Metabolic engineering in yeast demonstrates that S-adenosylmethionine controls flux through the methylenetetrahydrofolate reductase reaction in vivo.

One-carbon flux into methionine and S-adenosylmethionine (AdoMet) is thought to be controlled at the methylenetetrahydrofolate reductase (MTHFR) step. Mammalian MTHFRs are inhibited by AdoMet in vitro, and it has been proposed that methyl group biogenesis is regulated in vivo by this feedback loop. In this work, we used metabolic engineering in the yeast Saccharomyces cerevisiae to test this hypothesis. Like mammalian MTHFRs, the yeast MTHFR encoded by the MET13 gene is NADPH-dependent and is inhibited by AdoMet in vitro. This contrasts with plant MTHFRs, which are NADH-dependent and AdoMet-insensitive. To manipulate flux through the MTHFR reaction in yeast, the chromosomal copy of MET13 was replaced by an Arabidopsis MTHFR cDNA (AtMTHFR-1) or by a chimeric sequence (Chimera-1) comprising the yeast N-terminal domain and the AtMTHFR-1 C-terminal domain. Chimera-1 used both NADH and NADPH and was insensitive to AdoMet, supporting the view that the C-terminal domain is responsible for AdoMet inhibition. Engineered yeast expressing Chimera-1 accumulated 140-fold more AdoMet and 7-fold more methionine than did the wild-type and grew normally. Yeast expressing AtMTHFR-1 accumulated 8-fold more AdoMet. This is the first in vivo evidence that the AdoMet sensitivity and pyridine nucleotide preference of MTHFR control methylneogenesis. (13)C labeling data indicated that glycine cleavage becomes a more prominent source of one-carbon units when Chimera-1 is expressed. Possibly related to this shift in one-carbon fluxes, total folate levels are doubled in yeast cells expressing Chimera-1.

Methylenetetrahydrofolate reductase (MTHFR) 1 catalyzes the reduction of 5,10-methylenetetrahydrofolate (CH 2 -THF) to 5-methyltetrahydrofolate (CH 3 -THF), which serves as a methyl donor for the synthesis of methionine. Methionine gives rise to S-adenosylmethionine (AdoMet), which is used for numerous methylation reactions. Because the MTHFR reaction commits the methyl group carried by THF to methyl group biogenesis, the regulation of MTHFR is crucial for one-carbon (C 1 ) metabolism in all organisms (1,2). The MTHFR reaction in mammalian liver is NADPH-dependent and is consequently physiologically irreversible due to the high cytosolic NADPH/NADP ratio and the large standard free energy change for the reduction of CH 2 -THF (3)(4)(5). This reaction therefore has the potential to deplete the cytosolic CH 2 -THF pool (1,2). Mammalian MTHFRs have been shown to be inhibited by AdoMet (6,7), and this sensitivity is considered to be a key regulatory feature that prevents CH 2 -THF depletion (1,2,6). However, there is no direct evidence for the existence of this regulation in vivo.
Plant MTHFRs are NADH-dependent and are thought to be cytosolic. Due to the high cytosolic NAD/NADH ratio (8), the MTHFR reaction in planta is likely to be reversible, obviating a need for regulation by AdoMet (9). Consistent with this, plant MTHFRs are AdoMet-insensitive (9). Both Saccharomyces cerevisiae and Schizosaccharomyces pombe have two divergent copies of the MTHFR gene. The MTHFR genes of S. cerevisiae (MET12 and MET13) have been cloned and shown to encode functional enzymes (10), but neither their pyridine nucleotide requirement nor their AdoMet sensitivity has been investigated. Strains in which MET13 is disrupted require methionine for growth. However, met12 disruptants have no methionine requirement, and overexpressing Met12p does not eliminate the methionine requirement of met13 disruptants. These data indicate that the Met13p isozyme provides most of the MTHFR activity in yeast cells (10).
All known eukaryotic MTHFRs comprise an N-terminal domain that contains the catalytic site and a C-terminal domain that is implicated in the regulation of enzymatic activity (11)(12)(13). The C-terminal domain of mammalian MTHFRs contains the AdoMet-binding site and is therefore considered responsible for the enzyme's sensitivity to this effector (11)(12)(13)(14). The function of the C-terminal domain of plant MTHFRs, which are insensitive to AdoMet, is unknown (9). The NADH-dependent, AdoMet-insensitive MTHFR of Escherichia coli (MetF) lacks the C-terminal domain (15)(16)(17).
In this study, we first showed that S. cerevisiae Met13p is NADPH-dependent and inhibited by AdoMet, like mammalian MTHFRs. We then constructed a chimeric yeast-plant MTHFR (Chimera-1) and characterized its pyridine nucleotide dependence and AdoMet sensitivity. Finally, we used metabolically engineered yeast strains expressing AdoMet-insensitive MTHFRs (AtMTHFR-1 or Chimera-1) to establish the importance of the AdoMet feedback loop in vivo.
Enzyme Isolation and Assays for MTHFR Activity-Yeast cells were grown, washed, and extracted as described previously (9). Desalted, concentrated extracts were also prepared as described previously (9). MTHFR activity was assayed in the reductive direction using the NAD(P)H-CH 2 -THF oxidoreductase radioassay described previously (9). The assay buffer was 100 mM potassium phosphate, pH 7.2. Unless otherwise noted, substrates were saturating, and product formation was proportional to enzyme concentration and time.
Determination of AdoMet and Methionine-Yeast cells were grown in YMD medium supplemented with glycine, formate, leucine, histidine, tryptophan, and uracil. Cells were harvested in mid-log growth and washed with water. To measure AdoMet, the cell pellet (300 mg wet weight) was resuspended in 500 l of water and lysed by vortexing with glass beads for 5 ϫ 30 s. Protein from clarified lysates was precipitated with ethanol (final concentration, 70%) at Ϫ20°C for 15 min and removed by centrifugation (13,000 ϫ g, 20 min, 4°C). The supernatant was evaporated to dryness in vacuo, dissolved in water, and analyzed by isocratic HPLC (20) using a Waters Spherisorb ODS2 column (4.6 ϫ 250 mm) with a Beckman Coulter System Gold 126 Solvent Module and 168 Detector and a Rheodyne 7725i Injector. The mobile phase was 50 mM NaH 2 PO 4 , 10 mM 1-heptanesulfonic acid, and 20% methanol, pH 4.4; column temperature was 22°C, flow rate was 0.8 ml min Ϫ1 , and run time was 30 min. To determine methionine, 200-mg samples of wet cells were suspended in 180 l of water and heated at 100°C for 5 min. After cooling on ice, cells were extracted essentially as described above, except that protein was precipitated with 5-sulfosalicylic acid (final concentration, 3.5%). The deproteinized samples were submitted to Biosynthesis Inc. (Lewisville, TX) for amino acid analysis using a Beckman 7300 analyzer.
Preparation of Extracts for NMR Analysis-Yeast cells were grown in 500-ml cultures supplemented with [ 13 C]formate (250 mg/liter) and unlabeled glycine, harvested at an A 600 of 3-5, and washed with water. Extracts were prepared by resuspending the cell pellet from the whole culture in 500 l of water and lysing with glass beads by continuous vortexing at 4°C for 20 -30 min. Lysates were clarified by centrifugation, and proteins were precipitated with ethanol as described above. The deproteinized supernatant was dried in vacuo and redissolved in 1 ml of D 2 O.
NMR Analysis-NMR spectra were obtained on a Varian Unity-Inova Spectrometer with a 1 H frequency of 500 MHz and a 13 C frequency of 125 MHz. 13 C data were collected with an acquisition time of 1.3 s with a 3-s delay and 90°pulse angle. Two-level composite pulse was used for proton decoupling with a power level of 48 dB during acquisition and 40 dB during delay. A total of 200 -2000 scans of 64,000 data points were acquired over a sweep width of 25 kHz. Data processing included exponential line broadening of 1 Hz. [methyl-13 C]AdoMet was quantified by comparing NMR peak heights with a standard sample of known concentration.
Folate Determination-Yeast strains were grown, harvested, and washed as described for AdoMet and methionine determination. Pelleted cells were suspended in twice their wet weight of extraction buffer (50 mM HEPES, 50 mM 2-(cyclohexylamino)ethanesulfonic acid, 0.2 M 2-mercaptoethanol, and 2% (w/v) sodium ascorbate, pH 7.9). Glass beads were added at 1.5 times the wet weight of the pellet. The samples were heated at 100°C for 10 min and lysed by vortexing for 4 min. After centrifugation (20,000 ϫ g, 30 min), supernatants were decanted, their volumes were measured, and they were stored at Ϫ70°C until analysis. Folate species were separated by HPLC and quantified using the Lactobacillus casei microbiological assay (21,22).

NADPH Dependence and AdoMet Sensitivity of Yeast MTHFR Met13p-Recombinant
Met13p was overexpressed in RRY3, a met12 met13 double disruptant that totally lacks MTHFR activity and is a methionine auxotroph (10). The pyridine nucleotide preference and AdoMet sensitivity of recombinant Met13p were tested by measuring enzyme activity radiometrically in the forward (reductive) direction in 100 mM potassium phosphate buffer, pH 7.2 (9). Met13p was found to be strictly NADPH-dependent and to be inhibited by AdoMet (Fig. 1A). Activity with NADH as reductant was below the detection limit of the assay (Ͻ2% of activity with NADPH).
Construction of a Chimeric Yeast-Plant MTHFR-A chimeric MTHFR (Chimera-1) was constructed by fusing the N-terminal (catalytic) domain of Met13p to the C-terminal domain of a plant MTHFR (AtMTHFR-1 from Arabidopsis thaliana; Ref. 9). The domains were spliced together by overlap extension PCR (18) within a short, conserved amino acid sequence (VRPIFW) that is immediately downstream of a region identified as the interdomain bridge in mammalian MTHFRs (12). Fig. 2A shows the domain structure of the chimeric enzyme and its parents schematically; Fig. 2B shows the sequence of the Chimera-1 polypeptide.
Complementation of a Yeast ⌬met13 Mutant by CHIMERA-1-The CHIMERA-1 DNA was cloned into the yeast expression vector pVT103-U and introduced into strain SCY1 (MET12 ⌬met13). A MET12 ϩ strain was used to avoid any possible growth effects of a met12 mutation. The transformation yielded methionine-independent colonies whose growth was comparable with that of the wild-type strain DAY4, or SCY1 transformed with MET13-pVT101-U (Fig. 3). No complementation was observed with the vector alone, and retransformation of SCY1 with rescued plasmid containing CHIMERA-1 restored methionine prototrophy, confirming that the complementation is due to the plasmid-encoded chimeric enzyme (data not shown). Arabidopsis MTHFR (AtMTHFR-1), expressed from pVT103-U, was previously shown to complement the methionine requirement of a yeast met13 mutation (9).
Pyridine Nucleotide Preference and AdoMet Insensitivity of Chimera-1-Recombinant Chimera-1 overexpressed in yeast strain RRY3 was tested for pyridine nucleotide preference and AdoMet sensitivity as described above for Met13p. Chimera-1 was found to use both NADH and NADPH (Fig. 1B), which was unexpected because its yeast and plant parents use either one or the other (Fig. 1, A and C). However, as anticipated, Chimera-1 was totally insensitive to AdoMet (Fig. 1B), like the plant enzyme that contributes its C-terminal region (Fig. 1C).
Kinetic Properties of Met13p, AtMTHFR-1, and Chimera-1-Velocity versus [S] plots for both Met13p and AtMTHFR-1 showed substrate inhibition by CH 2 -THF that was especially marked in the case of Met13p (Fig. 4). The K m and K i values of these two enzymes for CH 2 -THF were accordingly calculated as described by Cleland (23). Chimera-1 showed no inhibition by CH 2 -THF within the range tested (Fig. 4). Its K m values for CH 2 -THF were therefore estimated from Hanes plots; these values were similar whether NADH or NADPH was the reductant and fell midway between those of the parent enzymes (Table I). Because the activities of Met13p and AtMTHFR-1 were inhibited by CH 2 -THF, apparent K m values of these en-zymes for NADPH and NADH were compared with those of Chimera-1 using a nonsaturating CH 2 -THF concentration of 50 M ( Table I). The apparent K m values of Chimera-1 for both NADPH and NADH were closer to that of AtMTHFR than to that of Met13p.
Construction of Yeast Strains with a Chromosomal Copy of CHIMERA-1 or AtMTHFR-1-To study how AdoMet sensitivity and pyridine nucleotide specificity impact the function of MTHFR in vivo, we engineered strains in which the coding sequence of AtMTHFR-1 or Chimera-1 replaced that of the chromosomal MET13 gene. The introduced sequences were thus present as single chromosomal copies under the control of the native promoter, ensuring an expression level comparable to that of MET13. Replacement of the MET13 open reading frame with Chimera-1 and AtMTHFR-1 gave strains SCY4 and SCY6, respectively. The growth rates of these strains in YMD medium supplemented with serine, histidine, leucine, tryptophan, and uracil were similar to that of the wild-type (doubling times, 2.5 Ϯ 0.2 h). Thus, both the chimeric enzyme and the plant enzyme complemented the methionine requirement when expressed from the chromosomal MET13 promoter.
Levels of AdoMet and Methionine in DAY4, SCY4, and SCY6 Yeast Strains-AdoMet and free methionine were determined by HPLC. In the experiment shown in Fig. 5, strain SCY6 expressing the AdoMet-insensitive, NADH-dependent At-MTHFR-1 accumulated 8-fold more AdoMet than the DAY4 strain expressing the native Met13p enzyme. Strain SCY4 expressing the AdoMet-insensitive, NADH-or NADPHdependent enzyme Chimera-1 accumulated even more AdoMet (140-fold more AdoMet than DAY4) (Fig. 5A). Along with the greatly elevated AdoMet level, strain SCY4 accumulated 7-fold more methionine than DAY4 (Fig. 5B). AdoMet hyperaccumulation was seen in SCY4 cultures harvested at a range of cell densities; its magnitude varied from 75-fold to 254-fold above wild-type in four independent experiments (data not shown).
Origin of the C 1 Units Used for AdoMet Synthesis-The yeast strains used in this study are all ser1 strains that are blocked at phosphoserine aminotransferase and thus require serine for growth; glycine (or glycine plus formate) can substitute for serine. Use of ser1 Ϫ strains allows controlled introduction of C 1 units via the cytosolic C 1 -THF synthase by providing formate or via the mitochondrial glycine cleavage system by providing glycine (24). To identify the source(s) of C 1 units for production of AdoMet, we determined the methyl-13 C enrichment of AdoMet in cultures supplemented with [ 13 C]formate and unlabeled glycine. The methyl-13 C enrichment under these conditions measures the relative contributions to methyl group synthesis of C 1 units originating from cytosol and mitochondria (25,26). The methyl- 13  wild-type expressing Met13p and 0.54 for a strain expressing Chimera-1. These data show that with wild-type Met13p, Ͼ90% of AdoMet comes from [ 13 C]formate, whereas only 54% comes from [ 13 C]formate when Chimera-1 replaces Met13p. Thus, AdoMet hyperaccumulation is associated with a far greater contribution from mitochondrial glycine cleavage to the C 1 unit pool.
Intracellular Folate Levels in Yeast Strains DAY4, SCY4, and SCY6 -The pools of 10-formyltetrahydrofolate, 5-formyltetrahydrofolate, 5-CH 3 -THF, and THF were determined by HPLC combined with microbiological assay (21,22). The most abundant folate in all strains was 5-CH 3 -THF, constituting Ն68% of the total folate pool. The other folates (10formyltetrahydrofolate, 5-formyltetrahydrofolate, and THF) each represented Ͻ15% of the total in all strains (Table II). The most striking difference between the strains was that SCY4 expressing Chimera-1 had more than twice the total folate content of the wild-type strain DAY4 (156 versus 66 nmol g Ϫ1 wet weight), although the relative amounts of the individual folates were essentially unchanged. Another difference was that strain SCY6 expressing AtMTHFR-1 showed only a slight increase in total folate relative to the wild-type DAY4, but its 10-formyltetrahydrofolate and THF pools were doubled. Chimera-1 comprises the N-terminal domain of Met13p and the Cterminal domain of AtMTHFR-1. The gray bar marks the hydrophilic bridge region between the domains, and the white triangle shows the position of the AdoMet-binding site identified by photoaffinity labeling of mammalian MTHFR (11,13). B, amino acid sequence of Chimera-1. The sequence from Met13p is in black, and the sequence from At-MTHFR-1 is in gray.

FIG. 4. Dependence of recombinant MTHFR activity on CH 2 -THF concentration.
Desalted extracts of RRY3 (⌬met12 ⌬met13) yeast cells expressing Met13p, AtMTHFR-1, or Chimera-1 were assayed for NAD(P)H-CH 2 -THF oxidoreductase activity using an NADH or NADPH concentration of 100 M and CH 2 -THF concentrations of 10 -1500 M. Enzyme activities are expressed as a percentage of the maximum activity obtained for each. Maximum activities (in nmol min Ϫ1 mg Ϫ1 protein) were as follows: Met13p, 12.1; AtMTHFR-1, 27.9; and Chimera-1, 92.2. Only the values obtained with NADPH as reductant are shown for Chimera-1; values with NADH were very similar. All data shown are means of duplicate observations; S.E. bars were smaller than the data points.

DISCUSSION
Characterization of yeast Met13p shows that its properties are very much like those of mammalian MTHFRs: Met13p is AdoMet-sensitive, NADPH-dependent, and susceptible to substrate inhibition by CH 2 -THF. Photoaffinity labeling evidence indicates that the AdoMet-binding site of porcine and human MTHFRs lies in the C-terminal domain, within a stretch of about 30 residues starting 30 residues after the junction between the domains (12,14). The sequence of Met13p in this region is far closer to that of human MTHFR (41% identity and 66% similarity) than to that of the AdoMet-insensitive plant enzyme AtMTHFR-1 (20% identity and 44% similarity). Out-side this region, all three MTHFRs share about ϳ40% identity. The primary structure of Met13p is thus consistent with the AdoMet-binding site being in the proximal part of the C-terminal domain.
The effects of allosteric inhibition by AdoMet and pyridine nucleotide preference on the in vivo control of MTHFR activity were assessed in metabolically engineered yeast strains SCY6 andSCY4.StrainSCY6expressesanAdoMet-insensitive,NADHdependent plant enzyme, AtMTHFR-1. Strain SCY4 expresses the chimeric enzyme Chimera-1, comprising the Met13p Nterminal domain and the AtMTHFR-1 C-terminal domain. Chimera-1 is an active, AdoMet-insensitive enzyme that uses both NADPH and NADH. The AdoMet insensitivity of Chimera-1 is consistent with the C-terminal domain of MTHFR being responsible for sensitivity to AdoMet. Because the determinants for pyridine nucleotide binding are located in the N-terminal domain (27), the loss of selectivity for NADPH in Chimera-1 could reflect distortion in the tertiary structure of its Met13p N-terminal domain caused by fusion to a foreign C terminus. Another possibility is that interaction between the catalytic and C-terminal domains guides the specificity of the enzyme for its pyridine nucleotide substrate. In either case, it is noteworthy that porcine MTHFR can use NADH as a substrate if the assay is carried out in the presence of phosphate, for this shows that the enzyme's pyridine nucleotide specificity can be readily altered (28). The apparent lack of substrate inhibition of Chimera-1 by CH 2 -THF may likewise be explained by distorted tertiary structure or by modified domain interaction.
The overaccumulation of methionine and AdoMet in yeast strains engineered with the AdoMet-insensitive AtMTHFR-1 or Chimera-1 (8-and 140-fold more AdoMet, respectively) provides the first in vivo evidence that allosteric inhibition of MTHFR by AdoMet controls the commitment of C 1 units to methylneogenesis. The more dramatic effect of Chimera-1 may be explained by its ability to utilize NADPH as well as NADH (Fig. 6). The redox state of the NAD(P) pool is known to affect the reversibility of the MTHFR reaction. In mammals, the almost completely reduced NADP pool (NADPH/NADP ratio of ϳ10 7 ) ensures that the MTHFR reaction is practically irreversible (3)(4)(5). In plants, the NAD pool is highly oxidized (NADH/ NAD ratio of ϳ10 Ϫ3 ) and is thought to make the MTHFR    6. Proposed mechanism of AdoMet accumulation in engineered yeast strains. A, in the wild-type strain DAY4 expressing Met13p, the MTHFR reaction is irreversibly driven toward CH 3 -THF formation by a high cytosolic NADPH/NADP, as in mammals. AdoMet overaccumulation is prevented by allosteric feedback inhibition (dotted line). B, in strain SCY6 expressing AtMTHFR-1, there is moderate AdoMet accumulation due to the lack of feedback inhibition by AdoMet. The relatively low cytosolic NADH/NAD ratio may limit the extent of this accumulation. C, in strain SCY4 expressing Chimera-1, adding the capacity to use NADPH as well as NADH to the lack of AdoMet feedback inhibition results in more CH 3 -THF synthesis (and hence AdoMet accumulation) because the aggregate cytosolic NAD(P)H/NAD(P) ratio (by which Chimera-1 is driven) is higher than the NADH/NAD ratio. reaction freely reversible (9). Reliable measurements of cytosolic NAD(P)H/NAD(P) ratios in yeast are lacking. However, studies with yeast cytosolic CH 2 -THF dehydrogenase show that the NADP-dependent reaction is reversible in vivo, whereas the NAD-dependent reaction runs only in the oxidative direction (29). This indirect evidence suggests that the yeast cytosolic NADP pool is far more reduced than the NAD pool. An AdoMet-insensitive MTHFR that is driven by the aggregate cytosolic NAD(P)H/NAD(P) ratio might therefore be expected to direct more C 1 units to CH 3 -THF, and thence to AdoMet, than one driven solely by the NADH/NAD ratio.
The finding that expressing an NADH-dependent, AdoMetinsensitive plant MTHFR in yeast causes methionine and AdoMet overaccumulation raises the question of what prevents this from occurring in plants. Plants have the S-methylmethionine cycle, an auxiliary feature exclusive to plant C 1 metabolism (30,31). In this apparently futile cycle, AdoMet donates a methyl group to methionine, giving S-methylmethionine. S-Methylmethionine then donates a methyl group to homocysteine, yielding two molecules of methionine. The net effect of the cycle is thus to convert AdoMet back to methionine. In vivo radiolabeling data and metabolic modeling have recently indicated that the S-methylmethionine cycle serves to stop accumulation of unutilized AdoMet and that it may be the main mechanism whereby plants achieve short-term control of AdoMet levels (31). Because S. cerevisiae does not have this cycle, AdoMet might be expected to accumulate in yeast expressing the plant enzyme, as in fact occurs.
The 140-fold hyperaccumulation of AdoMet has no apparent ill-effect on yeast growth, at least under our culture conditions. A likely reason is that most of the excess AdoMet is sequestered in the vacuole and hence excluded from the metabolically active pool, as occurs when AdoMet is overproduced in yeast cultured with a high level of methionine in the medium (32). It should be noted that although AdoMet hyperaccumulation was a robust phenomenon in strain SCY4 grown as we describe in medium containing formate and glycine, it did not occur when these supplements were replaced by serine. The basis for this effect is now being investigated.
In wild-type yeast, the dominant methylated products are phosphatidylcholine head groups in membranes (ϳ12 mol methyl units g Ϫ1 wet weight) and protein-bound methionine (ϳ14 mol g Ϫ1 wet weight, assuming 100 mg protein g Ϫ1 wet weight and a methionine content of 2 mol %) (33) (Proteome Analysis at EBI, 2001, www.ebi.ac.uk/proteome). The total level of methyl end products is therefore at least 26 mol g Ϫ1 wet weight. In relation to this, the AdoMet pool in the wild-type strain DAY4 (31 nmol g Ϫ1 wet weight) is very small (Ͻ0.2%). However, in the AdoMet-hyperaccumulating strain, the AdoMet pool expands to ϳ4 mol g Ϫ1 wet weight, which adds ϳ15% to the methyl budget. To investigate the source of the extra methyl groups, we compared the 13 C-enrichment of the AdoMet methyl group in these two strains supplied with [ 13 C]formate and unlabeled glycine. In the wild-type, DAY4, formate was the principal source of the AdoMet methyl group, consistent with the previous data on the origin of choline methyl groups in this strain (29). Surprisingly, even though the production of methyl groups in the AdoMet-hyperaccumulating strain changed by only ϳ15%, the formate contribution to the AdoMet methyl group dropped dramatically from Ͼ90% to 54%, with the remainder presumably coming from glycine cleavage in the mitochondria. If we make the reasonable assumption that AdoMet methyl-13 C enrichment faithfully reflects that of methionine, this finding shows that AdoMet hyperaccumulation is associated with a profound change in the relative importance of cytosolic and mitochondrial C 1 fluxes. Consistent with the above inference, and equally surprising, strain SCY4 expressing Chimera-1 showed a doubling of total cellular folate content compared with the wild-type. This folate pool expansion could be a response to the shift in C 1 fluxes occasioned by unregulated AdoMet synthesis. Whatever the mechanism, the data imply that cross-talk between C 1 metabolism and folate synthesis or degradation is an important factor in control of intracellular folate levels.