Mechanisms and rationales of SAM homeostasis

Abstract

S-Adenosylmethionine (SAM) is the primary methyl donor for numerous cellular methylation reactions. Its central role in methylation and involvement with many pathways link SAM availability to the regulation of cellular processes, whose dysregulation can contribute to disease states such as cancer or neurodegeneration. Emerging evidence indicates intracellular SAM levels are maintained within an optimal range by way of a variety of homeostatic mechanisms. This suggests that the need to maintain SAM homeostasis may represent a significant evolutionary pressure across all kingdoms of life. Here we review how SAM controls cellular functions at the molecular level, and strategies to maintain SAM homeostasis. We propose that SAM exerts a broad and underappreciated influence in cellular regulation that remains to be fully elucidated.

Significance of SAM homeostasis

The SAM precursor, methionine, is a sulfur-containing essential amino acid in mammals, therefore must be obtained through the diet by humans and other animals. It can, however, be synthesized de novo from sulfur-containing precursors in many microbial species, including the budding yeast Saccharomyces cerevisiae [1]. As a variety of nutrients are needed for its synthesis, SAM levels can be sensitive to nutritional and environmental changes. Nonetheless, SAM is a critical metabolite that serves as the major methyl donor for methylation reactions in the cell. This includes methylation of DNA, RNA, proteins, and lipids to regulate gene expression as well as protein and membrane functions. Moreover, SAM is also a precursor for synthesis of polyamines and glutathione, which are required for proliferation and redox balance. Therefore, SAM homeostasis and the regulation of SAM synthesis and consumption are of significant importance to cells.

SAM/methionine metabolism

SAM is synthesized from methionine in an ATP-dependent step by methionine adenosyltransferase (MAT), also known as SAM synthetase. The metabolism of SAM/methionine is reviewed elsewhere [1, 2] and is summarized in Figure 1. SAM and methionine are important players in one-carbon metabolism (see Glossary). Pathways linked to SAM also feed into synthesis of other key metabolites such as polyamines (methionine salvage), purines and pyrimidines (one-carbon/folate metabolism), as well as cysteine and glutathione (transsulfuration). As an activated carrier of methyl groups that feeds into a variety of metabolic pathways, SAM can reflect multiple aspects of cellular energy and redox state.

Figure 1. SAM/Methionine metabolism.

Methionine (Met) is an essential amino acid in mammals, but the sulfur assimilation pathway converts sulfate to homocysteine (Hcy) in other organisms such as fungi. Hcy can feed into methionine cycle that produces methionine through methionine synthase (MS). This step requires one-carbon donors, which can be provided via the folate cycle that is comprised of the recycling of tetrahydrofolate (THF) through intermediates such as 5,10-methylene-THF (5, 10-MTHF) and 5-methyl-THF (5-MTHF [32]. Methylthioadenosine (MTA), a byproduct of polyamine synthesis that consumes SAM, can be metabolized to produce Met (methionine salvage). Met is converted to SAM by methionine adenosyltransferase (MAT). SAM is then consumed by methyltransferases (MTases) during transmethylation to generate SAH, which can be hydrolyzed to form homocysteine. Hcy also fuels transsulfuration where Hcy is converted to cystathionine, cysteine, and finally glutathione.

Flux through the methionine cycle is influenced by pathways depicted in Figure 1. Homocysteine lies at a critical juncture in the sulfur metabolism pathway. It can either be re-methylated to methionine or enter transsulfuration to form cysteine. Upon entry into transsulfuration, homocysteine is committed to cysteine and can no longer be converted back to methionine. Multiple types of methionine synthases that differ in their cofactor dependency have been identified, suggesting versatility in controlling homocysteine re-methylation into methionine [3]. SAM inhibits methylene tetrahydrofolate reductase (MTHFR) [4], which generates 5-methyltetrahydrofolate (5-MTHD) utilized by methionine synthase (MS). In mammalian cells, SAM also activates cystathionine β-synthase (CBS) [5], promoting transsulfuration instead of methionine synthesis. Therefore, SAM levels mediate the coordination between the two pathways that utilize homocysteine. When SAM levels are high, flux through transsulfuration increases while re-methylation of homocysteine is diminished. When SAM concentrations are low, inhibition of MTHFR is released and homocysteine is allowed to re-methylate to methionine. Impairment of either of the above pathways will disrupt SAM homeostasis and lead to a metabolic disorder called “hyperhomocysteinemia”, characterized by high homocysteine levels in the blood [6].

Key strategies for maintaining SAM homeostasis

SAM levels are critical to organisms from different kingdoms of life. However, they each employ unique strategies to achieve tightly controlled SAM homeostasis. Here we describe examples of how yeast and mammalian cells modulate the expression and activities of factors involved in SAM synthesis. We then discuss the consumption of SAM and how bulk substrate methylation contributes to SAM homeostasis (Figure 2). Interestingly, bacteria also use their own unique strategy to sense and regulate SAM synthesis (Box 1).

Figure 2. Regulation of SAM homeostasis in different kingdoms of life.

Left. In yeast, the E3 ubiquitin ligase SCF^Met30 regulates the stability of Met4 according to sulfur (and possibly SAM) levels. Met4 is the transcriptional activator for sulfur metabolism genes. Right. In mammals, methionine is an essential amino acid. SAM synthetase (MAT2A) expression level is tightly controlled by the deposition of m⁶A mark on its 3’UTR by the m⁶A methyltransferase METTL16, whose activity is sensitive to SAM levels. Middle. The methyl sink function of key transmethylation substrates enables a large amount of SAM consumption in both yeast and mammalian cells. Higher eukaryotes have developed more complex, multi-layered regulatory mechanism than yeast, that could allow cells to fine-tune SAM levels and perhaps achieve tissue-specific regulation.

Box 1. SAM riboswitches in bacteria.

Riboswitches are structured RNA elements that sense metabolites and regulate expression of operons, through metabolite-induced confirmational changes. The majority of these elements are found in bacteria [121]. Riboswitches that recognize SAM (SAM riboswitches) are found in genes involved in sulfur metabolism and represent the most commonly occurring riboswitches in nature [122]. SAM riboswitches employ two common themes to regulate gene expression of proteins in sulfur metabolism: transcriptional control and translational control (Figure I). The former is achieved by formation of a secondary structure that terminates transcription when bound to SAM [123]. Genes controlled by this mechanism often lack apparent transcriptional regulator binding sites, indicating the importance of direct sensing and dynamic regulation of SAM levels in bacteria. Translational control involves interaction of the SAM riboswitch with the Shine-Dalgarno (SD) sequence. When the riboswitch is bound by SAM, it sequesters the SD sequence of the mRNA, thereby blocking ribosomes from binding to the mRNA [122] and limiting expression of sulfur metabolism genes. Recently, a SAM riboswitch was found to bind uncharged initiator Met-tRNA and derepress SAM’s translational inhibition effect [124]. This provides more complexity to SAM riboswitches and a possible utilization mechanism to sense and balance both methionine and SAM levels.

Box 1, Figure I. SAM riboswitches.

SAM riboswitches are sensors in bacteria that directly control the expression of sulfur metabolism genes.

The MET regulon in yeast

Methionine/SAM starvation in yeast triggers expression of sulfur metabolism genes (MET genes). Most MET genes are induced by the yeast-specific transcriptional activator Met4, and additional co-factors required for recruitment to target genes [7].However, under sulfur-replete conditions, Met4 is ubiquitinated by the E3 ubiquitin ligase SCF^Met30 (Skp1/ Cul1/ F-box protein Met30), leading to inactivation of Met4 and repression of MET genes [8] (Figure 2, left). Interestingly, Met4 harbors an internal ubiquitin interacting motif (UIM) that can insulate it from proteasomal degradation [9]. This may be important to preserve a Met4 protein pool that can be deubiquitinated and become transcriptionally active in a timely manner upon SAM and sulfur starvation.

To date, how Met30, the F-box protein recognizing Met4, is regulated by sulfur-containing metabolites in yeast remains incompletely understood. What could be the sentinel sulfur-containing metabolite that gauges Met30 activity? Although SAM has been implicated in the sensing mechanism, emerging evidence suggests that Met30 may instead sense other sulfur-containing metabolites, such as cysteine [10, 11]. Consistent with this hypothesis, a recent report shows that Met30 responds to flux through the transsulfuration pathway and the biological gas hydrogen sulfide that can be derived from cysteine [12].

Regulation of SAM synthetase (MAT2A) splicing in mammalian cells

Mammalian cells express MAT1A and MAT2A, which encode two MAT catalytic subunits, α1 and α2, respectively. MAT1A is liver-specific, whereas MAT2A is expressed in all human tissues except the adult liver [13]. Regulation of MAT1A and MAT2A expression in hepatocytes is reviewed by Ramani and Lu [13]. Notably, hepatocytes undergo MAT1A silencing and MAT2A induction during liver injury, fibrosis, and carcinogenesis. The two MATs, although catalyzing the same biochemical reaction, may enable distinct metabolic pathways to either promote normal liver function or proliferation. This is possibly achieved by interaction between the individual MAT with distinct factors/enzymes. Consistent with this idea, a recent report shows that a major one-carbon unit formaldehyde inhibits MAT1A catalytic activity by inducing methylation at Cys120 in MAT1A, but does not affect MAT2A [14].

Several studies have revealed that, in non-liver cells, MAT2A expression is regulated by the m⁶A methyltransferase METTL16, whose activity responds to cellular SAM levels [15, 16]. METTL16 methylates the conserved adenosines within six hairpin structures (hp1-hp6) in the 3’UTR of MAT2A. Two distinct mechanisms are proposed to explain METTL16’s regulatory roles.

First, METTL16 alters MAT2A splicing. Under SAM replete conditions, METTL16 transiently binds to and methylates hp1 in the 3’UTR of MAT2A. The intron proximal to hp1 is retained within a population of MAT2A transcripts, resulting in nuclear detention of the intron-retained isoform that undergoes nuclear decay. During methionine starvation, METTL16 shows increased binding to hp1, likely due to impaired enzymatic turnover. This is necessary and sufficient to promote co-transcriptional splicing of the retained intron, which leads to increased protein production [15, 17]. The cleavage factor I_m (CFI_m) complex acts downstream of METTL16 to facilitate MAT2A splicing. This novel function of the CFI_m complex is independent of its major role in poly(A) site selection [18]. It is notable that the concentration of extracellular methionine that triggers splicing of MAT2A in HCT116 cells is on the order of ~11 μM [18], suggesting a methionine:SAM threshold within cells that tightly governs splicing of the SAM synthetase gene transcript.

Second, METTL16 affects the stability of MAT2A mRNA. In SAM-rich conditions, hp1-hp6 in the 3’UTR of MAT2A are highly methylated by METTL16 and MAT2A transcripts are degraded. This may be partially mediated by the m⁶A-binding protein YTHDC1, suggesting a conventional ‘writer-reader’ paradigm for the role of m⁶A in MAT2A stability regulation [16]. Hunter et al. proposed that METTL16 utilizes both mechanisms (Figure 2, right). MAT2A splicing is regulated through hp1, and the stability of MAT2A mRNA is primarily controlled through hp 2–6 [19]. Through this two-tiered model, METTL16 can modulate MAT2A levels within a larger dynamic range. Interestingly, Shima et al. showed that reducing intracellular SAM concentrations does not alter MAT2A splicing [16], suggesting the choice of pathways could also be context-dependent.

The ability of METTL16 to fine-tune MAT2A levels is evident. Mutations that enhance or inhibit METTL16 activities to different degrees render the cells to express MAT2A at corresponding levels [20]. For example, a catalytically inactive mutation triggers the highest level of MAT2A even when SAM is sufficient, and hyperactive METTL16 mutants reduce MAT2A expression even when SAM levels are low [20]. A broader impact of METTL16 in SAM homeostasis is previously reviewed [21]. Interestingly, the Caenorhabditis elegans METTL16 ortholog, METTL10, regulates the alternative splicing and nonsense-mediated mRNA decay (NMD) of SAM synthetase transcripts through m⁶A modification at the distal 3’ splice site [22]. This suggests a conserved role of METTL16 and m⁶A in linking levels of SAM and SAM synthetase expression. However, budding yeast S. cerevisiae does not encode a METTL16 ortholog, and fission yeast Schizosaccharomyces pombe Mtl16 has not been implicated in SAM homeostasis.

Methyl sinks as consumers of SAM

Following SAM synthesis, methylation reactions consume SAM and represent another critical aspect in maintaining SAM homeostasis. Examples of major methylation substrates include histones, phosphatidylethanolamine (PE), and even metabolites such as nicotinamide (Figure 2, middle). An emerging role for these bulk SAM consumers is that they may function as methyl-sinks in both yeast and mammalian cells (reviewed in [23]). Methylation of these substrates can potentially buffer high flux of SAM to govern proper methylation status of other substrates and possibly to minimize aberrant methylations. Moreover, methylation reactions generate SAH, which feeds into the transsulfuration pathway to synthesize cysteine and glutathione. Therefore, these “methyl-sinks” may also be employed to cope with oxidative stress.

Recently, by analyzing expression datasets of cancer cell lines, it was shown that there is poor correlation between histone methylation and transcriptional activity in mammalian cells [24]. There was also an inverse correlation between expression of histone methyltransferases and two other SAM-consuming methyltransferases - Phosphatidylethanolamine N-methyltransferase (PEMT) and Nicotinamide N-Methyltransferase (NNMT) [24]. This indicates a possible hierarchy of different methyl-sink molecules, with histones being the primary SAM consumer in this case. It is worth noting that the methyl-sink functions are, in addition to the well-documented roles of histone methylation in chromatin structure and transcription [25], PE methylation in the production of the major phospholipid phosphatidylcholine (PC) to support membrane integrity and normal hepatic functions [26], and the emerging implications of nicotinamide methylation product N1-methylnicotinamide (MNAM) in maintaining liver and endothelial health [27].

The consumption of SAM via methylation sinks suggests a general model in which only when methionine and SAM are sufficient, can cells afford to utilize transsulfuration to convert sulfur equivalents to cysteine and glutathione. Consistent with this model, cancer cells depend on transsulfuration flux upon cysteine limitation [28], and when yeast cells are starved of methionine, ribosomes pause at cysteine codons instead of the methionine codon, suggesting that methionine takes priority over cysteine during conditions of sulfur amino acid limitation [29].

Sensing of SAM via signal transduction pathways

Mammalian target of rapamycin complex 1 (mTORC1) represents one of the most well-studied signaling pathways that controls cell growth, and can also reportedly sense SAM through multiple mechanisms [30]. One such mechanism is the direct binding of the MTase domain containing protein SAMTOR (S-Adenosylmethionine Sensor Upstream Of mTORC1) to SAM. The dissociation constant of SAMTOR:SAM interaction is ~7 μM, which suggests that the interaction may be sensitive to intracellular SAM levels [31]. Structural studies using Drosophila SAMTOR revealed conformational changes upon SAM binding that block its interaction with GATOR1, a negative regulator of mTORC1 [32]. This inhibits GATOR1 and activates mTORC1 [31]. Interestingly, mTORC1 was also shown to stimulate SAM synthesis [33], suggesting a possible positive feedback mechanism.

In yeast, methionine/SAM promotes methylation of the C-subunit of the general phosphatase PP2A [34]. This leads to dephosphorylated nitrogen permease regulator protein 2 (Npr2), a component of the SEACIT/GATOR1 complex, alleviating inhibition of yeast TORC1 [35, 36]. Methionine also stimulates the synthesis of nitrogen-containing metabolites such as nucleotides through activation of TORC1, again consistent with the hypothesis that maintenance of SAM takes priority over other biosynthetic routes [35]. In mammals, manganese (Mn) exposure induces cognitive impairment and showed lower SAM levels in neuronal cells. These cells show hypomethylation of PP2A, which can be reversed by methionine supplementation [37]. Using Neuronal N2a cells and Madin-Darby canine kidney (MDCK) cells, the Sontag group repeatedly showed that PP2A methylation levels are affected by variation in SAM/SAH ratios either due to supplementing SAM or disturbance of one-carbon metabolism [38–40]. Dysregulation of PP2A methylation has also been linked to Alzheimer’s disease, suggesting a link between SAM homeostasis and Alzheimer’s disease [41].

In addition to a key role for the (m)TORC1 pathway in SAM homeostasis, other signaling pathways also show connections to SAM. For example, cyclic GMP-AMP synthase (cGAS), a critical factor of the cGAS-STING pathway that functions in innate immune response, is recently found to be methylated [42], which is reportedly sensitive to methionine/SAM levels. Methylated cGAS is sequestered in the nucleus, thus inhibiting its activity as a cytosolic DNA sensor. Another example was observed by supplementing SAM to nonalcoholic fatty liver disease (NAFLD) cells increases the overall methylation levels of human antigen R (HuR). This promotes cytosolic localization of HuR and antagonizes a NAFLD-driving cell surface type 1 receptor (AT1R) [43]. Methionine depletion in HeLa cell cultures inhibits canonical Wnt signaling [44]. In line with this, treating mouse embryos with ethionine, a MAT inhibitor, reduces SAM levels and inhibits Wnt/β-catenin signaling, possibly through modulating m⁶A [45]. These examples suggest a broader impact of SAM homeostasis that requires more investigation.

SAM homeostasis as a balance between growth and survival

Building on current evidence, we propose a working model for how SAM fluctuations impact a cell or organism. We propose that SAM homeostasis is maintained when cells are trying to balance between growth and survival, in response to environmental cues or cellular stresses (Figure 3).

Figure 3. Significance of SAM homeostasis.

SAM levels can be influenced by various external and internal factors. Excess levels of SAM triggers a growth & consumption state when cells activate substrate methylation and (m)TORC1 signaling to promote proliferation. Substrate methylation includes hypermethylation of certain epigenetic marks and the RNA-binding protein HuR. Insufficient levels of SAM switches cells to a survival state where autophagy is activated due to (m)TORC1 inhibition, and cell cycle arrest. Some epigenetic marks appear to be hypomethylated, as well as the cGAS enzyme involved in immune response. Whether these changes in methylation levels are relevant to the cellular switch between growth and survival modes remains to be established. In coordination with the switch between these two states, cells also balance SAM levels by controlling production (SAM synthesis) and consumption (transmethylation and the methyl-sink effect). Therefore, cells sense nutrients and tightly control SAM homeostasis, and use that as a cue to mediate cellular processes to optimize fitness.

External supplementation or genetic perturbations can cause accumulation of SAM [46–48], which, as described above, turns on mTORC1 through binding to SAMTOR, or TORC1 through promoting methylation of PP2A [32, 34]. (m)TORC1 signaling is a major driver of cell growth and proliferation. Therefore, high levels of SAM may be a cue for the cells to grow and divide, consuming SAM as for biosynthesis of cellular building blocks (e.g., polyamines), thereby reducing the levels of SAM. Production of additional SAM when SAM has accumulated in cells is also inhibited through multiple routes. First, expression of SAM synthetase will decrease, decreasing SAM synthesis (see also Figure 2) [8, 19]. Secondly, high levels of SAM inhibit MTHFR, thereby inhibiting production of methionine, a SAM precursor [4]. Thirdly, SAM activates CBS [5], directing flux into transsulfuration to produce cysteine, one of the twenty amino acids, as well as glutathione, an abundant reductant required for proliferation. High levels of SAM also leads to hypermethylation of methyltransferase (MTase) substrates. This includes histones, in particular H3K36me3, H3K79me3, and H3K4me3, PE, and nicotinamide that function as methyl sinks to absorb excess SAM in the cells [23] (see above). Together, these strategies aid in rebalancing of SAM concentrations when SAM levels are high.

On the other hand, low levels of SAM may be a result of starvation, drug treatment, or genetic perturbations [49–52]. This inhibits (m)TORC1 and promotes autophagy, a process that degrades organelles and macromolecules to recycle vital building blocks of the cell [34]. The recycled metabolites can flow into sulfur metabolism to synthesize SAM. In line with this, cells turn on expression of SAM synthetases (see also Figure 2) in response to low SAM levels [19]. In addition to roles in TORC1 regulation, hypomethylated PP2A promotes histone demethylation in yeast, freeing up the methyl sink capacity of histones [23] and preparing the cell nucleus for the next influx of SAM [53]. Finally, SAM limitation arrests cells at G1-phase in an mTORC1-independent manner [1]. Therefore, under SAM-limitation, cells halt growth and switch to a pro-survival mode, which allows metabolic reprogramming to restore intracellular SAM concentrations.

Taken together, cells are constantly adapting to fluctuating intracellular SAM levels. Maintenance of SAM homeostasis is achieved while switching between a growth-promoting state or a survival state. Failure to do so may negatively impact fitness of the cell. For example, loss of Pbp1, a negative regulator of TORC1 in yeast leads to TORC1 de-repression and an elevated growth rate under methionine starvation. However, mutant pbp1Δ cells show decreased replicative health and survivability after prolonged culturing under the same condition [54].

Epigenetic or epitranscriptomic marks are also linked to SAM availability. CpG hypermethylation or hypomethylation occurs upon SAM treatment [46], knockout of MTHFR [49], or glutathione deficiency [55]. Decrease in histone methylations including H3K4me3, H3K9me3 and H3K27me3, and H3K79me2 [50, 51, 56–59], and mRNA N6-methyladenosine (m⁶A) [33, 45, 52] were observed under SAM-limited conditions in mammals. Changes in methylation modifications may be linked to SAM homeostasis and represent consequences of SAM imbalance, but a theme has yet to emerge for their functions. This may be due to the context-dependent nature of these modifications, but also points to a largely unexplored research area in linking methylation status to SAM homeostasis.

Given the aforementioned roles of SAM, it is not surprising that SAM homeostasis can have a significant influence on cell cycle, cancer, and aging. These connections are reviewed in depth elsewhere [1, 2].

Additional Perspectives and Considerations

SAM imbalance may be a common stress

There has been much effort in studying stress conditions, such as temperature stress, oxidative stress, hypoxia, etc. However, we propose that methionine/SAM imbalance is a physiologically relevant, commonly occurring, but understudied, nutritional stress condition. For example, (1) cells encounter oxidative stress in response to environmental insults or in the context of particular diseases (e.g., neurodegeneration), and may require increasing amounts of glutathione as an antioxidant [60, 61]. The synthesis of glutathione is dependent on methionine-derived cysteine [28, 62]. Therefore, cellular methionine (and SAM) levels may become limiting under oxidative stress or exposure to xenobiotic compounds that consume glutathione for detoxification. (2) Arsenic is an environmental toxin and chemotherapy drug, also used intensively in research to induce RNA granules. Metabolism and detoxification of arsenic and other heavy metals often consumes SAM and glutathione, which together can reduce SAM levels [63, 64]. In line with the notions above, treatments that increase SAM levels attenuate oxidative stress in neurodegenerative conditions or arsenic-induced NAFLD in rodent models [65, 66]. (3) Genetic disorders that alter one-carbon metabolism such as polymorphisms in the MTHFR gene cause SAM imbalance [67]. (4) Microbes and viruses, including the coronavirus, often scavenge SAM for capping of their own transcripts [68, 69], which may lead to low levels of host cellular SAM. (5) Over-the-counter SAM supplements are available, whereas consequences of dietary supplementation of SAM are largely undetermined or unexpected [70, 71]. Prescription medicines such as metformin have been shown to alter SAM levels. Differing reports suggest both increase and decrease in SAM levels via metformin administration, calling for more investigations [72, 73]. (6) Some tissues may also exhibit extraordinarily high requirements or demands for SAM, such as in the brain, where it is required for synthesis of particular neurotransmitters and methylation of abundant proteins. Lysosomal dysfunction or vitamin deficiencies may also lead to imbalances in methionine and SAM, as vitamin B₁₂/cobalamin, a cofactor required for the activity of methionine synthase, depends on lysosomes for proper intracellular trafficking [74].

Sensing of SAM through methylation modifications

Several examples discussed above show that methyltransferases and methylation modifications are responsive to SAM concentrations. SAM sensitivity of a methyltransferase may largely depend on its affinity to SAM (i.e., K_M). Table 1 shows reported K_M values of characterized methyltransferases, some of which show a range of values as reported by different groups. Interestingly, many of these enzymes have K_M’s greater than the intracellular SAM concentration in mammalian cells and yeast (~10 μM, with variable concentrations in mammalian cells) [23]. These methyltransferases would be sensitive to changes in SAM levels. Their connections to cellular adaptation upon SAM fluctuation are worth investigating. Beyond methylation, SAM is also utilized as a versatile cofactor in other ways by enzymes, reviewed elsewhere [75], such as by radical-SAM enzymes to carry out other interesting chemical transformations.

Table 1.

K_M values of methyltransferases.

MTases	Substrate	Km (μM)	System	References
DNMT3a	DNA	0.2–2.56	murine, human	[84–86]
DNMT1	DNA	0.4–4.4	murine, human	[87–91]
DNMT3b	DNA	0.7	murine	[86]
AvN4CMT	DNA	150	rotifer	[92]
SMYD2	histone, protein	0.06–0.12	human	[93, 94]
EZH2	histone, protein	0.1–1.62	human	[90, 94–96]
PRMT4	histone, protein	0.21–3.1	human	[90, 94, 97]
SET7	histone, protein	0.22–2	human	[90, 94]
PRMT1	histone, protein	0.28–21	human	[90, 94, 97]
DOT1L	histone	0.38	human	[94]
MLL1	histone, protein	0.5–1	human	[90, 94]
G9a	histone, protein	0.53–31.7	murine, human	[90, 94, 98–100]
SUV29H1	histone, protein	0.56–11.78	human	[90, 94, 98]
PRMT5	histone, protein	0.6–1	human	[90, 94, 97]
SUV29H2	histone, protein	0.74–9.85	human	[90, 94, 98]
MLL3	histone	0.85–1	human	[90, 94]
PRMT7	histone, protein	1.1–8.8	human	[97]
SETMAR	histone, protein	1.13	human	[94]
EZH1	histone, protein	1.24–2	human	[90, 94]
LCMT1	protein	1.25–1.3	porcine, human	[101]
PRMT6	histone, protein	1.8–3.2	human	[90, 94, 97]
NSD1	histone, protein	2	human	[102]
PRMT3	histone, protein	2–28.3	human	[94, 97]
PRMT8	histone, protein	2.2	human	[97]
GLP	histone, protein	2.21–13.51	human	[94, 98]
MLL2	histone, protein	3.17–5	human	[90, 94]
NSD3	histone, protein	3.6	human	[102]
NSD2	histone, protein	3.7	human	[102]
SET8	histone, protein	16.3	human	[94]
PRMT9	protein	37.7–40.5	human	[97]
PRDM9	histone, protein	60–240	human	[103]
METTL3/METTL14	RNA	0.11–0.23	human	[76, 104, 105]
Cov-2-Nsp14	RNA	0.25–3.5	viral	[106–109]
PCIF1	RNA	0.7–0.9	human	[110]
METTL5-Trm112	RNA	1	human	[110]
TkTrm10	RNA	3.0–6.3	archaea	[111]
ScTrm9	RNA	4.9	yeast	[112]
ScAbd1	RNA	6	yeast	[113]
RNMT1-RAM	RNA	11.4	human	[114]
EcEcm1	RNA	25	parasite	[115]
METTL16	RNA	132–400	human	[110, 116]
CeMETT10	RNA	551	worm	[117]
NNMT	nicotinamide	1.8–24	human	[118]
PEMT	PE	18.2	rat	[119]
GNMT	glycine	36	rat	[120]

Local production and compartmentalization of SAM

As with other cellular metabolites, it will be challenging but pertinent to consider the subcellular distribution of SAM. For example, the mammalian RNA m⁶A methyltransferase METTL3-METTL14 complex has high affinity for SAM (K_M ~ 100 nM) [76], which is lower than the intracellular SAM concentrations [23]. This is paradoxical considering the reported sensitivity of m⁶A levels to SAM homeostasis [45, 52], but could potentially be explained by local concentrations of freely available SAM. Differential concentrations of SAM in organelles may also be predicted due to the existence of SAM transporters on the mitochondrial membrane [77] and due to methyl-sink molecules that are located in different compartments of cells (e.g., histones, phospholipids and metabolites). In line with this, overexpression of solute carrier family 25 member 26 (SLC25A26), a carrier protein that imports SAM into the mitochondria, leads to increased mitochondrial SAM levels and hypermethylation of mitochondrial DNA in cervical cancer cells [78]. In yeast, a mitochondrial SAM transporter, Sam5, is critical for mitochondria to sense SAM levels and switch between an energetic and biosynthetic role upon methionine restriction [79]. Furthermore, recent studies in C. elegans suggest that different SAM synthetase enzyme isoforms may play different roles in provisioning SAM for histone methylation modifications such as H3K4me3 [80].

Studies on the in vitro properties of SAM reveal that SAM is highly unstable, subject to epimerization and degradation reactions and suggest that SAM is most stable in the presence of slightly acidic pH and in the presence of non-nucleophilic counteranions [81]. Consistent with these observations, an elegant study revealed that expression of a SAM-insensitive MTHFR in yeast resulted in more than a 100-fold hyperaccumulation of SAM and dependencies on both a vacuolar protein (Vps33) and polyphosphate accumulator (Vtc1) for viability [48]. Importantly, these studies implicate a role for vacuoles and possibly lysosomes in intracellular SAM homeostasis.

Concluding remarks

In summary, as a versatile and expensive metabolite cofactor second only to ATP in terms of utilization (~1% of proteomes encode SAM-dependent methyltransferase enzymes), cells employ a variety of homeostatic mechanisms to help ensure that intracellular SAM levels are neither too high nor too low. Future research will undoubtedly reveal additional SAM-dependent processes or modifications that are tuned to the availability of this sentinel metabolite and measure of the metabolic state (see Outstanding questions).

Outstanding Questions.

1. Is there a unifying model for various DNA and RNA methylation modifications in SAM homeostasis? DNA and RNA methylations can be responsive to intracellular SAM levels. Outside of MAT2A m⁶A methylation, their potential functions in SAM homeostasis have yet to be fully investigated.

2. What regulates the flux of one-carbon units between methionine/SAM versus nucleotides? What are the regulators of transsulfuration versus re-methylation of homocysteine?

3. What are the consequences of methionine supplementation or restriction? Methionine restriction, accompanied by the reduction of SAM levels, is shown to increase longevity in several species [82, 83], on the other hand, SAM is also a potential therapeutic supplement to treat diseases (reviewed in [1, 2]). The contributions of SAM homeostasis in these scenarios are unclear.

HIGHLIGHTS.

1. Synthesis of methionine may be prioritized over cysteine and nucleotides, and the methionine cycle may be prioritized over transsulfuration and nucleotide synthesis during cell survival. This may be due to the central role of methionine in protein synthesis and SAM in sustaining methylation potential of key macromolecules at all times, whereas pathways such as transsulfuration and nucleotide synthesis are irreversible and especially important during growth and proliferation.

2. SAM synthetase MAT2A expression in mammals is tightly controlled by m⁶A methyltransferase METTL16 via multi-layered regulation of pre-mRNA splicing and RNA stability, emphasizing the importance of SAM as the primary methyl donor synthesized using the essential amino acid methionine and the cellular energy currency ATP.

3. SAM homeostasis is, in part, maintained by transmethylation substrates acting as methyl-sinks; these include phospholipids, metabolites, and histones. Histone methylation status may, at least in part, correlate with its function as methyl-sink instead of local transcriptional activities.

4. SAM homeostasis is achieved in coordination with a switch between growth state versus survival state of the cell.

5. 5Many methyltransferases in cells may be sensitive to SAM limitation, which can represent an overlooked mode of nutritional stress.

Glossary

Autophagy

a “self-eating” process that degrades cellular components through lysosomes. The degradation products, such as amino acids, can be reused by the cell.

Epitranscriptome

the collection of over 300 chemical modifications found on RNA molecules in the cell.

Methionine salvage

a series of reactions that recycle the sulfur from 5'-methylthioadenosine, a byproduct of polyamine synthesis, to regenerate methionine.

Methyl-sink

molecules that serve as bulk methyl acceptors, allowing the conversion of SAM to SAH in the methionine cycle, proposed to govern SAM homeostasis when needed.

One-carbon metabolism

biochemical reactions centered around the folate and methionine cycles that enable cells to utilize one-carbon units for biosynthetic pathways and substrate methylation.

RNA granules

cellular structures formed by nucleation of RNAs and proteins that are microscopically visible.

Riboswitches

non-coding and structured RNA elements that directly bind specific ligands and regulate gene expression in response to the concentration of ligands. They are primarily found in bacteria.

Transsulfuration

a metabolic pathway that allows transfer of sulfur from homocysteine to cysteine through cystathionine.