Main

Regeneration of methionine from homocysteine plays a vital role in humans in maintaining adequate methionine levels, in the recycling of one carbon units for S-adenosylmethionine-dependent methylation reactions, as well as in the removal of potentially toxic homocysteine(2). This regeneration depends on the function of B12 requiring MeTHF:homocysteine methyltransferase (MS, EC 2.1.1.13,Fig. 1, reaction 1). MeTHF is produced from tetrahydrofolate by the action of a series of enzymes, the last of which is MR(EC 1.1.1.68, Fig. 1, reaction 5). Once formed, MeTHF must be demethylated to allow recycling of intracellular reduced folates(3).

Figure 1
figure 1

Interaction of folate and methionine metabolism. The numbers represent the following enzymes or sequences: 1, MeTHF:homocysteine methyltransferase; 2, 10-formyltetrahydrofolate synthase; 3, 5,10-methenyltetrahydrofolate cyclohydrolase;4, 5,10-methylenetetrahydrofolate dehydrogenase; 5, MR;6, serine hydroxymethyltransferase; 7, Protein synthesis;MeCbl, methyl-cbl; *radioactively marked compounds in the assay of methionine and serine formation. The shaded areas denote the substrates and products in this assay.

Full size image

Disturbance of MS activity leads to a variety of disease states. Inherited enzyme deficiencies that affect the function of MS lead to homocystinuria and low to normal plasma methionine levels. MS deficiency is associated with methylmalonyl CoA mutase (EC 5.4.99.2) deficiency in combined homocystinuria and methylmalonic aciduria due to the cblC, D, or F defect in which there is defective synthesis of both vitamin B12 coenzymes, deoxy-adenosyl-cbl and methyl-cbl(4). Isolated MS deficiency, associated with reduced synthesis of methylcbl, occurs in the cblE and cblG defects(5). These disorders usually result in neurologic abnormalities together with megaloblastic anemia, and all patients have shown some degree of clinical and biochemical improvement with vitamin B12(cbl), although the long-term outcome is poor(6).

MR deficiency which results in low levels of MeTHF is usually associated with severe neurologic abnormalities without hematologic abnormalities, and most patients have not responded to vitamin therapy(3). This disorder results in an abnormal distribution of folates in cultured fibroblasts(7) and liver and kidney(8) with normal intracellular folate concentrations but a much reduced proportion of the 5-methyl form. In contrast there was a reduced total folate concentration with a high proportion of the 5-methyl form in liver and kidney of a patient with MS deficiency due to the cblC defect(8).

Elevated plasma homocyst(e)ine has been well documented in both vitamin B12(9) and folate deficiency(10), reflecting disturbed MS function. In addition MS is inactivated dramatically in rats exposed to nitrous oxide(11). Administration of this gas has led to megaloblastic anemia in humans and to subacute combined degeneration of the spinal cord in monkeys, both mimicking clinical signs of MS deficiency states(12). In one related study, reduced purine synthesis was shown to accompany lowered MS activity in cultured lymphoblasts exposed to nitrous oxide(13). In addition, 14C-labeled formate was shown to be converted to serine, and this was lower in nitrous oxide-treated cells. It is likely that serine formation is due to the activity of serine hydroxymethyltransferase (EC 2.1.2.1). This enzyme has been characterized from rabbit liver(1) and is the final step in the conversion of formate to serine, a process dependent on the availability of reduced folate coenzymes. Previous studies have shown reduced synthesis of methionine from [14C]formate(14) and [35S]homocysteine(15) in cultured fibroblasts from patients with MR deficiency.

We have studied the formation of methionine and serine from radioactively labeled formate as a measure of both MS function and availability of reduced folates in intact fibroblasts in monolayer culture. These activities have been compared in fibroblasts from patients with genetic deficiencies of either MS or MR. This system provides an additional model to nitrous oxide-treated rats or lymphoblasts for the study of interactions of folate and cbl and also provides a valuable diagnostic tool.

METHODS

Cell culture. Cultured skin fibroblasts were established from pinch biopsies obtained with the informed consent of patients, or when appropriate their parents, and grown in Eagle's minimum essential medium containing undialyzed FCS (10% vol/vol), nonessential amino acids (1% wt/vol), and kanamycin (100 mg/L). This medium contains 1 mg/L folic acid and approximately 50 ng/L vitamin B12, which originates from the FCS. Cultures were maintained at 37°C in a 5% CO2/95% air atmosphere and periodically tested to exclude mycoplasma contamination. Cells were subcultured up to a maximum of 25 times or harvested for assay using trypsin(0.25% wt/vol)(16). Fibroblasts were set up for intact cell assay or harvested for specific enzyme assay (stored at -70°C) when just confluent, except for assay of MR, in which case cells were harvested 3 d after confluence. In some experiments the medium was changed after 4 d and replaced by medium supplemented with various vitamins or cofactors.

Specific enzyme assays. Cell pellets were disrupted either by sonication in a 10-fold dilution of the buffer used in the enzyme assay or by standing in the same buffer containing Lubrol (0.15%, wt/vol) for 30 min at 4°C. The supernatant obtained by centrifugation at 10,000 ×g for 10 min at 4°C was used for assays.

MR activity was determined by the method described by Rosenblatt and Erbe(17) except that the radioactive substrate was preextracted with dimedone and toluene to reduce radioactivity in the blank. The total assay volume was 100 μL, and extraction of the product formed in the reaction was achieved with 1 mL of toluene.

MS was assayed essentially as described previously(18) except that the separation of radioactive methionine produced in the assay was achieved batchwise with Dowex-1 anion exchange resin. Serine hydroxymethyltransferase activity was performed by measuring the conversion of [14C]serine and tetrahydrofolate to 5,10-methylenetetrahydrofolate as described by Taylor and Weissbach(19) except that DTT was used instead ofβ-mercaptoethanol as a reducing agent. Enzyme activities were related to protein determined by the Lowry method(20).

Intact cell assay of methionine and serine formation. The previously described method(15) for the determination of methionine formation from homocysteine in fibroblast monolayer cultures was modified by using [14C]formate as the labeled precursor instead of[35S]homocysteine. Duplicate 75-cm2 flasks of confluent cells were washed with 5 mL of Earle's basic salt solution, then incubated for 16 h at 37°C with 5 mL of Earle's basic salt solution that contained[14C]formate (2 μCi/mL), sodium formate (250 μmol/L), and L-homocysteine (200 μmol/L) prepared freshly from its thiolactone as previously described(16). Cells were harvested with 0.25% wt/vol trypsin solution, and proteins were precipitated with 5% trichloroacetic acid and oxidized with performic acid, then hydrolyzed with 6 M hydrochloric acid. Labeled amino acids were separated by high voltage paper electrophoresis, and radioactivity in the area of the electrophoretogram corresponding to oxidized methionine and serine standards, was quantitated by liquid scintillation counting(15). In some experiments the identity of the radioactively labeled amino acids was confirmed by analysis of hydrolysates, which were desalted on cation exchange resin using two-dimensional thin layer chromatography(21). After drying, chromatograms were stained with ninhydrin, and the areas corresponding to the individual amino acids were cut out and counted by liquid scintillation counting. Incorporation of label from [14C]propionate into protein was assayed in intact fibroblast monolayer cultures(22).

Cell lines. Control skin fibroblasts were obtained from healthy controls, from patients with a disease not related to homocysteine metabolism, or from their parents. Eleven cell lines from patients with clinical and biochemical evidence of MR deficiency were studied (designated MR mutant). Twelve cell lines were studied from patients with increased urinary or plasma homocystine and methylmalonic acid and accompanying clinical features(classified as cblC/D mutants). Three cell lines were studied from patients with biochemical and clinical features consistent with a functional deficiency of MS without methylmalonic aciduria. The cell lines were sent from various centers as follows: Dr. A. Burlina, Padua Italy, three lines; Dr. R. Cerone, Genoa Italy, five lines; Dr. A. Green, Birmingham UK, three lines; Dr. Knapp, Braunschweig, Germany; Dr. A. Kohlschutter, Hamburg Germany; Dr. E. Kvittingen, Oslo Norway; Prof. J. Leonard, London UK, three lines; Dr. H. Mandel, Haifa Israel; Dr. J. R. Moore, Leicester UK; Dr. I. B. Sardharwalla, Manchester UK; Dr. R. Schutgens, Amsterdam Netherlands, two lines; Dr. Shawinsky, Johannesburg Republic of South Africa Mr. R. Smith, Bradford UK, two lines; Dr. G. Zeman, Prague, Czech Republic.

Materials and chemicals. Medium, reagents, and disposable plastic ware for tissue culture were obtained from GIBCO Europe.[14C]Formate and 5-[14C]methyltetrahydrofolic acid were purchased from Amersham Corp. Hydroxo-cbl was obtained from Fluka Chemicals. All other chemicals were of the highest purity commercially available from Sigma Chemical Co. or Merck.

RESULTS

Cell line characteristics. MR activities in the 11 cell lines from patients with biochemical and clinical features of this enzyme deficiency and controls are shown in Table 1. Activities were greatly reduced in each patient cell line both with and without FAD, although activities tended to be higher in the presence of added coenzyme as also seen in controls. In mutant cells the level of activity assayed without FAD was 0-2.3% of the mean control value, confirming their classification as MR mutants, whereas activity was less than 6% of control when assayed with added FAD in all but two cell lines, which exhibited appreciable residual activity of 8 and 14%, respectively, of control values.

Table 1 MR activity in mutant and control cultured fibroblasts
Full size table

Incorporation of label from [14C]propionate into trichloroacetic acid-precipitable proteins in cells grown in normal and hydroxo-cbl-supplemented medium in the group of cell lines from patients with clinical and biochemical features of combined methylmalonic aciduria and homocystinuria is shown in Table 2. This activity was reduced in all cell lines when grown in normal medium ranging from 9 to 32% of the mean control value. Importantly this activity increased considerably in each cell line when grown in hydroxo-cbl-supplemented medium, reaching between 68 and 562% higher values in supplemented compared with values in basal medium. These findings together with homocystinuria in the patients confirm the combined defect and allow classification of these cell lines as cblC/D mutants.

Table 2 [14C]Propionate uptake in fibroblasts from patients with combined methylmalonic aciduria and homocystinuria (cblC/D mutants)
Full size table

Three cell lines were studied from patients with clinical features suggestive of a disorder of MS function with hypomethioninemia but without methylmalonic aciduria. In cells from these three patients MS activity was virtually absent in one cell line, which was shown by complementation studies to belong to the cblG class. The two other cell lines exhibited normal activity in the presence of excess, but reduced activity in the presence of low, concentrations of reducing agent and were confirmed as cblE by complementation studies.

Analysis of radioactively labeled amino acids. This is illustrated in hydrolysates of cell proteins by high voltage electrophoresis in Figure 2. In the control, clear peaks of radioactivity with migration which corresponds exactly with that of standards of methionine sulfone and serine are evident. In contrast, both peaks are much smaller in the cells from a cblC patient, whereas in the MR-deficient cells, low counts in the methionine position are accompanied by a large peak in the position of serine. In representative control cell lines two-dimensional thin layer chromatography of cell protein hydrolysates revealed a high number of counts in the positions of serine and methionine sulfone, confirming the identity of these two labeled amino acids. In contrast, only a few counts above background were found in the positions of threonine, aspartic acid, glycine, and histidine, whereas no radioactivity was detected in the remaining protein amino acids.

Figure 2
figure 2

Analysis of labeled amino acids in cell proteins. Fibroblast monolayers were incubated with [14C]formate, and cell proteins were precipitated, oxidized, hydrolyzed, and analyzed by high voltage electrophoresis as described in “Methods.” The positions of marker methionine sulfone (oxidation product of methionine, MetO2) and serine (Ser) are indicated by broken lines: A, control cell line; B, cblC mutant cell line; C, MR mutant cell.

Full size image

Formation of both methionine and serine was shown to be linear with time up to 24 h in duplicate flasks of confluent control fibroblasts. The coefficients of variation (R2) were 0.983 and 0.986 for methionine and serine, respectively.

Formation of methionine and serine in control and mutant fibroblasts. The formation of methionine and serine, in cells grown in normal medium from patients with the three types of defect and controls, is shown in Table 3 and Figure 3.Table 3 summarizes the range, mean, and SD of values obtained in control and mutant cells, whereas the individual values are shown in Figure 3.

Table 3 Formation of labeled methionine and serine in control and mutant fibroblasts grown in normal medium
Full size table
Figure 3
figure 3

Formation of methionine and serine in individual control and mutant cells grown in normal medium. Cells were incubated with labeled formate and harvested, and proteins were precipitated, oxidized, hydrolyzed, and analyzed by high voltage electrophoresis as described in“Methods.” Note the logarithmic scale.

Full size image

The control cells show clearly measurable formation of both methionine and serine with clearly, similarly reduced methionine formation in each type of remethylation defect with a mean value of 7.5, 3, and 3.5% of the mean control value for MR, cblC/D, and MS mutant cells, respectively. In both cblC/D and MS mutant cells, there is a simultaneous reduction of labeled serine formation of 9.5 and 10% of the mean control value, respectively. In contrast, MR mutant cells show on average 175% higher serine formation than control cells, although most values fall within the upper limit of the control range. Inspection of the values in the individual cell lines (Fig. 3) reveals that methionine formation is clearly lower than the lowest control value in all mutant cells (at the most 25%), whereas serine formation is also reduced in every cblC/D and MS mutant cell line (at the most 41%) without overlap.

The formation of methionine in 10 and serine in 9 of the cbl mutant cells, grown for 3 d in various concentrations of hydroxo-cbl before assay, is shown in Figure 4. All cell lines showed a significant increase of the reduced methionine and serine formation in the higher B12-containing media, sometimes reaching the control range values. Most required the highest concentration of 1000 μg/L to elicit the maximum response, whereas exceptional cell lines did this even at 100 μg/L.

Figure 4
figure 4

Formation of methionine and serine in cells grown with varying amounts of cbl in the medium. Cells were grown for 4 d in normal medium then for 3 d in medium supplemented with hydroxo-cbl at the depicted concentrations. Cells were incubated with labeled formate and harvested, and proteins were precipitated, oxidized, hydrolyzed, and analyzed by high voltage electrophoresis as described in “Methods.”

Full size image

Serine hydroxymethyltransferase activity in control cultured fibroblasts. Clearly measurable serine hydroxymethyltransferase activity was found in 10 control cell lines with a slight stimulation of activity in the presence of added pyridoxal 5′-phosphate (Table 4). This activity was dependent on time with linearity up to 60 min and on the amount of cell extract protein included in the assay with linearity up to 230 μg of protein (Fig. 5). The relation of activity to varying concentrations of the two substrates is shown in Lineweaver-Burke linear plots in Figure 6. These kinetic studies revealed Km values of 0.47 mmol/L for serine and 0.38 mmol/L for tetrahydrofolate.

Table 4 Serine hydroxymethyltransferase activity in control cultured fibroblasts
Full size table
Figure 5
figure 5

Effect of varying incubation time (A) and varying amount of cell extract (B) on serine hydroxymethyltransferase activity in cultured fibroblasts. The activity was measured as described under “Methods.” In experiment A, the protein content was 113 μg and in experiment B the incubation time was 60 min.

Full size image
Figure 6
figure 6

Effect of varying substrate concentrations on serine hydroxymethyltransferase activity in control cultured fibroblasts.(A) Varied serine concentrations with tetrahydrofolate(THF) maintained at 2 mmol/L. (B) Varied THF concentration with serine maintained at 1.25 mmol/L.

Full size image

DISCUSSION

These studies using cultured fibroblasts from patients with clinical, biochemical, and enzymologic evidence of three different types of remethylation defect provide evidence for the presence of the involved pathways and information on their disturbance in this in vitro model system. Incorporation of label from [14C]formate into methionine and serine, but not to any significant extent in other amino acids, reflects the existence in cultured fibroblasts of the pathway from formate to MeTHF illustrated in Figure 1. This probably involves the coupled cyclic system described by Strong and Schirch(1) based on studies in rabbit liver. In this system formate is converted to 5,10-methylenetetrahydrofolate by the action of 10-formyltetrahydrofolate synthase, 5,10-methenyltetrahydrofolate cyclohydrolase, and 5,10-methylenetetrahydrofolate dehydrogenase. In the next step, 5,10-methylenetetrahydrofolate is reduced by MR to MeTHF, whose methyl group, derived from formate, is transferred to homocysteine by MS. Finally, 5,10-methylenetetrahydrofolate donates its C1 unit to glycine, catalyzed by serine hydroxymethyltransferase, yielding serine and releasing tetrahydrofolate, which can react once more with formate, thereby completing this cycle. That serine formation reflects activity of serine hydroxymethyltransferase is supported by the demonstration, in this study, of the presence of this enzyme activity in cultured fibroblasts by specific assay.

The clearly reduced formation of labeled methionine, using labeled formate as a precursor, in cells with each of the different remethylation defects indicates the critical role of both MR and MS in remethylation of homocysteine. This extends previous observations of defective methionine formation in MR mutants using labeled homocysteine(15), or labeled formate(14).

An important additional finding in the present study is the clearly reduced formation of serine in cells with deficient MS activity, so far shown only in nitrous oxide-treated lymphoblasts(13). This is explained by a lack of availability of reduced folate coenzymes, especially 5,10-methylenetetrahydrofolate, as shown in the liver of such patients(8) due to lack of release of C1 units from MeTHF,i.e. methyl group trapping. This idea is supported by the opposite finding of elevated serine formation in MR-deficient cells. Presumably this is due to 5,10-methylenetetrahydrofolate accumulation, which would be expected because non-methyl forms of folate accumulate in tissues and fibroblasts in this defect, leading to increased serine hydroxymethyltransferase activity.

Taken together these findings in control and different types of mutant cells provide clear evidence for the existence of the formate to serine pathway of Strong and Schirch(1) in cultured fibroblasts. They also provide further evidence that reduced availability of reduced folate coenzymes is associated with disturbed MS function in cell lines with a specific genetic disorder.

There is much evidence that reduced availability of folate coenzymes occurs in functional MS deficiency due to the so-called methyl trap hypothesis(12). According to this theory, once MeTHF is formed it cannot be reoxidized by MR, because this reaction is irreversible under physiologic conditions. The ensuing functional folate deficiency leads to low levels of folate coenzymes, thus impairing other folate-dependent reactions. However, this theory has been reputed by some workers who provided evidence that disturbed folate metabolism in cbl deficiency is best explained by formate starvation rather than methyl group trapping(23, 24). Also the reduction of quinonoid dihydrobiopterin by MeTHF, catalyzed by MR, has been proposed as an alternative mechanism for recycling of reduced folate coenzymes(25).

This genetic model of MS deficiency, in a system plentiful in other related nutrients and with normal activities of other enzymes, clearly demonstrates the consequences of a single enzyme block on folate coenzyme availability. Importantly, the correction of MS activity, in the cells with the cbl defect, by hydroxo-cbl was also associated with improvement of serine formation and presumably correction of the deficiency of reduced folates. Thus in this study vitamin B12 corrected methyl trapping, independent of changes in other constituents of the medium, for example folate. The idea that formate deprivation is the cause of decreased availability of reduced folates seems less likely, because methionine and formate levels in this culture system and assay, respectively, are relatively high and no different in those experiments where hydroxo-cbl improved both MS activity and serine formation. This idea is supported by other studies, for example the observation by Hoffbrand and Jackson(26) that the DNA synthesis defect in vitamin B12 deficiency can be corrected by tetrahydrofolate administration.

The increase of the abnormally low serine formation in response to supplementation with hydroxo-cbl in the cblC/D mutant cell lines correlates well with the clinical response in the patients. For example, the megaloblastic anemia, which is a key feature of the cblC/D defect(3, 27), probably results from low availability of reduced folates and consequent scarcity of purine and pyrimidine precursors of DNA synthesis. Correction of this defect in vivo and in vitro by vitamin B12 provides strong evidence that the low serine formation in fibroblasts reflects low reduced folates in these mutant cells with a functional deficiency of MS and therefore fits in well with the methyl trap hypothesis.

Finally these studies demonstrate that normal and mutant cultured skin fibroblasts provide an appropriate model for the study of the interaction of methionine and folate metabolism and cofactor and vitamin influences. Also the measurement of methionine and serine from labeled formate in cultured fibroblasts is invaluable in the enzymatic confirmation of remethylation defects and can also be used for prenatal diagnosis(28).