., Savidor, A., Gabashvili, A. & Levin, Y. MS1-based label-free proteomics using a quadrupole orbitrap mass spectrometer. J. Proteome Res. 14, 1979–1986 (2015).
Article CAS PubMed Google Scholar

Download references

Acknowledgements

We thank the Kahn Family Foundation for their continuous support. This work was supported in part by grants from the Flight Attendant Medical Research Institute (FAMRI); the Ernest and Bonnie Beutler Research Program, by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Grant agreement No. 743168) and by the German–Israeli Project Cooperation (DIP) of the German Federal Ministry of Education and Research. GR is a member of the Sagol Neuroscience Network and holds the Djerassi Chair in Oncology at the Sackler Faculty of Medicine, Tel Aviv University. HC is a the head of the SAGOL center of healthy human aging. MSBH was supported by the Herczeg Institute of Aging. This work was performed in partial fulfillment of the requirements for a PhD degree to MSBH, Sackler School of Medicine, Tel-Aviv University. We thank Hannah Kanety for helpful comments and antibody assistance and Yshai Levin of the De Botton Protein Profiling institute of the Nancy and Stephen Grand Israel National Center for Personalized Medicine, Weizmann Institute of Science for proteomic profiling services.

Author information

These authors contributed equally: Moshe Shay Ben-Haim, Yishay Pinto.

Authors and Affiliations

Cancer Research Center, Chaim Sheba Medical Center, Tel Hashomer, Israel
Moshe Shay Ben-Haim, Sharon Moshitch-Moshkovitz, Vera Hershkovitz, Nitzan Kol, Tammy Diamant-Levi, Ninette Amariglio & Gideon Rechavi
Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel
Moshe Shay Ben-Haim, Sharon Moshitch-Moshkovitz & Gideon Rechavi
The Wohl Institute for Translational Medicine, Chaim Sheba Medical Center, Tel Hashomer, Israel
Moshe Shay Ben-Haim, Sharon Moshitch-Moshkovitz, Vera Hershkovitz, Nitzan Kol, Tammy Diamant-Levi, Ninette Amariglio & Gideon Rechavi
Department of Psychology, The Hebrew University of Jerusalem, Jerusalem, Israel
Moshe Shay Ben-Haim
The Mina and Everard Goodman Faculty of Life Sciences, Bar-Ilan University, Ramat-Gan, Israel
Yishay Pinto, Ninette Amariglio & Haim Y. Cohen
The Joseph Sagol Neuroscience Center, Chaim Sheba Medical Center, Tel Hashomer, Israel
Michal Schnaider Beeri
Mount Sinai School of Medicine, New York, NY, USA
Michal Schnaider Beeri

Authors

Moshe Shay Ben-Haim
View author publications
You can also search for this author in PubMed Google Scholar
Yishay Pinto
View author publications
You can also search for this author in PubMed Google Scholar
Sharon Moshitch-Moshkovitz
View author publications
You can also search for this author in PubMed Google Scholar
Vera Hershkovitz
View author publications
You can also search for this author in PubMed Google Scholar
Nitzan Kol
View author publications
You can also search for this author in PubMed Google Scholar
Tammy Diamant-Levi
View author publications
You can also search for this author in PubMed Google Scholar
Michal Schnaider Beeri
View author publications
You can also search for this author in PubMed Google Scholar
Ninette Amariglio
View author publications
You can also search for this author in PubMed Google Scholar
Haim Y. Cohen
View author publications
You can also search for this author in PubMed Google Scholar
Gideon Rechavi
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

G.R., H.Y.C. and M.S.B.H. conceived the idea and designed experiments. M.S.B.H. performed all experiments with help from S.M.M., V.H., T.D.L. and M.S.B. Y.P. performed the HFD exposure in mice. M.S.B.H. and Y.P. conducted all bioinformatic analysis with help from N.K. M.S.B.H. and G.R. wrote the paper with meaningful insight and editing from S.M.M., N.A. and Y.P.

Corresponding authors

Correspondence to Haim Y. Cohen or Gideon Rechavi.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review information

Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Description of Additional Supplementary Files

Data S1

Data S2

Data S3

Data S4

Data S5

Data S6

Reporting Summary

Source data

Source data file

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Ben-Haim, M.S., Pinto, Y., Moshitch-Moshkovitz, S. et al. Dynamic regulation of N⁶,2′-O-dimethyladenosine (m⁶Am) in obesity. Nat Commun 12, 7185 (2021). https://doi.org/10.1038/s41467-021-27421-2

Download citation

Received: 04 January 2019
Accepted: 05 November 2021
Published: 10 December 2021
DOI: https://doi.org/10.1038/s41467-021-27421-2

This article is cited by

RNA modification-mediated mRNA translation regulation in liver cancer: mechanisms and clinical perspectives
- Shuibin Lin
- Ming Kuang
Nature Reviews Gastroenterology & Hepatology (2024)
RNA modifications in physiology and disease: towards clinical applications
- Sylvain Delaunay
- Mark Helm
- Michaela Frye
Nature Reviews Genetics (2024)
EMDL_m6Am: identifying N6,2′-O-dimethyladenosine sites based on stacking ensemble deep learning
- Jianhua Jia
- Zhangying Wei
- Mingwei Sun
BMC Bioinformatics (2023)
Epitranscriptomics in metabolic disease
- Yoshihiro Matsumura
- Fan-Yan Wei
- Juro Sakai
Nature Metabolism (2023)
Biological roles of adenine methylation in RNA
- Konstantinos Boulias
- Eric Lieberman Greer
Nature Reviews Genetics (2023)

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.