Abstract

Various questions have been raised about labile methyl balance and total transmethylation fluxes, and further discussion has been encouraged. This report reviews and discusses some of the relevant evidence now available. The fact that, if needed, labile methyl balance is maintained by methylneogenesis appears to be established, but several aspects of transmethylation remain uncertain: definitive measurements of the rate of total transmethylation in humans of both sexes on various diets and at various ages; the extent to which synthesis of phosphatidylcholine has been underestimated; and the relative contributions of the 2 pathways for the formation of sarcosine (ie, N-methylglycine). The available evidence indicates that the quantitatively most important pathways for S-adenosylmethionine–dependent transmethylation in mammals are the syntheses of creatine by guanidinoacetate methyltransferase, of phosphatidylcholine by phosphatidylethanolamine methyltransferase, and of sarcosine by glycine N-methyltransferase. Data presented in this report show that S-adenosylmethionine and methionine accumulate abnormally in the plasma of humans with glycine N-methyltransferase deficiency but not of those with guanidinoacetate N-methyltransferase deficiency or in the plasma or livers of mice devoid of phosphatidylethanolamine N-methyltransferase activity. The absence of such accumulations in the latter 2 conditions may be due to removal of S-adenosylmethionine by synthesis of sarcosine. Steps that may help clarify the remaining issues include the determination of the relative rates of synthesis of sarcosine, creatine, and phosphatidylcholine by rapid measurement of the rates of radiolabel incorporation into these compounds from l-[methyl-3H]methionine administered intraportally to an experimental animal; clarification of the intracellular hepatic isotope enrichment value during stable-isotope infusion studies to enhance the certainty of methyl flux estimates during such studies; and definitive measurement of the dietary betaine intake from various diets.

INTRODUCTION

In a recent issue of this Journal, Stead et al (1) asked, “Is it time to reevaluate methyl balance in humans?” Their article, also a commentary, raised questions about existing views on this subject and expressed the hope that their comments would generate a discussion on labile methyl balance in humans. Here, we attempt to respond to that hope by further discussion of some of the questions raised by Stead et al and by presentation of additional data bearing on labile methyl balances. The available evidence leads us to conclude that, if needed, labile methyl balance is maintained by methylneogenesis, but that several aspects of transmethylation rates remain uncertain. Among the most serious of these aspects are the possibility that the rate of synthesis of phosphatidylcholine has been underestimated and the uncertainty about the rates and relative contributions of the 2 pathways for formation of sarcosine (N-methylglycine). We present relevant new data on the concentrations of many of the metabolites of the transsulfuration and transmethylation pathways in plasma and livers of mice devoid of phosphatidylethanolamine N-methytransferase (PEMT) activity or in plasmas of humans with severe deficiencies of either glycine N-methyltransferase (GNMT) or guanidinoacetate N-methyltransferase (GAMT). Steps that may help clarify the remaining issues are suggested.

THE OCCURRENCE OF METHYL BALANCE AND ESTIMATES OF MINIMAL TRANSMETHYLATION RATES

Initial attempts to estimate methyl balances in humans and the total fluxes of labile methyl groups originating either in methionine [proceeding through S-adenosylmethionine (AdoMet)] or in choline (proceeding through betaine, methionine, and AdoMet) to various methylated products were reported by Mudd and Poole (2). The relevant pathways are summarized in Figure 1. Young adult volunteers adapted to normal diets providing ≈9.9 (males) or ≈9.2 mmol (females) methionine/d and ≈3.6 (males) and ≈2.5 mmol (females) choline/d were found to excrete in the urine 17.5 and 12.9 mmol, respectively, methylated products/d, chiefly in the form of creatinine, but including 1.9 (males) or 2.6 (females) mmol creatine/d plus 31 additional methylated compounds. The excretions of the latter 31 compounds (all relatively minor) were estimated from published values. When the subjects were adapted to experimental diets containing 4.7 (males) or 4.5 (females) mmol methionine/d, ≈0.2 mmol choline/d, and negligible creatine plus creatinine, they remained in slightly positive nitrogen balance or within the zone of nitrogen equilibrium and continued to excrete in the urine 17.6 (males) and 12.7 (females) mmol methylated products/d, again chiefly in the form of creatinine. Thus, when following the latter diets, all subjects put out far more methylated products than they took in as labile methyls in their diets.

FIGURE 1.

Major reactions involved in transmethylation flux and methylneogenesis. The total transmethylation flux is equivalent to the total flux occurring through reactions that convert S-adenosylmethionine to S-adenosylhomocysteine. The 3 S-adenosylmethionine–dependent reactions thought to contribute quantitatively most to this flux are methylation of guanidinoacetate by guanidinoacetate methyltransferase (GAMT) to form creatine; methylation of phosphatidylethanolamine by phosphatidylethanolamine methyltransferase (PEMT) to form phosphatidylcholine; and methylation of glycine by glycine N-methyltransferase (GNMT) to form sarcosine (N-methylglycine). A large number of additional S-adenosylmethionine–dependent methyltransferases also occur in mammals [see Katz et al (3)], but their collective quantitative contribution to transmethylation flux may be small compared with those mentioned above. The final steps in methylneogenesis are the reduction of a methylene group of 5,10-methylenetetrahydrofolate (methylene-THF) by methylenetetrahydrofolate reductase (MTHFR) to form 5-methyltetrahydrofolate (methyl-THF), followed by transfer by methionine synthase of the newly formed methyl moiety to homocysteine, forming methionine and tetrahydrofolate (THF). Sarcosine is formed not only by GNMT, but also by oxidation of choline to betaine, formation of dimethylglycine by betaine homocysteine methyltransferase (BHMT), and oxidation of dimethylglycine to sarcosine. Sarcosine is oxidized by sarcosine dehydrogenase (SDH). During the reaction, glycine is produced, and a 1-carbon unit is transferred to THF, forming methylene-THF. MAT, methionine adenosyltransferase; CBS, cystathionine β-synthase; CGL, cystathionine γ-lyase; SAHH, S-adenosylhomocysteine hydrolase.

These observations provide very strong evidence that, when labile methyl intake is less than methyl output, the difference is balanced by the formation of new methyl groups. This conclusion is now amply supported by the stable-isotope infusion studies reviewed in this article and by observation of patients with genetic defects in the pathway that carries out such methylneogenesis—namely, formation of methyltetrahydrofolate (methyl-THF) from methylene-THF followed by transfer of the methyl group of methyl-THF to homocysteine to form methionine. For example, such patients tend to have abnormally low or borderline low plasma methionine and to have elevated plasma total homocysteine (tHcy) (4). Thus, the fact is well established that, when dietary intake of labile methyls is less than that required for ongoing methylation reactions, methyl balance is normally maintained by de novo formation of methyl groups.

What remains less clear is the total amount of transmethylation flux and the constituents of that flux. Mudd and Poole (2) pointed out that their estimates of total transmethylation were preliminary and probably minimal, because methylated products in addition to those specified may be found, and, as is equally important, because no account had been taken of the utilization of methyl groups to form products that were subsequently demethylated [eg, sarcosine (N-methylglycine)], the use of choline to form sarcosine, and the use of methyl esters such as the protein methyl ester that may be the chief consumer of AdoMet in human red blood cells (5). Furthermore, excretion of methylated compounds other than in the urine was not taken into account. Stead et al pointed out that, according to current estimates, a typical human secretes into the bile an amount of phosphatidylcholine the synthesis of which would consume some 5 mmol AdoMet/d. Although a substantial, but uncertain, amount of this phosphatidylcholine is reabsorbed (1), biliary secretion may contribute significantly to the total output of compounds formed from AdoMet.

In an effort to address one of these problems, Mudd et al (6) measured sarcosine excretion by a 16-y-old sarcosinemic girl with a genetic deficiency in her ability to metabolize sarcosine. Weighing 50 kg and following a normal diet containing 7.1 mmol methionine/d and 2.9 mmol choline/d (as lecithin), she had a urinary sarcosine excretion of 4.3 mmol/d, excretion of creatinine plus creatine of 10.2 mmol/d, and that of other methylated compounds [those included in the earlier study (2)] of 1.4 mmol/d, for a total excretion of 15.9 mmol methyl groups/d. When labile methyl group intake from methionine plus choline was increased beyond that required for formation of creatinine and the specified methylated compounds, sarcosine excretion rose to balance the excess intake. A limitation was that it was not possible to be sure what portion of the sarcosine excreted was formed by each of 2 alternative pathways: AdoMet-dependent methylation of glycine or oxidation of choline to betaine, which was followed by conversion of betaine to dimethylglycine and then to sarcosine (Figure 1). To the extent that the latter pathway contributed, in a steady state, the utilization of a choline moiety may be balanced at least to some extent by AdoMet-dependent conversion of phosphatidylethanolamine to phosphatidylcholine, a process using 3 methyl groups, only 1 of which would be regained from the betaine methyl. Such phosphatidylcholine synthesis would require further upward revision of the labile methyl flux, as would also, of course, the other factors discussed above. Two additional adjustments of the results for this patient are pointed out by Stead et al (1). First, the recovery of sarcosine in her urine was only 60–80% of an administered dose, which suggested that she retained some capacity to metabolize sarcosine. The value for sarcosine formation might then be adjusted to 4.3/0.7 = 6.1 mmol/d. Second, customary diets in the United States, such as the basal diet of the subject, provide ≈5.0 mmol creatine/d to young adult females (1). On the basis of data reported for the absorption of creatine (79), we estimate that ≈80% of dietary creatine will be absorbed. The uptake from a dietary intake of 5.0 mmol/d will then be 4.0 mmol/d. Subtraction of this uptake from the observed creatinine-plus-creatine excretion yields a corrected value for creatine synthesis of 10.2 − 4.0 = 6.2 mmol/d, and the minimal total adjusted labile methyl flux becomes 6.2 + 1.4 + 6.1 = 13.7 mmol/d. According to these values, the contribution of creatine synthesis to the total labile methyl flux would be (6.2/13.7) × 100% = 45%.

TOTAL LABILE METHYL FLUXES ESTIMATED BY STABLE-ISOTOPE INFUSION STUDIES

An alternative experimental approach that overcomes some of the limitations inherent in the above balance studies was devised in the Laboratory of Human Nutrition, School of Science and Clinical Research Center at the Massachusetts Institute of Technology. Constant intravenous infusion of [methyl-2H3] and [1-13C]methionine permitted calculation of total methyl fluxes, ie, transmethylation rates. In an initial study, Storch et al (10) found that, for healthy males following diets that were adequate in methionine and cystine and receiving supplementation with 500 mg choline/d, transmethylation rates were 14 ± 1.3 μmol · kg−1 · h−1 in the fed state and 5.85 ± 0.6 μmol · kg−1 · h−1 in the postabsorptive state. These values converted to an estimated daily flux of 238 ± 22 μmol/kg or, for a 70-kg person, 16.7 mmol/d (10). Using the same methodology, Fukagawa et al (11) found somewhat similar postabsorptive rates: 7.61 ± 1.02 μmol · kg−1 · h−1 for men and 8.43 ± 1.15 μmol · kg−1 · h−1 for women. Total daily fluxes were not estimated because studies were not done when subjects were in the fed state. MacCoss et al (12) did further similar studies of postabsorptive subjects. A still unresolved difficulty with this approach is the need to estimate the isotopic enrichment of methionine in liver. For this purpose, Storch et al (10) used the enrichment of plasma methionine, but MacCoss et al (12) measured the enrichments of plasma methionine as well as those of homocysteine and cystathionine and found the enrichments of the latter compounds to be less than the enrichment of methionine. Using the homocysteine and cystathionine enrichments, MacCoss et al calculated transmethylation rates ≈40% higher than those reported by Storch et al—ie, 16.7 × 1.4 = 23.4 mmol/d. However, Davis et al (13) subsequently found that the enrichment value for the M + 1 ion used by MacCoss et al is difficult to determine accurately because of its high background abundance. Using instead the M + 4 ion (which is far less subject to such inaccuracy), Davis et al found the isotopic enrichments of plasma homocysteine and methionine to be similar. On the basis of the methionine enrichment, they estimated that the correct value of MacCoss et al would have agreed closely (93%) with the earlier value of Storch et al. Most recently, Mercier et al (14) found fed and postabsorptive rates of 13.5 ± 0.8 and 8.2 ± 0.6 μmol · kg−1 · h−1, respectively, in young subjects and of 12.5 ± 0.4 and 6.0 ± 0.3 μmol · kg−1 · h−1, respectively, in subjects aged 66–76 y (again calculated by using methionine isotopic enrichment values). Converted to daily rates in the manner used by Storch et al (10), these rates indicate daily fluxes of 260 and 222 μmol/kg, or 18.2 and 15.5 mmol for a 70-kg person. Thus, after adjustment for uncertainty as to the correct intracellular hepatic isotopic enrichment value, a reasonable range for the transmethylation rate in young persons based on the stable-isotope infusion studies is 16.7–23.4 mmol/d, and that in elderly subjects is somewhat less: 15.5–21.7 mmol/d.

This rate of transmethylation in young persons may be compared with that calculated by Mudd et al (6) in the sarcosinemic girl. If the transmethylation rate is directly proportional to body weight, that subject's final weight-adjusted rate would become 12.7 × (70/50) = 17.8 mmol/d. Because (as discussed above) this is a minimal rate, it appears that the major portion of methylated compounds have been accounted for, although factors causing the utilization of as much as 5.6 mmol AdoMet/d (ie, 23.4 − 17.8 mmol/d) may remain to be identified. Data are available to indicate that, although it may account for at least 50% of the AdoMet consumption of erythrocytes, the rate of AdoMet utilization in one such reaction, the isoaspartyl methyltransferase of red blood cells, is lower relative to overall transmethylation (5). By using the values in Table 6 in the report by Barber et al (5), one may calculate an AdoMet consumption rate between 0.06 and somewhat more than 0.12 mmol/d by fresh human red blood cells.

IMPORTANT REMAINING UNCERTAINTIES

Among the more important remaining uncertainties in the above analysis are 1) the correct values for total transmethylation fluxes in humans of both sexes following various diets and at various ages; 2) the distribution between the metabolic pathways leading to sarcosine—either methylation of glycine or catabolism of choline, the latter of which will require compensatory synthesis; and 3) the actual rate of phosphatidylcholine synthesis, especially in humans. Stead et al (1) drew attention to the observation that plasma tHcy is decreased by 50% in mice devoid of the activity of PEMT, the enzyme that catalyzes AdoMet-dependent formation of phosphatidylcholine (15), and they interpreted this observation to suggest a major role for phosphatidylcholine synthesis in determining plasma tHcy in mice (1). This suggestion is supported by the observations that hepatocytes isolated from PEMT/ mice have less Hcy secretion than do those from their wild-type counterparts (16) and that, in CTP:phosphocholine cytidylyltransferase-α (CTα) knockout mice, PEMT flux is increased and plasma tHcy is raised by 20–40% (17). Furthermore, the estimated rate of hepatic synthesis of PEMT-derived phosphatidylcholine in rats (18), converted to the rate expected for a 250-g rat (17), would be 108–150 μmol/d, which exceeds the rate of creatine formation [based on the measured rate of creatinine excretion (19)] of 58 μmol/d for a 250-g rat. The next section provides further data relevant to the roles in AdoMet consumption of PEMT, GNMT (the enzyme that catalyzes methylation of glycine to form sarcosine), and GAMT (the enzyme that catalyzes methylation of guanidinoacetate, the last step in creatine synthesis).

METABOLITES IN SUBJECTS DEFICIENT IN A TRANSMETHYLATION ENZYME

Values for many of the metabolites most directly involved in the transmethylation pathway under circumstances in which the enzyme activities in question are severely decreased are shown in Table 1. AdoMet and S-adenosylhomocysteine (AdoHcy) were assayed as described by Capdevila and Wagner (22). Other metabolites were quantitated by capillary gas chromatography–mass spectrometry (2325).

TABLE 1

Metabolite concentrations in phosphatidylethanolamine N-methyltransferase knockout (PEMT−/−) and wild-type mice and in glycine N-methyltransferase (GNMT)- or guanidinoacetate N-methyltransferase (GAMT)-deficient humans1

Tissue and conditionMetAdoMetAdoHcytHcyCystatCysMGDMG
Liver samples         
    Mouse liver (nmol/g)         
        Wild-type (n = 5) 420 ± 35 34 ± 3 53 ± 12 14.6 ± 2.2 40 ± 12 458 ± 119 62 ± 21 88 ± 13 
        PEMT−/− (n = 5) 441 ± 63 32 ± 13 35 ± 19 19.3 ± 6.8 57 ± 21 512 ± 155 63 ± 27 58 ± 23 
Plasma samples2         
    Mouse plasma         
        Wild-type (n = 9) 119 ± 35 146 ± 393 33 ± 8 7.6 ± 2.4 613 ± 153 257 ± 33 2.7 ± 1.1 7.6 ± 0.6 
        PEMT−/− (n = 9) 106 ± 26 156 ± 643 42 ± 19 4.5 ± 1.44 877 ± 322 197 ± 24 4.0 ± 0.85 5.7 ± 1.4 
    Human plasma         
        GNMT-deficient (n = 3) 668 ± 31 2213 ± 36 50 ± 19 11.6 ± 1.4 609 ± 184 222 ± 11 1.8 ± 0.3 9.1 ± 2.5 
        GAMT-deficient (n = 7) 19 ± 9 65 ± 41 13 ± 8 4.5 ± 1.7 162 ± 43 272 ± 23 3.0 ± 16 5.2 ± 3.2 
        Reference range 13–457 93 ± 16 15–45 5.5 ± 2.1–6.4 ± 0.28 44–3427 203–3697 1.36 ± 0.557,9 1.4–5.37 
Tissue and conditionMetAdoMetAdoHcytHcyCystatCysMGDMG
Liver samples         
    Mouse liver (nmol/g)         
        Wild-type (n = 5) 420 ± 35 34 ± 3 53 ± 12 14.6 ± 2.2 40 ± 12 458 ± 119 62 ± 21 88 ± 13 
        PEMT−/− (n = 5) 441 ± 63 32 ± 13 35 ± 19 19.3 ± 6.8 57 ± 21 512 ± 155 63 ± 27 58 ± 23 
Plasma samples2         
    Mouse plasma         
        Wild-type (n = 9) 119 ± 35 146 ± 393 33 ± 8 7.6 ± 2.4 613 ± 153 257 ± 33 2.7 ± 1.1 7.6 ± 0.6 
        PEMT−/− (n = 9) 106 ± 26 156 ± 643 42 ± 19 4.5 ± 1.44 877 ± 322 197 ± 24 4.0 ± 0.85 5.7 ± 1.4 
    Human plasma         
        GNMT-deficient (n = 3) 668 ± 31 2213 ± 36 50 ± 19 11.6 ± 1.4 609 ± 184 222 ± 11 1.8 ± 0.3 9.1 ± 2.5 
        GAMT-deficient (n = 7) 19 ± 9 65 ± 41 13 ± 8 4.5 ± 1.7 162 ± 43 272 ± 23 3.0 ± 16 5.2 ± 3.2 
        Reference range 13–457 93 ± 16 15–45 5.5 ± 2.1–6.4 ± 0.28 44–3427 203–3697 1.36 ± 0.557,9 1.4–5.37 
1

All values are ± SD. Met, methionine; AdoMet, S-adenosylmethionine; AdoHcy, S-adenosylhomocysteine; tHcy, total homocysteine; Cysta, cystathionine; tCys, total cysteine; MG, methylglycine (sarcosine); DMG, dimethylglycine. PEMT−/− and wild-type mice were maintained on stock laboratory diet as previously described (15). The patients with GAMT deficiency had a mean age of 6.5 y (range: 6 mo–9 y) when the samples were obtained. All were receiving supplemental dietary creatine as part of the treatment for their enzyme deficiency.

2

Met, tHcy, tCys, MG, and DMG are reported in μmol/L; AdoMet, AdoHcy, and Cysta are reported in nmol/L.

3

n = 7.

4

Significantly lower than in wild-type mice (P = 0.004).

5

Significantly higher than in wild-type mice (P = 0.01).

6

Significantly higher than reference range (P < 0.05).

7

Reference range determined in adults aged 18–65 y.

8

The reference value of 5.5 ± 2.1 (n = 150) is that for Italian children aged 7.24 ± 4.9 y (20); that of 6.4 ± 0.25 (n = 28) is for Spanish children aged 4–9 y (21).

9

n = 60.

TABLE 1

Metabolite concentrations in phosphatidylethanolamine N-methyltransferase knockout (PEMT−/−) and wild-type mice and in glycine N-methyltransferase (GNMT)- or guanidinoacetate N-methyltransferase (GAMT)-deficient humans1

Tissue and conditionMetAdoMetAdoHcytHcyCystatCysMGDMG
Liver samples         
    Mouse liver (nmol/g)         
        Wild-type (n = 5) 420 ± 35 34 ± 3 53 ± 12 14.6 ± 2.2 40 ± 12 458 ± 119 62 ± 21 88 ± 13 
        PEMT−/− (n = 5) 441 ± 63 32 ± 13 35 ± 19 19.3 ± 6.8 57 ± 21 512 ± 155 63 ± 27 58 ± 23 
Plasma samples2         
    Mouse plasma         
        Wild-type (n = 9) 119 ± 35 146 ± 393 33 ± 8 7.6 ± 2.4 613 ± 153 257 ± 33 2.7 ± 1.1 7.6 ± 0.6 
        PEMT−/− (n = 9) 106 ± 26 156 ± 643 42 ± 19 4.5 ± 1.44 877 ± 322 197 ± 24 4.0 ± 0.85 5.7 ± 1.4 
    Human plasma         
        GNMT-deficient (n = 3) 668 ± 31 2213 ± 36 50 ± 19 11.6 ± 1.4 609 ± 184 222 ± 11 1.8 ± 0.3 9.1 ± 2.5 
        GAMT-deficient (n = 7) 19 ± 9 65 ± 41 13 ± 8 4.5 ± 1.7 162 ± 43 272 ± 23 3.0 ± 16 5.2 ± 3.2 
        Reference range 13–457 93 ± 16 15–45 5.5 ± 2.1–6.4 ± 0.28 44–3427 203–3697 1.36 ± 0.557,9 1.4–5.37 
Tissue and conditionMetAdoMetAdoHcytHcyCystatCysMGDMG
Liver samples         
    Mouse liver (nmol/g)         
        Wild-type (n = 5) 420 ± 35 34 ± 3 53 ± 12 14.6 ± 2.2 40 ± 12 458 ± 119 62 ± 21 88 ± 13 
        PEMT−/− (n = 5) 441 ± 63 32 ± 13 35 ± 19 19.3 ± 6.8 57 ± 21 512 ± 155 63 ± 27 58 ± 23 
Plasma samples2         
    Mouse plasma         
        Wild-type (n = 9) 119 ± 35 146 ± 393 33 ± 8 7.6 ± 2.4 613 ± 153 257 ± 33 2.7 ± 1.1 7.6 ± 0.6 
        PEMT−/− (n = 9) 106 ± 26 156 ± 643 42 ± 19 4.5 ± 1.44 877 ± 322 197 ± 24 4.0 ± 0.85 5.7 ± 1.4 
    Human plasma         
        GNMT-deficient (n = 3) 668 ± 31 2213 ± 36 50 ± 19 11.6 ± 1.4 609 ± 184 222 ± 11 1.8 ± 0.3 9.1 ± 2.5 
        GAMT-deficient (n = 7) 19 ± 9 65 ± 41 13 ± 8 4.5 ± 1.7 162 ± 43 272 ± 23 3.0 ± 16 5.2 ± 3.2 
        Reference range 13–457 93 ± 16 15–45 5.5 ± 2.1–6.4 ± 0.28 44–3427 203–3697 1.36 ± 0.557,9 1.4–5.37 
1

All values are ± SD. Met, methionine; AdoMet, S-adenosylmethionine; AdoHcy, S-adenosylhomocysteine; tHcy, total homocysteine; Cysta, cystathionine; tCys, total cysteine; MG, methylglycine (sarcosine); DMG, dimethylglycine. PEMT−/− and wild-type mice were maintained on stock laboratory diet as previously described (15). The patients with GAMT deficiency had a mean age of 6.5 y (range: 6 mo–9 y) when the samples were obtained. All were receiving supplemental dietary creatine as part of the treatment for their enzyme deficiency.

2

Met, tHcy, tCys, MG, and DMG are reported in μmol/L; AdoMet, AdoHcy, and Cysta are reported in nmol/L.

3

n = 7.

4

Significantly lower than in wild-type mice (P = 0.004).

5

Significantly higher than in wild-type mice (P = 0.01).

6

Significantly higher than reference range (P < 0.05).

7

Reference range determined in adults aged 18–65 y.

8

The reference value of 5.5 ± 2.1 (n = 150) is that for Italian children aged 7.24 ± 4.9 y (20); that of 6.4 ± 0.25 (n = 28) is for Spanish children aged 4–9 y (21).

9

n = 60.

PEMT−/− mice

The data in Table 1 confirm the earlier observation of Noga et al (15) that, in PEMT/ mice, plasma homocysteine is significantly lower, with a mean tHcy 59% of that of the wild-type mice. The plasma sarcosine (methylglycine), which was not reported by Noga et al (15), was significantly increased from 2.7 to 4.0 μmol/L. GNMT activity did not differ significantly between the livers of the PEMT/ mice and those of the wild-type mice (data not shown). Although sarcosine is a poor indicator because of its rapid metabolic turnover, the rise in question is consistent with an increased overflow of AdoMet methyls to methylate glycine when methylation of phosphatidylethanolamine cannot occur. In the PEMT/ mice plasmas, no significant changes were seen in methionine, AdoMet, or AdoHcy. Marginal changes were seen in plasma cystathionine and dimethylglycine. In the livers of the PEMT/ mice, however, AdoMet, AdoHcy, and tHcy were not changed (Table 1). These observations are consistent with the studies of liver-specific CTα knockout mice (17) in which PEMT flux increased and plasma tHcy was raised by 20–40%. However, hepatic AdoMet or AdoHcy concentrations did not change (17); tHcy was not measured. Taking these observations together, it is evident that altered flux through hepatic PEMT may influence plasma tHcy independently of liver AdoMet, AdoHcy, or tHcy concentrations. This finding raises the issue of total transmethylation flux in these animals and the question of the extent to which the decreased PEMT flux in the PEMT/ mice and the increased PEMT flux in the CTα knockout mice are compensated by, respectively, increased and decreased fluxes through GNMT. In liver of rats, the ratio between free and protein-bound Hcy is close to 1.5:1 if care is taken to freeze the liver in vivo, and the half-life for Hcy (free and bound regarded as a single pool) is 3 s (26). However, we draw attention to the lack of information as to the mechanism(s) whereby homocysteine is transported from liver to plasma.

Humans deficient in glycine N-methyltransferase activity

The plasma changes in GNMT-deficient humans (27, 28) differ quite strikingly from the pattern in PEMT−/− mice. These differences reflect an important distinction between GNMT and the other methyltransferases: fluxes through methyltransferases such as PEMT are limited by the physiologic need for their products, but GNMT is not constrained in this way. In such humans, marked rises are seen in plasma AdoMet and methionine (Table 1). The same pattern is observed in liver of GNMT−/−mice (29). The rise in AdoMet has previously been interpreted as due to a lack of consumption of AdoMet to methylate glycine and the rise of methionine attributed to one or more feedback mechanisms whereby accumulated AdoMet down-regulates the flux from methionine to AdoMet (27). The GNMT-deficient patients were following normal diets when the samples reported in Table 1 were obtained. Thus, the changes observed were not due to methionine intakes that would be considered excessive by current standards. Even under these circumstances, the methylation of glycine may contribute substantially to the transmethylation flux.

Plasma tHcy is known to rise somewhat during childhood. Therefore, the mean plasma tHcy value for the GNMT-deficient children was compared with the most relevant reference ranges we could find in the published literature (20, 21). The mean was found to be significantly higher than the means in these reference series (P = 0.001) and was also higher than the mean of 6.9 ± 2.2 μmol/L (n = 96) for boys aged 12.9 ± 0.3 y (P < 0.001) reported by Thomas et al (30). The reason for this elevation is not clear.

Patients with severe, genetically determined deficiencies of guanidinoacetate N-methyltransferase

In the patients in question, the diagnosis of GAMT deficiency was established by the presence of elevated guanidinoacetate concentrations. In those for whom the GAMT mutations have been identified (4 of 7), truncating changes were found, so their GAMT deficiencies were probably very severe (SH Mudd et al, unpublished observations, 2006). At the time the samples reported in Table 1 were obtained, all the patients were receiving supplemental creatine, a treatment that was beneficial with respect to their genetic abnormality (31). Because creatine down-regulates l-arginine:glycine amidinotransferase, the enzyme that forms guanidinoacetate (32, 33), this treatment should accentuate, not diminish, any effect or effects of the existing GAMT deficiency on AdoMet utilization for creatine synthesis. Indeed, creatine administration has been shown to decrease plasma tHcy by ≈25% in rats (33) and by 10% in humans, according to one report (34); in another report, this reduction in humans was not significant (35). The metabolic data in Table 1 show that the plasmas of these severely GAMT-deficient humans follow a pattern not unlike that in the PEMT−/− mice: a rise in sarcosine but no elevations of AdoMet or methionine. The GAMT-deficient children for whom values are reported in Table 1 were young, European, and mostly boys (6 of 7). The mean plasma tHcy value in these patients was lower, but not significantly so, than the mean in either of the relevant reference series listed in Table 1. It was significantly (P < 0.05) lower than the mean of 6.9 ± 2.2 μmol/L found by Thomas et al (30) in boys aged 12.9 ± 0.3 y.

One interpretation of the overall results in Table 1 is that, when GNMT activity is severely diminished or absent, increases in plasma AdoMet and methionine indicate disruption of a pathway that is important to AdoMet disposal; in contrast, in the presence of intact GNMT activity, such a disruption is indicated by a rise in sarcosine with a decrease in plasma tHcy (as in the PEMT−/− mice and, perhaps, the GAMT-deficient humans). These results do not clearly establish the relative contributions to the overall transmethylation flux of the reactions forming phosphatidylcholine, creatine, or sarcosine.

FOR THE FUTURE

Clearly, more evidence is necessary to answer the remaining questions about total transmethylation in humans and the roles therein of various reactions using AdoMet. A definitive determination of the use of AdoMet for phosphatidylcholine synthesis in humans would be an important step forward. Resolution of the uncertainty about hepatic isotopic enrichment in stable-isotope infusion experiments would be another important step, and more precise data on the absorption of dietary creatine would be useful. An approach that may empirically settle the question of the relative contributions of phosphatidylcholine, creatine, and sarcosine syntheses (at least in an experimental animal) would be the use of the procedure (or a modification thereof) described by Datko et al (36). They administered l-[methyl-3H]methionine intraportally to a rat for 25 s. At 30 s, the liver was rapidly removed and freeze-clamped. It was then possible to quantitate in the liver the amounts of radiolabel in phosphatidylmonomethylethanolamine, phosphatidyldimethylethanolamine, and phosphatidylcholine, whereas at longer times of labeling, the radioactivity in phosphatidylcholine far exceeded that in the monophosphatidylethanolamine and dimethylphosphatidylethanolamine derivatives. If labeling of creatine and sarcosine were also quantitated in such an experiment, it should be possible to judge the relative importance of these pathways. The results of this experiment will be most readily interpretable if the enzymes in question share the same pool of substrate AdoMet. This is likely to be so because both GNMT and GAMT use cytosolic AdoMet. Furthermore, although PEMT is located chiefly in liver on the endoplasmic reticulum (37), but also in the biliary canalicular membrane (38), Shields et al (39) proposed a model in which the AdoMet-binding portions of PEMT extend into the cytosol and would thus share the same pool of AdoMet as the other 2 enzymes.

Additional worries would be that ≥1 of the methylated products is exported from the liver so rapidly that, even in the short time of 30 s, a significant amount escapes, or that sarcosine is turning over so rapidly as to lose significant label. To allay these problems, it may be necessary to check the blood for labeled creatine and sarcosine and to determine the labeling of methylene-THF, the first product of sarcosine degradation, or to carry out the labeling even more rapidly. Although difficult, such an experiment could go a long way toward establishing the relative roles of the major transmethylation reactions. Ideally, this experiment should be carried out in the context of different intakes of methionine, from deficient to generous, which would permit investigators to determine whether there is a constitutive, physiologic flux through GNMT or whether this enzyme functions chiefly to catabolize excess methionine. Finally, Stead et al (1) mentioned that clarification of the role of betaine as a dietary source of labile methyls is also needed. On the basis of unpublished observations, Craig (40) estimated that the dietary intake of betaine for humans may range from 8.5 to 21.3 mmol/d, amounts that are very significant with respect to the intakes of methionine and choline. However, more recent assessments of betaine dietary intakes have been 1.8 ± 0.8 and 2.54 ± 0.3 mmol/d in New Zealand (41) and the United States (42), respectively. If the latter values turn out to be correct, dietary betaine will make a significant but relatively small contribution to the normal intake of labile methyl groups.

SUMMARY

The following are the major conclusions reached in this discussion. First, methyl balance is maintained in normal humans, if necessary, by the process of methylneogenesis in which a methyl group newly formed on methyl-THF is transferred to homocysteine, which forms methionine (Figure 1). Second, early (and minimal) estimates of total rates of human transmethylation, based on balance studies, are in reasonable agreement with total rates calculated from more recent stable isotope infusion studies, although uncertainties exist as to the upper limit established by the latter approach and as to the effects of sex and age on these rates. Third, of the probable major contributors to total transmethylation (if dietary intakes of creatine in normal diets in the United States are taken into account), the estimated rate of creatine synthesis is significantly reduced from that calculated on the basis of balance studies of humans on diets virtually free of creatine. Fourth, in mice devoid of PEMT activity, plasma tHcy decreases by 40–50% without a significant change in hepatic tHcy concentration. These observations suggest that phosphatidylcholine synthesis is a relatively important contributor to total transmethylation flux, although the extent to which flux through GNMT may increase in these mice is unknown. Definitive measurement of the flux through either PEMT or GNMT in human liver remains to be achieved. We note also the paucity of information on the mechanism or mechanisms of homocysteine export from the liver to the plasma. Not only do we not know the specific transporter or transporters involved, we also require further information on the form of homocysteine that is exported. Fifth, N-methylglycine (sarcosine) is formed by 2 pathways, whose relative contributions are still unclear (Figure 1). However, the important contribution of AdoMet-dependent glycine methylation, especially under circumstances of increased dietary methionine intake, is indicated not only by balance studies of suitable subjects but also by the fact that, in humans deficient in glycine N-methyltransferase, abnormally elevated concentrations of plasma AdoMet are seen. The lack of elevated plasma AdoMet in mice devoid of activity of phosphatidylethanolamine methyltransferase or in humans severely deficient in guanidinoacetate N- methyltransferase activity may possibly be explained by disposition of AdoMet under both circumstances by increased methylation of glycine.

All authors participated in structuring the content of this commentary. SHM wrote the first draft of the manuscript, and all other authors participated in the revision and editing of the manuscript and preparation of the final manuscript. DEV is a Scientist of the Alberta Heritage Foundation for Medical Research and holds the Canada Research Chair in Molecular and Cell Biology of Lipids. JTB is a Canadian Institutes for Health Research Senior Investigator. None of the authors had a personal or financial conflict of interest.

REFERENCES

1

Stead
LM
,
Brosnan
JT
,
Brosnan
ME
,
Vance
DE
,
Jacobs
RL
.
Is it time to reevaluate methyl balance in humans?
Am J Clin Nutr
2006
;
83
:
5
10
.

2

Mudd
SH
,
Poole
JR
.
Labile methyl balances for normal humans on various dietary regimens
.
Metabolism
1975
;
24
:
721
35
.

3

Katz
JE
,
Dlakic
M
,
Clarke
S
.
Automated identification of putative methyltransferases from genomic open reading frames
.
Mol Cell Proteom
2003
;
2
:
525
40
.

4

Rosenblatt
DS
,
Fenton
WA
.
Inherited disorders of folate and cobalamin transport and metabolism
. In:
Scriver
CR
,
Beaudet
AL
,
Valle
D
et al. 
eds.
The metabolic and molecular bases of inherited disease.
New York, NY
:
McGraw-Hill
,
2001
:
3897
933
.

5

Barber
JR
,
Morimoto
BH
,
Brunauer
LS
,
Clarke
S
.
Metabolism of S-adenosyl-L-methionine in intact human erythrocytes
.
Biochim Biophys Acta
1986
;
886
:
361
72
.

6

Mudd
SH
,
Ebert
MH
,
Scriver
CR
.
Labile methyl group balances in the human: the role of sarcosine
.
Metabolism
1980
;
29
:
707
20
.

7

Harris
RC
,
Söderlund
K
,
Hultman
E
.
Elevation of creatine in resting and exercised muscle of normal subjects by creatine supplementation
.
Clin Sci
1992
;
83
:
367
74
.

8

Harris
RC
,
Nevill
M
,
Harris
DB
,
Fallowfield
JL
,
Bogdanis
GC
,
Wise
JA
.
Absorption of creatine supplied as a drink, in meat or in solid form
.
J Sports Sci
2002
;
20
:
147
51
.

9

MacNeil
L
,
Hill
L
,
MacDonald
D
, et al. 
Analysis of creatine, creatinine, creatine-d3 and creatinine-d3 in urine, plasma, and red blood cells by HPLC and GC-MS to follow the fate of ingested creatine-d3
.
J Chromatogr B
2005
;
827
:
210
5
.

10

Storch
KJ
,
Wagner
DA
,
Burke
JF
,
Young
VR
.
Quantitative study in vivo of methionine cycle in humans using [methyl-2H3]- and [1-13C]methionine
.
Am J Physiol Endocrinol Metab
1988
;
255
:
E322
31
.

11

Fukagawa
NK
,
Martin
JM
,
Wurthmann
A
,
Prue
AH
,
Ebenstein
D
,
O'Rourke
B
.
Sex-related differences in methionine metabolism and plasma homocysteine concentrations
.
Am J Clin Nutr
2000
;
72
:
22
9
.

12

MacCoss
MJ
,
Fukagawa
NK
,
Matthews
DE
.
Measurement of intracellular sulfur amino acid metabolism in humans
.
Am J Physiol Endocrinol Metab
2001
;
280
:
E947
55
.

13

Davis
SR
,
Stacpoole
PW
,
Williamson
J
, et al. 
Tracer-derived total and folate-dependent homocysteine methylation and synthesis rates in humans indicate that serine is the main one-carbon donor
.
AmJ Physiol Endocrinol Metab
2004
;
286
:
E272
9
.

14

Mercier
S
,
Breuillé
D
,
Buffière
C
, et al. 
Methionine kinetics are altered in the elderly both in the basal state and after vaccination
.
Am J Clin Nutr
2006
;
83
:
291
8
.

15

Noga
AA
,
Stead
LM
,
Zhao
Y
,
Brosnan
ME
,
Brosnan
JT
,
Vance
DE
.
Plasma homocysteine is regulated by phospholipid methylation
.
J Biol Chem
2003
;
279
:
5952
5
.

16

Shields
DJ
,
Lingrell
S
,
Agellon
LB
,
Brosnan
JT
,
Vance
DE
.
Localization-independent regulation of homocysteine secretion by phosphatidylethanolamine N-methyltransferase
.
J Biol Chem
2005
;
280
:
27339
44
.

17

Jacobs
RL
,
Stead
LM
,
Devlin
C
, et al. 
Physiological regulation of phospholipid methylation alters plasma homocysteine in mice
.
J Biol Chem
2005
;
280
:
28299
305
.

18

Reo
NV
,
Adinehzadeh
M
,
Foy
BD
.
Kinetic analyses of liver phosphatidylcholine and phosphatidylethanolamine biosynthesis using 13C NMR spectroscopy
.
Biochim Biophys Acta
2002
;
1580
:
171
88
.

19

Loo
G
,
Goodman
PJ
,
Hill
KA
,
Smith
JT
.
Creatine metabolism in the pyridoxine-deficient rat
.
J Nutr
1986
;
116
:
2403
8
.

20

Rendeli
C
,
Ausili
E
,
Castorina
M
,
Antuzzi
D
,
Tabacco
F
,
Caldarelli
M
.
Homocysteine, folate, lipid profile and MTHFR genotype and disability in children with myelomeningocele
.
Childs Nerv Syst
2006
;
22
:
1316
21
.
Epub 2000 April 7
.

21

Martos
R
,
Valle
M
,
Morales
R
,
Canete
R
,
Gavilan
MI
,
Sánchez-Margalet
V
.
Hyperhomocysteinemia correlates with insulin resistance and low-grade systemic inflammation in obese prepubertal children
.
Metab Clin Exp
2006
;
55
:
72
7
.

22

Capdevila
A
,
Wagner
C
.
Measurement of plasma S-adenosylmethionine and S-adenosylhomocysteine as their fluorescent isoindoles
.
Anal Biochem
1998
;
264
:
180
4
.

23

Allen
RH
,
Stabler
SP
,
Lindenbaum
J
.
Serum betaine, N,N-dimethylglycine and N-methylglycine levels in patients with cobalamin and folate deficiency and related inborn errors of metabolism
.
Metabolism
1993
;
42
:
1448
60
.

24

Stabler
SP
,
Marcell
PD
,
Podell
ER
,
Allen
RH
.
Quantitation of total homocysteine, total cysteine, and methionine in normal serum and urine using capillary gas chromatography-mass spectrometry
.
Anal Biochem
1987
;
162
:
185
96
.

25

Stabler
SP
,
Lindenbaum
J
,
Savage
DG
,
Allen
RH
.
Elevation of serum cystathionine levels in patients with cobalamin and folate deficiency
.
Blood
1993
;
81
:
3404
13
.

26

Svardal
A
,
Refsum
H
,
Ueland
PM
.
Determination of in vivo protein binding of homocysteine and its relation to free homocysteine in the liver and other tissues of the rat
.
J Biol Chem
1986
;
261
:
3156
63
.

27

Mudd
SH
,
Cerone
R
,
Schiaffino
MC
, et al. 
Glycine N-methyltransferase deficiency: a novel inborn error causing persistent isolated hypermethioninemia
.
J Inherit Metab Dis
2001
;
24
:
448
64
.

28

Augoustides-Savvopoulou
P
,
Luka
Z
,
Karyda
S
, et al. 
Glycine N-methyltransferase deficiency: a new patient with a novel mutation
.
J Inherit Metab Dis
2003
;
26
:
745
59
.

29

Luka
Z
,
Capdevila
A
,
Mato
JM
,
Wagner
C
.
A glycine N-methyltransferase knockout mouse model for humans with deficiency of this enzyme
.
Transgenic Res
2006
;
15
:
393
7
.

30

Thomas
N-E
,
Cooper
S-M
,
Williams
SRP
,
Baker
JS
,
Davies
B
.
Fibrinogen, homocysteine, and C-reactive protein concentrations relative to sex and socioeconomic status in British young people
.
Am J Hum Biol
2005
;
17
:
809
13
.

31

von Figura
K
,
Hanefeld
F
,
Isbrandt
D
,
Stöckler-Ipsiroglu
S
.
Guanidinoacetate methyltransferase deficiency
. In:
Scriver
CR
,
Beaudet
AL
,
Sly
WS
, et al. 
eds.
The metabolic and molecular bases of inherited disease.
New York, NY
:
McGraw-Hill
,
2001
:
1897
908
.

32

Guthmiller
P
,
Van Pilsum
JF
,
Boen
JR
,
McGuire
DM
.
Cloning and sequencing of rat kidney L-arginine:glycine amidinotransferase
.
J Biol Chem
1994
;
269
:
17556
60
.

33

Stead
LM
,
Au
KP
,
Jacobs
RL
,
Brosnan
ME
,
Brosnan
JT
.
Methylation demand and homocysteine metabolism: effects of dietary provision of creatine and guanidinoacetate
.
Am J Physiol Endocrinol Metab
2001
;
281
:
E1095
100
.

34

Korzun
WJ
.
Oral creatine supplements lower plasma homocysteine concentrations in humans
.
Clin Lab Sci
2004
;
17
:
102
6
.

35

Steenge
GR
,
Verhoef
P
,
Greenhaff
PL
.
The effect of creatine and resistance training on plasma homocysteine concentrations in healthy volunteers
.
Arch Intern Med
2001
;
161
:
1455
6
.

36

Datko
AH
,
Aksamit
RR
,
Mudd
SH
.
Phosphatidylcholine synthesis in the rat: the substrate for methylation and regulation by choline
.
Lipids
1990
;
25
:
135
42
.

37

Shields
DJ
,
Lehner
R
,
Agellon
LB
,
Vance
DE
.
Membrane topography of human phosphatidylethanolamine N-methytransferase
.
J Biol Chem
2003
;
278
:
2956
62
.

38

Sehayek
E
,
Wang
R
,
Ono
JG
, et al. 
Localization of the PE methylation pathway and SR-BI to the canalicular membrane: evidence for apical PC biosynthesis that may promote biliary excretion of phospholipid and cholesterol
.
J Lipid Res
2004
;
44
:
1605
13
.

39

Shields
DJ
,
Altarejos
JY
,
Wang
X
,
Agellon
LB
,
Vance
DE
.
Molecular dissection of the S-adenosylmethionine-binding site of phosphatidylethanolamine N-methyltransferase
.
J Biol Chem
2003
;
278
:
35826
36
.

40

Craig
SAS
.
Betaine in human nutrition
.
Am J Clin Nutr
2004
;
80
:
539
49
.

41

Slow
S
,
Donaggio
M
,
Cressey
PJ
,
Lever
M
,
George
PM
,
Chambers
ST
.
The betaine content of New Zealand foods and estimated intake in the New Zealand diet
.
J Food Comp Anal
2005
;
18
:
473
85
.

42

Cho
E
,
Zeisel
SH
,
Jacques
P
,
Selhub
J
,
Dougherty
L
,
Colditz
GA
.
Dietary choline and betaine assessed by food-frequency questionaire in relation to plasma total homocysteine concentration in the Framingham Offspring Study
.
Am J Clin Nutr
2006
;
83
:
905
11
.

FOOTNOTES

2

The views expressed here do not necessarily represent those of the National Institute of Mental Health, the National Institutes of Health, or the US Department of Health and Human Services.

3

Supported by grants from the Canadian Institutes for Health Research (to DEV, JTB, and MEB), grant no. AG-09834 from the National Institute on Aging (to SPS), grant no. DK15289 from the National Institutes of Health and the Research Service of the Department of Veterans Affairs (to CW), and by postdoctoral fellowships from the Canadian Institutes of Health Research and Alberta Heritage Foundation for Medical Research (to RLJ).