S-adenosyl-L-homocysteine hydrolase and methylation disorders: yeast as a model system.

S-adenosyl-L-methionine (AdoMet)-dependent methylation is central to the regulation of many biological processes: more than 50 AdoMet-dependent methyltransferases methylate a broad spectrum of cellular compounds including nucleic acids, proteins and lipids. Common to all AdoMet-dependent methyltransferase reactions is the release of the strong product inhibitor S-adenosyl-L-homocysteine (AdoHcy), as a by-product of the reaction. S-adenosyl-L-homocysteine hydrolase is the only eukaryotic enzyme capable of reversible AdoHcy hydrolysis to adenosine and homocysteine and, thus, relief from AdoHcy inhibition. Impaired S-adenosyl-L-homocysteine hydrolase activity in humans results in AdoHcy accumulation and severe pathological consequences. Hyperhomocysteinemia, which is characterized by elevated levels of homocysteine in blood, also exhibits a similar phenotype of AdoHcy accumulation due to the reversal of the direction of the S-adenosyl-L-homocysteine hydrolase reaction. Inhibition of S-adenosyl-L-homocysteine hydrolase is also linked to antiviral effects. In this review the advantages of yeast as an experimental system to understand pathologies associated with AdoHcy accumulation will be discussed.

AdoMet — a principal methyl group donor and more

Beyond its role in protein synthesis and structure, methionine, after its activation to AdoMet by methionine adenosyltransferase, plays a crucial role in many aspects of cellular metabolism. AdoMet is synthesized in virtually all living organisms [1]. It is the second most widely used enzyme substrate after ATP [2] and, thus, one of the most versatile biomolecules. Being a high-energy sulfonium compound AdoMet can serve as a source of all three ligands linked to the sulfur atom. AdoMet is the methyl group donor used by all organisms in most methyl transfer reactions [3]. It was also shown to be used in a number of other reactions, serving as a source of propylamine for the synthesis of spermidine and spermine, amino groups for the synthesis of biotin, or ribosyl groups in the synthesis of queuosine, a hypermodified tRNA, as well as 5′ deoxyadenosyl radicals required for a number of radical reactions [3], [4].

The fact that under normal conditions more than 90% of total AdoMet in mammalian cells is used for methylation reactions catalyzed by AdoMet-dependent methyltransferases in which AdoMet donates its methyl group to a large variety of acceptors including nucleic acids, proteins and lipids underscores the importance of methylation in downstream reactions of AdoMet [3], [5]. AdoHcy, which is released as a by-product of AdoMet-dependent methyl transfer reactions, is a strong product inhibitor of AdoMet-dependent methyltransferases and is hydrolyzed by S-adenosyl-L-homocysteine hydrolase to homocysteine and adenosine (Fig. 1a). In mammals, homocysteine produced by this reaction can be remethylated to methionine by methionine synthase or betaine–homocysteine methyltransferase and, thus, retained in the methylation cycle. Alternatively, it can be withdrawn from the methylation cycle by its conversion to cysteine, the precursor of glutathione, via the two-step transsulfuration pathway. Cystathionine β-synthase is the first and rate-limiting enzyme of the latter. Both cystathionine β-synthase and the enzyme catalyzing the second step of the transsulfuration pathway, cystathionine γ-lyase, are vitamin B6 requiring enzymes.

Fig. 1

AdoMet-dependent methylation: the role of AdoHcy and S-adenosyl-L-homocysteine hydrolase a) in yeast and b) in mammals. Reverse transsulfuration pathway in yeast converts cysteine into homocysteine via cystathionine γ-synthase, Str2, and cystathionine β-lyase, Str3. AdoMet, S-adenosyl-L-methionine; AdoHcy, S-adenosyl-L-homocysteine; Hcy, homocysteine; Met, methionine; Sah1, S-adenosyl-L-homocysteine hydrolase in yeast; AHCY, S-adenosyl-L-homocysteine hydrolase in mammals; CTT, cystathionine; Sam1, AdoMet synthetase 1; Sam2, AdoMet synthetase 2; Sam4, AdoMet-homocysteine methyltransferase; Mht1, S-methylmethionine–homocysteine methyltransferase; Met6, methionine synthase; Met25, O-acetylhomoserine sulfhydrylase; Str1, cystathionine γ-lyase; Str2, cystathionine γ-synthase; Str3, cystathionine β-lyase; Str4, cystathionine β-synthase; Gsh1, γ-glutamylcysteine synthetase; MAT, methionine adenosyltransferase; MS, methionine synthase; BHMT, betaine-homocysteine methyltransferase; CBS, cystathionine β-synthase; CTH, cystathionine γ-lyase; GCS, γ-glutamylcysteine synthetase.

AdoMet is synthesized by methionine adenosyltransferase from ATP and methionine. In mammals, methionine can be either ingested in the diet or formed by methylation of homocysteine using either a methyl group of betaine or that of 5-methyltetrahydrofolate. The latter is formed by methylenetetrahydrofolate reductase from 5,10-methylenetetrahydrofolate in a cobalamin-dependent reaction [5]. Deficiency of methionine synthase that catalyzes remethylation of homocysteine to methionine using 5-methyltetrahydrofolate links homocysteine and folate metabolism leading to accumulation of methyltetrahydrofolate at the expense of the methylenetetrahydrofolate and tetrahydrofolate pools required for thymidylate and purine biosynthesis [6].

The enzymes required for AdoMet synthesis, AdoHcy catabolism, homocysteine remethylation and transsulfuration are conserved between yeast and mammals except for betaine–homocysteine methyltransferase and the reverse transsulfuration pathway (Fig. 1a, b). A further difference between sulfur metabolism in mammals and in yeast is the presence of the sulfur assimilation pathway in the latter (Fig. 1b); however, this pathway can be efficiently eliminated in yeast by the introduction of a mutation disrupting the pathway [7]. Deficiencies in the enzymes involved in the methylation cycle or homocysteine catabolism e.g. cystathionine β-synthase, methionine adenosyltransferase, methylenetetrahydrofolate reductase and S-adenosyl-L-homocysteine hydrolase show that many of them are linked to a number of pathologies including deregulation of lipid metabolism [7], [8], [9], [10], [11], [12], [13], neurological abnormalities [10], [14], [15], [16], tumorigenicity [8], [9], [11], oxidative stress [8] and myopathy [14].

AdoMet-dependent methyltransferases

Computational analysis of the human proteome revealed 208 known and putative human AdoMet-dependent methyltransferases that make up about 0.9% of genes in the human genome [17]. Noteworthy, 30% of these proteins have been linked to disease states e.g. cancer and mental disorders [17]. Analysis of the yeast proteome revealed 81 known and putative AdoMet-dependent methyltransferases, making up about 1.2% of genes in the yeast genome [18]. The increased number of AdoMet-dependent methyltransferases in humans as compared to yeast reflects partial redundancy as well as novel functions present in the human methyltransferasome in comparison with yeast [17]. AdoMet-dependent methyltransferases such as DNA-, glycine-, guanidinoacetate- and isoaspartyl methyltransferases are found only in humans. However, sterol-24C-methyltransferase, Erg6p, is found only in yeast. Nevertheless, the majority of AdoMet-dependent methyltransferases are conserved between yeast and humans. In Table 1, yeast and human AdoMet-dependent methyltransferases grouped by their substrate specificity are compared. The whole list of known and putative yeast and human AdoMet-dependent methyltransferases can be found in [18], [17], respectively.

Table 1

Comparison of yeast and human AdoMet-dependent methyltransferases grouped by substrate specificity. The whole list of known and putative yeast and human AdoMet-dependent methyltransferases can be found in [18], [17], respectively.

AdoMet-dependent methyltransferases	Biological function/process	Yeast	Mammalian orthologs
DNA
DNA (cytosine-5-)-methyltransferases	Gene silencing	–	DNMT1, DNMT2, DNMT3a, DNMT3b
RNA
mRNA (guanine-N7-)-methyltransferases	Methylation of the 5′ cap structure of mRNA	Abd1	RNMT
mRNA (adenosine-N6)-methyltransferases	Entry into meiosis	Ime4	METTL3
Trimethylguanosine synthases	Hypermethylation of m(7)G to the m(2,2,7)G 5′ cap of snRNAs, snoRNAs and telomerase TLC1 RNA	Tgs1	TGS1
Probable rRNA (cytosine-5-)-methyltransferases	27S pre-rRNA processing and 60S ribosome biogenesis	Nop2	NOP2
rRNA (guanine-N7)-methyltransferases	rRNA processing	Bud23	WBSCR22
rRNA (adenine-N6,N6-)-dimethyltransferases	18S pre-ribosomal rRNA processing	Dim1	DIMT1, TFB1M
rRNA (2′-O-ribose)-methyltransferases	27S pre-rRNA and 25S rRNA processing, 60S ribosomal subunit maturation	Spb1	FTSJ1, FTSJ2, FTSJ3
Mitochondrial rRNA (2′-O-ribose)-methyltransferases	Mitochondrial 21S rRNA processing	Mrm1, Mrm2	MRM1
tRNA (guanine)-methyltransferases	tRNA modification	Trm1, Trm5, Trm10, Trm8/Trm82, Trm11/Trm112, Trm12	TRMT1, TRMTL1, TRMT5, RG9MTD2, METTL1, TRMT11, TRMT12
tRNA (uracil-5-)-methyltransferases	tRNA modification	Trm2, Trm9	TRMT2A, ALKBH8, KIAA145
tRNA (cytosine-5-)-methyltransferases	tRNA modification	Ncl1 (Trm4)	NSUN2, NSUN3, NSUN5
tRNA (adenine-N1-)-methyltransferases	Maturation of initiator methionyl-tRNA	Trm6/Trm61	TRMT6
tRNA (2′-O-ribose)-methyltransferases	tRNA modification	Trm3, Trm7, Trm13, Trm44	TARBP1, FTSJ1, CCDC76
tRNA methyltransferases	Methylation of N-4 position of yW-86 in wybutosine biosynthesis	Tyw3	TYW3
Carboxyl methyltransferases	Methylation of the α-carboxy group of yW-72 in wybutosine biosynthesis	Ppm2	LCMT2
Proteins
Histone lysine N-methyltransferases	Transcriptional elongation and silencing	Set1, Set2, Dot1	SET superfamily consisting of SUV39, SET1, SET2, RIZ, SMYD, EZ and SUV4-20 families, SET7/9, SET8, DOT1L
Ribosomal protein lysine N-methyltransferases	Ribosome biogenesis	Rkm1, Rkm2, Rkm3, Rkm4, Rkm5	SETD6
ω-NG-monomethylarginine and asymmetric ω-NG,NG-dimethylarginine methyltransferases	Methylation of hnRNPs affecting their activity and nuclear export; methylation of U1 snRNP protein Snp1p and ribosomal protein Rps2p	Hmt1 (Rmt1)	PRMT1, PRMT3, PRMT4 (CARM1), PRMT6, PRMT8
ω-NG-monomethylarginine and symmetric ω-NG,NG-dimethylarginine methyltransferases	Recruitment and inactivation of Swe1 (Wee1 ortholog) during mitotic entry	Hsl7	PRMT5 (JBP1), PRMT7
δ-NG-monomethylarginine methyltransferase	Methylation of ribosomal protein Rpl12	Rmt2	–
Protein histidine methyltransferases	3-Methylhistidine modification of ribosomal protein Rpl3p	Hpm1	METTL18
N-terminal methyltransferases	Methylation of ribosomal proteins Rpl12 and Rps25	Tae1	METTL11A
C-terminal leucine carboxyl methyltransferases	Methylation of C-terminal leucine of PP2A catalytic subunit, complex formation	Ppm1	LCMT1
Protein S-isoprenylcysteine O-methyltransferases	C-terminal methylation of CAAX proteins	Ste14	ICMT
Protein-l-isoaspartate (d-aspartate) O-methyltransferase	Repair of damaged proteins	−	PCMT1
Cytochrome c lysine N-methyltransferase	Trimethylation of cytochrome c, not required for respiratory growth	Ctm1	−
eRF1 methyltransferases	Peptidyl-glutamine methylation, translation release	Mtq2	N6AMT1
Mrf1methyltransferases	Peptidyl-glutamine methylation, mitochondrial translation release	Mtq1	HEMK1
Lipids
Phospholipid N-methyltransferases	Phosphatidylcholine de novo synthesis	Cho2, Opi3	PEMT
Sterol 24-C-methyltransferase	Ergosterol biosynthesis	Erg6	–
Others
S-AdoMet-homocysteine S-methyltransferase	Regulation of methionine/AdoMet ratio	Sam4	–
Trans-aconitate methyltransferase	Leucine biosynthesis	Tmt1	–
Uroporphyrin-III C-methyltransferase	Siroheme biosynthesis	Met1	–
Glycine/sarcosine N-methyltransferase	Regulation of hepatic AdoMet metabolism	–	GNMT
Guanidinoacetate N-methyltransferase	Creatine synthesis	–	GAMT
Ubiquinone methyltransferases	Ubiquinone biosynthesis, respiration	Coq3, Coq5	COQ3, COQ5
Phenylethanolamine N-methyltransferase	Adrenaline (epinephrine) biosynthesis	–	PNMT
Acetylserotonin O-methyltransferase	Melatonin biosynthesis	–	ASMT
Histamine N-methyltransferase	Histamine degradation	–	HNMT
Catechol O-methyltransferase	Degradation of catecholamines	–	COMT
Indolethylamine N-methyltransferase	Tryptamine methylation	–	INMT
Nicotinamide N-methyltransferases	Nicotinamide metabolism	Nnt1	NNMT

The largest group of AdoMet-dependent methyltransferases methylates proteins, predominantly at the ε-amine group of lysine and ω- or δ-guanidine groups of arginine as well as on the C-terminal leucine and isoprenylated cysteine residues. Targets of AdoMet-dependent protein methyltransferases include histones, ribosomal proteins, transcription and translation factors, signal transduction proteins etc. [19], [20], [21], [22]. For instance, the SET domain protein lysine methyltransferase family, which makes up 27% of human methyltransferasome and 14% of yeast methyltransferasome, methylates predominantly histones at lysine residues [17], [23]. Comparison of a number of histone lysine methyltransferases from various organisms ranging from yeast to human underlines the similarity of their roles in the maintenance of normal cellular functions in eukaryotes [24].

Protein arginine methyltransferases methylate in addition to histones [19], [20] numerous non-histone proteins [20], [22], [25] both in yeast and in mammals, including heterogeneous nuclear riboproteins (hnRNPs) required for eukaryotic transcription elongation, mRNA processing and export [26], [27], [28]. Further targets for AdoMet-dependent methylation are the protein phosphatase 2A (PP2A), which is methylated at the highly conserved C-terminal leucine of its catalytic C subunits [29], and G proteins, in particular Ras proteins, that are methylated at C-terminal isoprenylcysteine residues [30].

A big group of AdoMet-dependent methyltransferases methylate nucleic acids. However, while many AdoMet-dependent methyltransferases are involved in the methylation of mRNA, rRNA and tRNA both in yeast and in mammals, there are no orthologs to mammalian DNA (cytosine-5)-methyltransferases that methylate DNA in yeast. While methylation of rRNA is crucial for ribosomal processing and maturation [31], methylation of tRNA helps to control tRNA folding and to ensure decoding specificity and efficiency [32]. Cap methylation at the 5′-terminal guanosine by mRNA (guanine-7-)-methyltransferase is an essential process both in yeast and mammals [33], [34].

In addition to proteins and nucleic acids AdoMet-dependent methyltransferases methylate lipids and a number of small molecules, in particular, phosphatidylethanolamine, ubiquinone and nicotinamide (Table 1). It is noteworthy that de novo biosynthesis of phosphatidylcholine from phosphatidylethanolamine catalyzed by AdoMet-dependent phospholipid methyltransferases is the major AdoMet consumer in mice and humans [35], [36] as well as in yeast in the absence of choline/ethanolamine supplementation.

The role of S-adenosyl-L-homocysteine hydrolase in AdoMet-dependent methylation

AdoHcy that is released as a by-product of AdoMet-dependent methyltransferase reactions, is a potent product inhibitor of most AdoMet-dependent methyltransferases with Ki values in the submicromolar to low micromolar range [37]. AdoHcy was shown to inhibit a number of AdoMet-dependent methyltransferases in vitro, such as mammalian DNA (cytosine-5-)-methyltransferase, mRNA cap (guanine-N7-)-methyltransferase, tRNA (uracil-5-)-methyltransferase, protein isoprenylcysteine carboxylmethyltransferase and phospholipid methyltransferases [38], [39], [40], [41], [42], [43], [44], [45], and in vivo, such as DNA (cytosine-5-)-methyltransferase and phospholipid methyltransferases [7], [46], [47]. However, not all AdoMet-dependent methyltransferases exhibit equal sensitivity against AdoHcy. While some AdoMet-dependent methyltransferases have a Ki value for AdoHcy that is lower than the Km value for AdoMet and are especially sensitive to AdoHcy [37], others, in particular glycine N-methyltransferase, which regulates AdoMet levels in mammals [1], are only weakly inhibited by AdoHcy [48]. Being a potent product inhibitor for most AdoMet-dependent methyltransferases, AdoHcy is a sensitive marker for the cellular methylation status that is reflected by AdoMet/AdoHcy ratio, and can be used to predict methylation deficiency e.g. DNA hypomethylation [49].

The only eukaryotic enzyme capable of AdoHcy catabolism both in yeast and mammals is S-adenosyl-L-homocysteine hydrolase (Sah1 in yeast, AHCY in mammals), which catalyzes the reversible hydrolysis of AdoHcy to homocysteine and adenosine. Similarly to AdoHcy accumulation, interference with S-adenosyl-L-homocysteine hydrolase was shown to affect methylation of different targets of AdoMet-dependent methyltransferases. In particular, inhibition of S-adenosyl-L-homocysteine hydrolase and the resulting increase of AdoHcy levels were shown to be associated with DNA hypomethylation in cultured endothelial cells [46] and to interfere with RNA methylation in cultured lymphocytic leukemia cells [50]. Furthermore, inhibition of S-adenosyl-L-homocysteine hydrolase was also shown to result in decreased methylation of poly(A)(+) RNA in Xenopus laevis oocytes [51] as well as in inhibition of phospholipid methylation in yeast [7]. In accordance with the crucial role of AdoMet-dependent methylation in many biological processes, accumulation of AdoHcy and/or S-adenosyl-L-homocysteine hydrolase dysfunction is linked to numerous pathologies including neurological and vascular disorders, myopathy, fatty liver, cancer, renal insufficiency and diabetic nephropathy [11], [14], [52], [53], [54], [55], [56], [57], [58]. The pathological consequences of AdoHcy accumulation as well as the molecular mechanisms triggered by elevated AdoHcy levels are largely not understood. Considering the role of AdoMet-dependent methylation in biological processes, inhibition of a number of AdoMet-dependent methyltransferase reactions appears to play a central role in AdoHcy-associated pathology. However, evidence of in vivo sensitivity of individual AdoMet-dependent methyltransferase reactions as well as the responsiveness of methylation-dependent biological processes to AdoHcy accumulation is limited. Since S-adenosyl-L-homocysteine hydrolase is a crucial physiological regulator of AdoHcy levels we will discuss dysfunction of S-adenosyl-L-homocysteine hydrolase, its evolutional conservation, structural organization and regulation in the next sections.

S-adenosyl-L-homocysteine hydrolase dysfunction

Deletion of a locus overlapping the AHCY gene in mouse results in embryonic death [59]. Homozygous insertion mutations in the S-adenosyl-L-homocysteine hydrolase gene in Arabidopsis thaliana result in zygotic lethality [60]. In contrast, yeast mutants lacking Sah1 are viable, due to the presence of an alternative pathway for homocysteine synthesis via the sulfur assimilation. Homocysteine that is synthesized by this pathway is further used both for the synthesis of cysteine and glutathione as well as for the synthesis of methionine and AdoMet (Fig. 1) [7]. Introduction of an additional mutation in the sulfur assimilation pathway indeed renders the resulting yeast sah1 mutants inviable [7], consistent with an essential function of S-adenosyl-L-homocysteine hydrolase.

S-adenosyl-L-homocysteine hydrolase deficiency in humans, a rare genetic disorder, is characterized by up to 150-fold elevated plasma AdoHcy levels with a significant decrease in the AdoMet/AdoHcy ratio and severe pathological consequences that can lead to death during childhood [14], [61], [62], [63], [64]. In these patients, three tissues are predominantly affected: muscles, brain and liver, manifested by severe myopathy, slow myelination, developmental delay [61], [62], [63] as well as mild hepatitis [61] and lipid droplet accumulation in the liver [63]. Biochemical analyses revealed that AHCY deficient patients display 3–20% of mean control S-adenosyl-L-homocysteine hydrolase activity in the liver, red blood cells and fibroblasts [61], [62], [63]. They exhibit a decrease of phosphatidylcholine, choline and albumin levels as well as an elevation of creatine kinase, alanine transferase, aspartate transaminase [61], [62], [63] and lactate dehydrogenase [63] levels in the plasma. At the cellular level these patients show a proliferation of smooth and a decrease of rough endoplasmic reticulum [61], altered mitochondria [61], [63] as well as hypermethylation of leukocyte DNA [61], [62], [63].

S-adenosyl-L-homocysteine hydrolase: conservation, structure and catalytic mechanism

S-adenosyl-L-homocysteine hydrolase is an exceptionally well-conserved enzyme exhibiting over 70% identity at the protein level between human and yeast orthologs, and is among the 90 most highly conserved yeast proteins, including actin (90%), ubiquitin (90%), histones (70–90%), ribosomal (70%) and heat shock proteins (70%) [65]. Fig. 2 displays a comparison of S-adenosyl-L-homocysteine hydrolase sequences from Archaea, Bacteria and Eucarya. Emphasizing the similarity of yeast and human orthologs, many bacterial and some eukaryotic sequences, but not the mammalian and yeast proteins, exhibit an insertion segment of 40 amino acids of as yet unknown function in the catalytic domain of the enzyme [66].

Fig. 2

S-adenosyl-L-homocysteine hydrolase: sequence and structural conservation. The sequence conservation based on ConSurf calculations [160] using 166 unique S-adenosyl-L-homocysteine hydrolase sequences is color coded (from blue – identity to red – least conservation) and mapped onto the structure of the Pf-SAHH (PDB 1v8b [77]). Shown is one monomer of the functional tetrameric S-adenosyl-L-homocysteine hydrolase protein in cartoon drawing, the NAD+ cofactor (yellow) is shown in a stick representation and the substrate (Ado) was omitted for clarity. Note that the 40 amino acid insertion consisting of 3 helices and a loop (black circle), which is not present in mammalian and yeast S-adenosyl-L-homocysteine hydrolases, shows the least degree of conservation.

S-adenosyl-L-homocysteine hydrolase belongs to the large family of NAD(P)H/NAD(P)+-binding proteins that share a Rossmann-fold [67]. The NAD(P)H/NAD(P)+ binding domain is found in numerous dehydrogenases as well as in many other redox enzymes, but is rather unusual for a hydrolase. The atomic structures of S-adenosyl-L-homocysteine hydrolase from different sources have been resolved: Homo sapiens (PDB 1a7a [68], PDB 1li4 [69] and PDB 3nj4 [70]), Rattus norvegicus (PDB 1b3r [71], PDB 1ky4 and PDB 1ky5 [72], PDB 1k0u [73], PDB 2h5l [74], PDB 1xwf [75] and PDB 1d4f [76]), Plasmodium falciparum (PDB 1v8b [77]), Mycobacterium tuberculosis (PDB 3ce6, PDB 3dhy, PDB 2zj1, PDB 2zj0 and PDB 2ziz [78]), Burkholderia pseudomallei (PDB 3d64), Trypanosoma brucei (PDB 3h9u) and Leishmania major (PDB 3g1u) and the plant Lupinus luteus (PDB 3ond, PDB 3one and PDB 3onf [79]).

All structurally characterized Sah1/AHCY proteins except plant S-adenosyl-L-homocysteine hydrolase are tetramers with NADH/NAD+ cofactor bound in the active site of each subunit [80], [81]. Plant S-adenosyl-L-homocysteine hydrolase from L. luteus has been shown to function as a homodimer [79]. The monomeric subunits of the protein contain three domains: N-terminal substrate-binding domain, cofactor-binding domain and C-terminal tail. The C-terminal tails of two subunits reciprocally protrude into opposite subunits and form a part of their cofactor-binding sites [68]. The two dimers then form a tetramer with the four cofactor-binding domains molding the central core of the tetramer structure, and the substrate-binding domains being exposed on the surface.

The thermodynamic equilibrium of the reaction catalyzed by S-adenosyl-L-homocysteine hydrolase favors the synthesis of AdoHcy from adenosine and homocysteine in vitro [82]. In vivo rapid enzymatic removal of homocysteine (Hcy) and adenosine (Ado) enables the net hydrolysis of AdoHcy [82], [83]. The reaction cycle requires reciprocal oxidation–reduction of the substrate and NAD+ [84]. In the first step, the Ado 3′ hydroxyl group of adenosine (synthetic reaction: Ado + Hcy → AdoHcy) or AdoHcy (hydrolytic reaction: AdoHcy → Ado + Hcy) is oxidized to ketone by NAD+ resulting in the formation of NADH (Fig. 3). In the next steps the proton is removed from C4′ forming the carbanion intermediate, followed by its cleavage and the release of water or Hcy, respectively. The catalytic cycle is finished by the addition of Hcy or water to the C4′C5′ double bond and reduction of the 3′ keto group under regeneration of NAD+, forming AdoHcy or Hcy, respectively. Depending on the presence of the substrate the protein undergoes large conformational rearrangements [71], [85], [86], [87]. Substrate (AdoHcy/Ado) binding induces a structural transition from the open to the closed form of the enzyme, and product (Ado/AdoHcy) release induces the transition back to the open conformation [85].

Fig. 3

Catalytic activity of S-adenosyl-L-homocysteine hydrolase.

S-adenosyl-L-homocysteine hydrolase: regulation and localization

S-adenosyl-L-homocysteine hydrolase requires NAD+ to maintain its quaternary structure and catalytic activity, making the enzyme sensitive to the redox status of the cell that is reflected by the NAD+/NADH ratio. As described above NAD+ bound to S-adenosyl-L-homocysteine hydrolase is subject to reduction–oxidation cycles during the reaction [84] (Fig. 3). Moreover, the form of NAD bound to the enzyme determines binding of adenosine, the substrate for the synthetic reaction and the inhibitor of the hydrolytic reaction. It was shown that S-adenosyl-L-homocysteine hydrolase from bovine kidney possesses two binding sites for adenosine: a low affinity binding site located in the catalytic domain and a high affinity binding site located in the NAD+/NADH binding domain of the enzyme [88]. The NAD+ form of S-adenosyl-L-homocysteine hydrolase binds adenosine with low affinity leading to the formation of 3′-keto-adenosine, reduction of the tightly bound NAD+ to NADH and adoption of the closed conformation by the enzyme [88]. However, the enzymatically inactive NADH form of the enzyme binds adenosine with high affinity, suggesting a role for adenosine in NADH-induced inhibition of S-adenosyl-L-homocysteine hydrolase and potential function of the enzyme as intracellular Ado-binding protein [88].

S-adenosyl-L-homocysteine hydrolase was also shown to bind copper with high affinity and is indeed the major copper binding protein in mouse liver [89], [90]. It has a similar dissociation constant for copper as albumin [89] and was suggested to serve as an intracellular copper transporter with a putative function in delivering copper to superoxide dismutase [91]. Copper inhibits S-adenosyl-L-homocysteine hydrolase activity in a non-competitive manner due to the release of NAD+ [92] that results from interruption of subunit interactions, which leads to a large decrease in the affinity for the NAD+ cofactor [92]. However, the significance of regulation of S-adenosyl-L-homocysteine hydrolase by NAD+, adenosine and copper is not clear.

In yeast S-adenosyl-L-homocysteine hydrolase is primarily found in the cytosol [93]. However, AdoMet-dependent methylation takes place in different cellular compartments e.g. nucleus and ER, suggesting that subcellular translocation of the enzyme may play a regulatory role. Indeed, S-adenosyl-L-homocysteine hydrolase was found in the nucleus in association with mRNA (guanine-7-)-methyltransferase and RNA polymerase II in transcriptionally active X. laevis oocytes [51], [94]. S-adenosyl-L-homocysteine hydrolase was also shown to translocate into the nucleus in adult mammalian cells under hypoxia conditions known to induce transcriptional activity [95]. Localization analyses in A. thaliana also demonstrated that the enzyme is capable of localizing to the cytoplasm and the nucleus [96], and was proposed to be targeted to the nucleus in a complex with adenosine kinase, another enzyme required for AdoHcy catabolism [96]. The S-adenosyl-L-homocysteine hydrolase/adenosine kinase complex is, presumably, targeted to the nucleus by the nuclear localization signal of mRNA (guanine-7-)-methyltransferase, with which it interacts [96].

AdoHcy in hyperhomocysteinemia

The pathological accumulation of homocysteine above the normal plasma homocysteine level of 10 μmol/L, hyperhomocysteinemia, can be caused by genetic or nutritional factors. Deficiency in cystathionine β-synthase (Fig. 1) or genetic defects of folate and cobalamin metabolisms lead to severe hyperhomocysteinemia characterized by the plasma homocysteine levels of more than 100 μmol/L [97] and are rare in comparison to mild hyperhomocysteinemia. Mild hyperhomocysteinemia, in turn, may be due to moderate deficiencies of key enzymes involved in homocysteine metabolism, such as the thermolabile variant of 5,10-methylenetetrahydrofolate reductase gene that has been reported in 5–15% of the general population [98]. Additionally, mild hyperhomocysteinemia may be a result of dietary deficiencies of vitamin cofactors required for homocysteine catabolism — folic acid, vitamins B6 and B12, and is characterized by the plasma total homocysteine levels of 15–25 μmol/L [99]. Hyperhomocysteinemia is linked to a number of disorders. It is recognized as a strong, independent and causal risk factor for cardiovascular and vascular diseases [100], [101], [102], [103]. In particular, it has been estimated that a 2.5 μM rise in homocysteine levels corresponds to a 10% increase in cardiovascular disease risk [104] and it was shown that about 40% of patients diagnosed with premature coronary artery disease, peripheral vascular disease or venous thrombosis exhibit hyperhomocysteinemia [105]. In addition to vascular diseases hyperhomocysteinemia is also associated with a number of other pathologies including neurological disorders, cancer, renal insufficiency, diabetes, birth defects and aging [55], [56], [104], [106], [107], [108], [109]. A number of potential mechanisms that are aimed at the explanation of the pathological consequences resulting from homocysteine accumulation were proposed [100], [105], [110]; however, there is still a controversy on the underlying metabolic connections [111], [112], [113].

Since certain vitamin deficiencies are characterized by mild hyperhomocysteinemia the possibility that vitamin supplementation could reduce/prevent cardiovascular events was studied in several large trials. However, it was observed that vitamins, while capable of lowering elevated plasma homocysteine levels, do not reduce the rates of vascular events [114]. This result may be due to the promotion of cell proliferation by folic acid through its role in the synthesis of thymidine, the increase of the methylation potential leading to changes in gene expression, or an increase in the levels of asymmetric dimethylarginine that inhibits the activity of nitric oxide synthase, offsetting any positive effect of homocysteine lowering [114]. An additional possibility is that a related metabolite rather than homocysteine could be a trigger of some pathological changes associated with elevated homocysteine levels. Since the thermodynamic equilibrium of the S-adenosyl-L-homocysteine hydrolase reaction favors the synthesis of AdoHcy [82], the accumulation of hydrolytic products of the reaction, in particular, homocysteine leads to the reversal of the physiological direction of the Sah1 catalyzed reaction, resulting in AdoHcy synthesis and its accumulation in vivo, as shown both in mammals as well as in yeast [7], [83], [115]. Indeed, patients with even mild hyperhomocysteinemia (95% CI values for total homocysteine 11.0–14.7 μmol/L) that overlap those of controls (95% CI values 9.8–12.2 μmol/L) exhibit significant elevation of plasma AdoHcy (95% CI values 32.3–47.7 nmol/L) compared to controls (95% CI values 24.9–30.7 nmol/L) [116]. Therefore, even mild hyperhomocysteinemia is characterized by an accumulation of AdoHcy, which may lead to a decreased methylation potential and the inhibition of a number of AdoMet-dependent methyltransferases and, thus, contribute to the pathology associated with homocysteine accumulation.

Notably, the mechanisms that have been proposed to explain the pathological changes associated with elevated homocysteine levels do not address the accumulation of AdoHcy and the inhibition of AdoMet-dependent methylation in homocysteine-associated pathology, with the exception of an inhibition of DNA methylation [117], [118], [119]. However, AdoHcy rather than homocysteine was shown to be a more sensitive indicator of cardiovascular disease [116], [120] as well as of renal insufficiency [56] suggesting that AdoHcy is indeed a crucial pathological factor in homocysteine-associated disorders. It was also reported that the supplementation with B-vitamins including folate does not efficiently lower plasma AdoHcy levels in contrast to homocysteine levels in elderly people with hyperhomocysteinemia [121], possibly explaining the failure of homocysteine lowering vitamins to reduce vascular events mentioned earlier. Furthermore, supporting the pathogenic role of AdoHcy, studies in yeast showed that indeed AdoHcy but not homocysteine is more toxic to cells that are deficient in homocysteine catabolism [122].

Induced AdoHcy accumulation as a strategy towards antiviral therapy

During the last three decades S-adenosyl-L-homocysteine hydrolase was intensively studied as a target for antiviral drug design [123]. Inhibitors that block the enzyme are efficient against many types of viruses, including Ebola, that are dependent, in particular, on mRNA methylation [123], [124], [125]. In addition to antiviral activity inhibitors targeted at S-adenosyl-L-homocysteine hydrolase show also other effects of pharmacological importance, for instance, antimalarial and immunosuppressive activity [126], [127]. Although AdoHcy accumulation may be responsible for a number of the effects of inhibitors targeted at S-adenosyl-L-homocysteine hydrolase, the ability of some of these inhibitors to undergo metabolic phosphorylation to nucleotides may account for a part of their biological activities, making it difficult to validate the mode of their action [128]. Furthermore, the interference with central metabolic pathways associated with high cytotoxicity is an additional hindrance in the utilization of S-adenosyl-L-homocysteine hydrolase inhibitors in the potential future clinical applications [125]. Nevertheless, suppression of viral replication in plants by controlled downregulation of S-adenosyl-L-homocysteine hydrolase confirms the antiviral activity as a result of AdoHcy accumulation [129].

Yeast as a model system to study methylation deficiency

Numerous disorders associated with AdoHcy accumulation suggest that this metabolite is capable of triggering multiple pathological mechanisms, presumably via inhibition of different AdoMet-dependent methyltransferases. However, the role of AdoHcy in these diseases as well as the AdoHcy-dependent molecular mechanisms contributing to the consequences inherent to these pathologies is poorly understood. Of course, understanding the responsiveness of methylation-dependent processes requiring AdoMet-dependent methyltransferases that are present only in mammals, e. g. protein L-isoaspartate methyltransferase, requires the use of a mammalian model system. However, due to the complexity of the methylation metabolism its dissection in a multicellular organism as well as in higher cells is rather difficult.

The yeast Saccharomyces cerevisiae is a unicellular eukaryote with an about 4 times lower number of genes compared to humans, but shares the complexity of the cellular architecture of higher cells. Yeast is especially amenable to experimentation, in particular for genetic manipulation and whole genome studies. As a result, this model system is prized with the highest genome annotation level and was successfully used to characterize a number of fundamental biological processes, including secretion, organelle biogenesis and cell cycle [130], [131], [132], [133]. Yeast exhibits a highly conserved methylation metabolism [17], [65] and as such is an advantageous system to understand fundamental toxicity of AdoHcy at the cellular level. For instance, yeast mutants capable of reproducible down-regulation of S-adenosyl-L-homocysteine hydrolase and, thus, in vivo modulation of AdoHcy levels can be used as a valuable tool to understand downstream mechanisms triggered by AdoHcy accumulation. Usage of a yeast mutant that is deficient in homocysteine remethylation to methionine, met6, provides another independent system to understand the role of AdoHcy. In this mutant, grown under homocysteine supplementation, elevated homocysteine leads to AdoHcy synthesis via the reversal of S-adenosyl-L-homocysteine hydrolase reaction resulting in an accumulation of the latter. However, in this mutant elevated homocysteine cannot be remethylated to methionine, impeding an additional increase in AdoMet levels and an elevation of the AdoMet/AdoHcy ratio. Further advantage of this system is that it uncouples the S-adenosyl-L-homocysteine hydrolase function in the regulation of AdoMet-dependent methylation from its potential effect on glutathione levels. Both systems were already shown to help in the identification and characterization of critical cellular processes that respond to AdoHcy accumulation [7]. In the next section the deregulation of lipid metabolism, the major consumer of AdoMet both in yeast and in mammals, will be discussed, as a special aspect of AdoHcy toxicity.

Role of AdoHcy in lipid metabolism

Changes at the epigenetic level are the most extensively studied consequences of methylation deficiency [117], [118], [119]. However, quantitatively the most important AdoMet-dependent methyltransferases take part in the primary metabolism. Phospholipid methylation requires three sequential AdoMet-dependent methylation steps to synthesize one molecule of phosphatidylcholine (PC) from phosphatidylethanolamine (PE) de novo. Whereas in yeast phospholipid methylation is the predominant way to synthesize PC (in particular, in the absence of choline/ethanolamine in the culture medium), phospholipid methylation in mammals that occurs primarily in the liver covers only 30% of hepatic PC synthesis and accounts for estimated 10 μmol and 1.65 mmol PC secreted into bile per day in mice and humans, respectively [36]. Phospholipid methylation is the major AdoMet consumer in yeast in the absence of choline/ethanolamine supplementation. Despite of the relatively low contribution of phospholipid de novo synthesis to PC production in mammals this reaction is also the major consumer of AdoMet in mice [35]. Reexamination of the methylation metabolism in humans also revealed that phospholipid methylation, but not creatine synthesis as was assumed previously, accounts for the major part of AdoMet being utilized in the human body [36].

AdoHcy in yeast lipid metabolism

The involvement of S-adenosyl-L-homocysteine hydrolase in the regulation of phospholipid biosynthesis in yeast was found by showing that yeast SAH1 is regulated at the transcriptional level in coordination with phospholipid biosynthesis [134]. S-adenosyl-L-homocysteine hydrolase controls phospholipid de novo methylation in yeast by regulating the hydrolysis of AdoHcy, which is a competitive inhibitor of Cho2 and Opi3 phospholipid methyltransferases that catalyze methylation of PE to PC [7], [45] (Fig. 4). However, not only phospholipid methylation, but also a methylation-independent branch of lipid metabolism is affected by AdoHcy accumulation in yeast: yeast cells deficient in AdoHcy catabolism massively accumulate triacylglycerols (TAG) [7]. Supporting the causal role of impaired phospholipid methylation in the deregulation of TAG metabolism in response to AdoHcy accumulation, it was found that yeast mutants that are deficient in the enzymatic activities required for methylation of PE to PC, cho2 and opi3, also accumulate TAG [7]. Additionally, homocysteine supplementation in yeast that leads to a 10-fold accumulation of AdoHcy results in the inhibition of phospholipid methylation and accumulation of TAG, confirming a crosstalk between homocysteine and lipid metabolism [7].

Fig. 4

Role of AdoHcy and S-adenosyl-L-homocysteine hydrolase in lipid metabolism in yeast. AdoMet, S-adenosyl-L-methionine; AdoHcy, S-adenosyl-L-homocysteine; Hcy, homocysteine; Met, methionine; Sah1, S-adenosyl-L-homocysteine hydrolase; CTT, cystathionine; Sam1, AdoMet synthetase 1; Sam2, AdoMet synthetase 2; Sam4, AdoMet-homocysteine methyltransferase; Mht1, S-methylmethionine–homocysteine methyltransferase; Met6, methionine synthase; Met25, O-acetylhomoserine sulfhydrylase; Str1, cystathionine γ-lyase; Str2, cystathionine γ-synthase; Str3, cystathionine β-lyase; Str4, cystathionine β-synthase; Gsh1, γ-glutamylcysteine synthetase; Gro, glycerol; DHAP, dihydroacetone phosphate; LPA, lysophosphatidic acid; PA, phosphatidic acid; DAG, diacylglycerol; TAG, triacylglycerol; CDP-DAG, CDP-diacylglycerol; Cho, choline; Etn, ethanolamine; CL, cardiolipin; PG, phosphatidylglycerol; PGP, phosphatidylglycerol phosphate; Glc, glucose; Ins, inositol; PI, phosphatidylinositol; PS, phosphatidylserine; PE, phosphatidylethanolamine; PC, phosphatidylcholine; Cho2, phosphatidylethanolamine methyltransferase catalyzing first methylation from PE to PC; Opi3, phospholipid methyltransferase catalyzing two last methylation steps from PE to PC.

In addition to TAG metabolism, transcriptional regulation of phospholipid biosynthesis is also affected in mutants impaired in AdoHcy catabolism. Deficient phospholipid methylation in Sah1 depleted mutant cells unable to hydrolyze AdoHcy, or in cho2 and opi3 mutants, leads to induced expression of phospholipid biosynthetic (UASINO-responsive) genes, indicating accumulation of the phospholipid precursor, phosphatidic acid, in the endoplasmic reticulum (ER) [7]. ACC1 encoding acetyl-CoA carboxylase, the first and rate-limiting enzyme of fatty acid biosynthesis, is also subject to UASINO-mediated regulation. Therefore, the inhibition of phospholipid methylation in mutants deficient in AdoHcy catabolism as well as in cells supplemented with Hcy appear to result in an accumulation of fatty acids and their channeling into TAG, which play a crucial role in buffering excess fatty acids [135]. Moreover, in addition to a deregulated phospholipid and triacylglycerol metabolism, down-regulation of SAH1 expression in yeast also impairs sterol synthesis leading to 4-fold elevated squalene levels. This suggests an accumulation of early precursors of ergosterol biosynthesis under these conditions, presumably as a result of inhibition of sterol 24-C-methyltransferase (Tehlivets, Kohlwein, unpublished).

AdoHcy and homocysteine in lipid metabolism in mammals

AdoHcy also inhibits mammalian phosphatidylethanolamine N-methyltransferase (PEMT) [136], [137]. Similarly as in yeast, deficiency of phospholipid methylation in PEMT−/− knockout mice leads to a rapid decrease of the hepatic PC/PE ratio and accumulation of TAG in the liver, in the absence of choline supplementation [138]. However, TAG accumulation in the livers of these animals is also due to a decreased TAG secretion from hepatocytes [139].

Elevated levels of homocysteine are also linked to the deregulation of lipid metabolism in mammals. Cystathionine β-synthase−/− knockout mice exhibit severe hyperhomocysteinemia and accumulate AdoHcy in all tissues tested [53], [140]. These mutant animals show elevated TAG and nonesterified fatty acid levels in the liver and serum, and develop hepatic steatosis [12]. Another genetic disorder that results in moderately elevated homocysteine levels, methylenetetrahydrofolate reductase deficiency, leads to fatty liver development as well as to neuropathology and aortic lipid deposition in mouse model systems [10], [141]. Additionally, dietary-induced hyperhomocysteinemia in mice also causes fatty liver [142] suggesting that AdoHcy accumulation and impaired phospholipid methylation are common denominators in homocysteine-associated pathologies.

Activation of UPR and deregulation of unsaturated fatty acid metabolism in AdoHcy-related deficiencies

Supporting the role of elevated homocysteine levels in the deregulation of lipid metabolism in mammals, homocysteine supplementation was shown to lead to the activation of the sterol regulatory element-binding proteins (SREBPs), which function to activate genes encoding enzymes in cholesterol, fatty acid and triacylglycerol biosynthesis and uptake pathways, both in cultured mammalian cell lines as well as in the livers of CBS−/− mice that lack cystathionine β-synthase [142], [143]. The finding that overexpression of the ER chaperone GRP78/BiP inhibits homocysteine-induced gene expression [142], [144] suggests that unfolded protein response (UPR) induction plays a direct role in the activation of triacylglycerol and cholesterol biosynthesis in response to homocysteine accumulation in mammals. One of the possible mechanisms responsible for the induction of ER stress and UPR activation in response to homocysteine accumulation is the accumulation of saturated fatty acids in membrane phospholipids [145], [146], [147]. In accordance, overexpression of stearoyl-CoA desaturase was shown to attenuate palmitate-induced ER stress and protect from lipoapoptosis in mammalian cells [148], [149], [150].

Phospholipid methyltransferases preferentially catalyze synthesis of PC containing di-C16:1 species in yeast both in vitro and in vivo [151]. In line, the PEMT-derived PC pool is also enriched in unsaturated fatty acids [152], [153]. Supporting the role of PEMT in the metabolism of unsaturated fatty acids, PEMT−/− knockout mice were reported to accumulate more saturated PC molecular species in the liver compared with the control littermates [154], and to exhibit dramatically reduced concentrations of polyunsaturated fatty acids in the plasma and in hepatic PC, independently of the choline status [155]. Moreover, both AdoHcy and homocysteine accumulation are linked to decreased levels of unsaturated fatty acids. It was shown in particular, that elevated plasma AdoHcy levels are negatively correlated with both PC content and the level of polyunsaturated fatty acids in PC in red blood cells in Alzheimer patients [156]. Furthermore, elevated plasma homocysteine levels were shown to be associated with a decrease in polyunsaturated (docosahexaenoic) fatty acids in the plasma of healthy humans [157] and in the plasma and erythrocytes of cystic fibrosis patients [158]. It was proposed that deficiency of phospholipid methylation due to AdoHcy accumulation leads to an increase in saturated PC molecular species in ER membranes followed by ER stress, protein misfolding, induction of UPR and activation of lipid metabolism [159].

Concluding remarks and future perspectives

With 70% identity between yeast and human orthologs S-adenosyl-L-homocysteine hydrolase, which catalyzes the reversible hydrolysis of AdoHcy to homocysteine and adenosine, is among the most conserved S. cerevisiae proteins, pointing to its central role in the cellular metabolism. Moreover, accumulation of homocysteine, characteristic to the common pathological condition, hyperhomocysteinemia, “mimics” S-adenosyl-L-homocysteine hydrolase dysfunction, due to the reversal of the S-adenosyl-L-homocysteine hydrolase reaction resulting in the synthesis and accumulation of AdoHcy.

Homocysteine is linked to numerous diseases, however, the mechanisms responsible for the pathological consequences associated with these disorders are still elusive. Most importantly, AdoHcy rather than homocysteine is recognized since recently as a more sensitive marker of several homocysteine-associated pathologies.

As a product inhibitor of a number of AdoMet-dependent methyltransferases AdoHcy can trigger multiple pathological mechanisms. Due to the complexity of methylation metabolism comprehension of these mechanisms requires a systematic approach. Yeast as a validated unicellular model system particularly amenable to genomic approaches offers new possibilities to understand cellular responses to AdoHcy accumulation. Our recent work confirms that this system can be successfully used for identification and characterization of AdoHcy responsive cellular processes that are relevant for humans.