-
PDF
- Split View
-
Views
-
Cite
Cite
Matthew J. Rowling, Mary H. McMullen, Kevin L. Schalinske, Vitamin A and Its Derivatives Induce Hepatic Glycine N-Methyltransferase and Hypomethylation of DNA in Rats, The Journal of Nutrition, Volume 132, Issue 3, March 2002, Pages 365–369, https://doi.org/10.1093/jn/132.3.365
Close -
Share
Abstract
Regulation of S-adenosylmethionine (SAM) and the SAM/S-adenosylhomocysteine (SAH) ratio by the key cytosolic enzyme glycine N-methyltransferase (GNMT) is essential in optimizing methyl group supply and subsequent functioning of methyltransferase enzymes. Therefore, inappropriate activation of GNMT may lead to the loss of methyl groups vital for many SAM-dependent transmethylation reactions. Previously, we demonstrated that the retinoid derivatives 13-cis- (CRA) and all-trans-retinoic acid (ATRA) mediated both the activity of GNMT and its abundance. The present study was conducted to determine whether vitamin A had a similar ability to up-regulate GNMT and to assess the biological importance of GNMT modulation by examining both the transmethylation and transsulfuration pathways after retinoid treatment. Rats were fed a control (10% casein + 0.3% l-methionine) diet and orally given retinyl palmitate (RP), CRA, ATRA or vehicle daily for 10 d. RP, CRA and ATRA elevated hepatic GNMT activity 32, 74 and 124%, respectively, compared with the control group. Moreover, the retinoid-mediated changes in GNMT activity were reflected in GNMT abundance (38, 89 and 107% increases for RP-, CRA-, and ATRA-treated rats, respectively). In addition, hepatic DNA, a substrate for SAM-dependent transmethylation, was hypomethylated (∼100%) after ATRA treatment compared with the control group. In contrast, the transsulfuration product glutathione was unaffected by retinoid treatment. These results provide evidence of the following: 1) vitamin A, like its retinoic acid derivatives, can induce enzymatically active GNMT; and 2) inappropriate induction of GNMT can lead to a biologically important loss of methyl groups and the subsequent impairment of essential transmethylation processes.
Methyl groups, provided from the diet or the folate-dependent one-carbon pool, are essential for ensuring optimal health as well as in the prevention of disease. Methyl group deficiency leads to down-regulation of important transmethylation reactions, and subsequent pathological conditions such as hepatocarcinogenesis (1,2). Methyl groups supplied from the diet, in the form of choline and methionine, or from the folate-dependent one-carbon pool, must be activated to S-adenosylmethionine (SAM)3 to serve as substrates in numerous transmethylation reactions. Because S-adenosylhomocysteine (SAH) is a product of transmethylation reactions and a potent inhibitor of most SAM-dependent methyltransferases (3), the ratio of SAM/SAH is considered an important index of transmethylation potential (4,5). Therefore, not only is the intracellular supply of SAM critical for transmethylation, but the regulation of this ratio is important as well.
The cytosolic enzyme, glycine N-methyltransferase (GNMT), functions to optimize transmethylation reactions by regulating the SAM/SAH ratio (Fig. 1). When methyl groups are abundant and SAM levels are elevated, GNMT disposes of the excess methyl groups by forming the essentially inactive metabolite sarcosine from glycine. SAM also reduces the supply of methyl groups originating from the one-carbon pool by inhibiting 5,10-methylenetetrahydrofolate reductase (MTHFR) (6,7), the enzyme responsible for the synthesis of 5-methyltetrahydrofolate (5-methyl-THF), the folate coenzyme that donates its methyl group to homocysteine to form methionine. Because 5-methyl-THF also binds to GNMT and inhibits its activity (8,9), a decrease in 5-methyl-THF levels due to inhibition of MTHFR by SAM results in an increase in the activity of GNMT. Conversely, under conditions of decreased methyl groups and SAM, the inhibition of MTHFR is removed, leading to an increase in 5-methyl-THF concentrations and subsequent inhibition of GNMT. This ensures that methyl groups are conserved for important transmethylation reactions when methyl group availability is compromised. Therefore, factors that inappropriately activate GNMT may lead to the down-regulation of numerous methyltransferases that are important in the maintenance of optimal health.
Methyl group metabolism. Methyl groups, as methionine, can be converted to SAM and function as a substrate in numerous SAM-dependent transmethylation reactions. GNMT regulates the transmethylation potential by controlling the SAM/SAH ratio. After the hydrolysis of SAH, a branch point exists at which homocysteine can either be remethylated by 5-methyl-THF to generate methionine, or used for the synthesis of biologically significant metabolites such as cysteine and glutathione via the transsulfuration pathway. Abbreviations used: GNMT; glycine N-methyltransferase, SAH; S-adenosylhomocysteine, SAM; S-adenosylmethionine, THF; tetrahydrofolate.
In addition to transmethylation, activation of GNMT may have important consequences on the metabolism of homocysteine, a subsequent product of SAH hydrolysis. After SAH is converted to homocysteine, a metabolic branch point exists at which homocysteine is either remethylated to generate methionine or irreversibly committed to the transsulfuration pathway where it can serve as a precursor to biologically important compounds such as cysteine and glutathione (Fig. 1). Currently, there is no evidence of an association between activation of GNMT and alterations in components of the transsulfuration pathway such as glutathione or homocysteine.
Previously, we have shown that methionine catabolism appears to be enhanced in rats given the retinoid compound 13-cis-retinoic acid (CRA) (10). Recently, we reported that a potential mechanism for the retinoid-induced increase in methionine/SAM catabolism was the activation and induction of hepatic GNMT mediated by both CRA and all-trans-retinoic acid (ATRA) (11). Two obvious questions emerged from the results of that study. First, does vitamin A modulate hepatic GNMT? Second, is the induction of enzymatically active GNMT by retinoids sufficient to compromise the hepatic transmethylation and/or transsulfuration pathways? In the present study, we have found that vitamin A does mediate an increase in GNMT, resulting in an inability to maintain SAM-dependent methylation of DNA. Furthermore, the increase in GNMT abundance by retinoid administration produces GNMT protein that exists primarily in its enzymatically active tetrameric state.
MATERIALS AND METHODS
Chemicals and reagents.
Reagents were obtained from the following: S-adenosyl-L-[methyl-3H]methionine for GNMT activity, New England Nuclear (Boston, MA); S-adenosyl-L-[methyl-3H]methionine for DNA methylation assay and ECL Western blotting detection reagents, Amersham Pharmacia (Piscataway, NJ); phenylmethylsulfonylflouride and dimethyl sulfoxide (DMSO), Calbiochem (La Jolla, CA); disuccinimidyl suberate (DSS), Pierce Chemical (Rockford, IL); goat anti-rabbit immunoglobulin G horseradish peroxidase, Southern Biotechnology (Birmingham, AL); S-adenosyl-L-methionine and glutathione reductase, Sigma Chemical (St. Louis, MO); and Sss I methylase, New England Biolabs (Beverly, MA). GNMT antibodies were kindly provided by Conrad Wagner, Vanderbilt University. Retinyl palmitate (RP), CRA and ATRA were provided courtesy of Hoffman-LaRoche (Nutley, NJ). All other chemicals were of analytical grade.
Animals and diets.
All animal experiments were approved and conducted in accordance with Iowa State University Laboratory Animal Resources Guidelines. Male Sprague-Dawley (Harlan Sprague Dawley, Indianapolis, IN) rats were housed in plastic cages with a 12-h light:dark cycle and given free access to food and water. The control diet was the same as described previously (10). After a 7-d acclimation period during which rats were adapted to both the diet and oral administration of corn oil, rats were divided into four treatment groups (5 rats/group) and were orally given vehicle (corn oil, 1 μL/g body), RP, CRA or ATRA on a daily basis for 10 d. Retinoids were prepared in corn oil and administered at a level of 30 μmol/kg body. Although pharmacologic in magnitude, this dosage was similar to levels used previously (10,11), thus maintaining the continuity of this research and allowing comparisons to be drawn. However, current research using more physiologic levels of retinoids (∼1 μmol/kg body) is important as well. After the 10-d treatment period, rats were anesthetized and portions of liver were removed for analysis of GNMT activity, GNMT protein abundance, DNA methylation and total glutathione concentration.
Measurement of GNMT activity and protein abundance.
The enzymatic activity of GNMT was measured using the method of Cook and Wagner (12) with minor modifications. Portions of liver were homogenized in three volumes of ice-cold phosphate buffered (10 mmol/L, pH 7.0) sucrose (0.25 mol/L) containing 1mmol/L EDTA, 1 mmol/L sodium azide and 0.1 mmol/L phenylmethylsulfonylflouride. After centrifugation at 20,000 × g for 30 min, the resulting supernatant was removed and 2-mercaptoethanol was added to a final concentration of 10 mmol/L. The assay reaction mixture (100 μL) contained 0.2 mol/L Tris buffer (pH 9.0), 5 mmol/L dithiothreitol, 2 mmol/L glycine and 0.2 mmol/L S-adenosyl-L-[methyl-3H]methionine (47.7 kBq/μmol). The reaction was initiated upon the addition of 250 μg of sample protein and was performed in triplicate. GNMT protein abundance was measured using immunoblotting followed by chemiluminescence detection as described previously (11). Samples (75 μg total protein) were subjected to SDS-PAGE using a 10–20% gradient gel and were electrophoretically transferred to a nitrocellulose membrane. For the chemiluminescent detection of GNMT, the membrane was incubated with affinity-purified polyclonal GNMT antibody followed by goat anti-rabbit horseradish peroxidase. Densitometric analysis was performed using the NIH Image software. For the determination of total soluble protein concentration in liver extracts, a commercial kit (Coomassie Plus, Pierce, Rockford, IL) based on the method of Bradford (13) was used with bovine serum albumin as the standard.
Chemical cross-linking of endogenous GNMT.
For determining the oligomeric state of GNMT after retinoid treatment, liver extracts were incubated with 1.5 mmol/L DSS in DMSO for 30 min at room temperature followed by the addition of 6 μL ethanolamine to terminate the cross-linking reaction (14). Samples were subjected to SDS-PAGE followed by immunoblotting and chemiluminescence detection as described above.
DNA methylation.
To determine whether retinoids induced hypomethylation of hepatic DNA (i.e., altered transmethylation), an assay to measure the in vitro incorporation of methyl groups into DNA was used (15) with minor modifications. DNA was purified from liver samples based on the method of Miller et al. (16) using a commercial kit (Promega, Madison, WI). All isolated DNA samples exhibited an A260/280 ratio > 1.8 and consisted of DNA > 20 kb as determined by agarose gel electrophoresis and ethidium bromide staining. The assay mixture consisted of 1.0 μg DNA, 2.7 μmol/L S-adenosyl-L-[methyl3H]methionine (555 GBq/mmol) and reaction buffer [10 mmol/L Tris buffer (pH 7.9), 50 mmol/L NaCl, 10 mmol/L EDTA, 1 mmol/L dithiothreitol] in a total volume of 50 μL. The reaction was initiated upon the addition of Sss I methylase (4 U) and was carried out at 30°C for 1 h. After the reaction was stopped by heating the samples at 65°C for 20 min, the mixture was applied to Whatman DE-81 ion exchange filters (Fisher Scientific, Itasca, IL) fixed on a suction apparatus and washed successively with 20 mL 500 mmol/L sodium phosphate buffer (pH 7.0), 2 mL 70% ethanol and 2 mL absolute ethanol. The filters were allowed to air dry and subjected to liquid scintillation counting.
Total liver glutathione.
For the measurement of total hepatic glutathione concentrations, portions of liver were homogenized in 2 volumes of 0.4 mol/L perchloric acid followed by centrifugation at 10,000 × g for 10 min at 4°C. The resulting supernatant was diluted 100-fold with 125 mmol/L sodium phosphate buffer (pH 7.5) containing 6.3 mmol/L EDTA, and 30 μL was used to spectrophometrically measure total glutathione concentrations as described by Tietze (17).
Statistical analysis.
The means of each treatment group were subjected to a one-way ANOVA (P ≤ 0.05) and compared using Fisher's least significant difference procedure (18). For the GNMT activity data in Figure 2A, the mean of each retinoid-treated group was also individually compared to the control group using Dunnett's Test. This second post-hoc test determines significant differences between a given treatment group and the control group only, and does not compare mean values across treatment groups. For Figure 3, an association between GNMT activity and protein abundance across treatment groups was determined using the Pearson Correlation procedure. All statistical analysis was performed using SigmaStat (SPSS, Chicago, IL).
Retinoid administration activated hepatic glycine N-methyltransferase (GNMT) in rats administered retinyl palmitate (RP), 13-cis-retinoic acid (CRA) or all-trans-retinoic acid (ATRA). Panel A: elevation of hepatic GNMT activity. All assays were performed in triplicate. Panel B: induction of GNMT protein. The immunoblot above the bar graph (panel B) is a representative example with the relative fold induction located below each lane. Data are means ± SEM, n = 5. Bars without a common letter differ, P ≤ 0.05. Bars with an asterisk differ from the control, P ≤ 0.05.
Correlation between glycine N-methyltransferase (GNMT) activity and abundance in rats administered retinyl palmitate (RP), 13-cis-retinoic acid (CRA) or all-trans-retinoic acid (ATRA). Symbols represent individual values within each group. Results of a Pearson correlation test are indicated by the solid line (r = 0.772, P = 0.00007).
RESULTS
Retinoids did not alter the growth rate or relative liver size of rats.
As we reported previously (11), all rat groups exhibited similar growth patterns (data not shown) and no significant differences in initial (65 ± 1 g) or final (157 ± 2 g) body weights were observed among groups. Similarly, relative liver size was not affected by retinoid administration (data not shown).
Retinoids activate hepatic GNMT.
The ability of RP, CRA and ATRA to activate GNMT is shown in Figure 2A. Both CRA and ATRA significantly elevated the activity of GNMT 74 and 124%, respectively, compared with controls. Similarly, RP administration tended to increase (P = 0.052) GNMT activity (32%). The increase in GNMT activity due to RP administration was significant (P = 0.020) when the retinoid treatment groups were compared with the control group alone (denoted by the asterisk).
Retinoid compounds, including vitamin A, induce GNMT abundance.
Figure 2B illustrates the ability of RP, CRA and ATRA to induce significant hepatic production of GNMT protein. The bar graph reflects the mean from all experimental rats, whereas the immunoblot above is a representative example with the relative fold induction located under each lane. All three retinoids, including RP, significantly induced GNMT protein abundance. The mean induction of GNMT protein due to retinoid treatment ranged from 38 to 107%. Interestingly, it appears that all of the retinoid-induced synthesis of GNMT resulted in enzymatically active protein. As shown in Figure 3, GNMT protein abundance was positively correlated with the increase in GNMT activity (r = 0.772, P < 0.001). The ability of retinoids to induce hepatic GNMT in its enzymatically active tetrameric state is further illustrated in Figure 4. Using DSS to cross-link subunits, we determined the oligomeric state of GNMT after retinoid treatment. Similar to Figure 2B, all three retinoids induced GNMT monomer (32 kDa) synthesis (lanes 1–4). The same samples were also analyzed after incubation with the cross-linking reagent DSS (lanes 5–8). DSS treatment of both control and retinoid samples demonstrated that little, if any, of the newly synthesized GNMT protein remains as a monomer, and the majority of the GNMT protein was in its tetrameric (128 kDa) enzymatically active form. No GNMT was detected in the dimeric (64 kDa) form (data not shown).
The enzymatically active tetrameric form of glycine N-methyltransferase (GNMT) was mediated in rats administered retinyl palmitate (RP), 13-cis-retinoic acid (CRA) or all-trans-retinoic acid (ATRA). Cytosolic extracts were subjected to the cross-linking reagent disuccinimidyl suberate (DSS) before SDS-PAGE and immunoblotting using a polyclonal GNMT antibody. As indicated by the arrows, bands at ∼32 kDa represent the monomeric form of GNMT, whereas the tetrameric form is shown at a molecular weight of ∼128 kDa monomer. No immunoreactive bands were discernible at 64 kDa, representing the dimeric form of the protein (data not shown). Hepatic extracts from control, RP-, CRA- and ATRA-treated rats in the absence of DSS treatment are shown in lanes 1–4, respectively, whereas the same samples subjected to DSS cross-linking before SDS-PAGE are shown in lanes 5–8.
Retinoid treatment induced hepatic DNA hypomethylation, but failed to alter hepatic glutathione levels.
The effects of RP, CRA, and ATRA on hepatic DNA methylation status are shown in Figure 5. Compared with control rats, hepatic DNA isolated from ATRA-treated rats exhibited a greater (∼100%) ability to incorporate methyl groups from SAM into hepatic DNA, indicating that a significant reduction in endogenous methylation status was present after retinoid treatment. The mean level of DNA methylation exhibited by RP- and CRA-treated rats was not different from that of controls. To determine whether the disruption in the transmethylation pathway due to retinoid administration had a potential effect on specific components of the transsulfuration pathway, the hepatic concentration of total glutathione was assessed. No significant differences in glutathione levels or the total glutathione content of the liver were observed among groups (data not shown).
Administration of all-trans-retinoic acid (ATRA) to rats resulted in hepatic DNA hypomethylation. In this assay, the ability to serve as an acceptor for methyl groups is inversely proportional to endogenous methylation status. Data are means ± SEM, n = 5. Bars without a common letter differ, P < 0.05.
DISCUSSION
The regulation of transmethylation potential by GNMT is vital to ensure that an optimal supply of methyl groups is available for SAM-dependent transmethylation reactions. However, if GNMT function is not regulated in a manner that is appropriate for the nutritional and/or physiologic conditions of the cell, unwarranted catabolism and subsequent loss of methyl groups may occur. Ultimately, this may contribute to a number of adverse conditions in the liver, as has been reported for ethanol administration (19) and folate deficiency (20). Previously, we reported that the activation and induction of GNMT by CRA and ATRA (11) may serve as the mechanistic basis for the increased methionine/SAM catabolism associated with retinoid treatment (10,21,22). The current studies extend these findings by demonstrating that vitamin A, as retinyl palmitate, also has the ability to enhance methyl group metabolism via the induction of GNMT, although to a lesser degree. The metabolism of vitamin A to its retinoic acid derivative and the regulation of this process is a likely factor in the reduced sensitivity of methyl group metabolism to the actions of retinyl palmitate.
Understanding specifically how retinoids regulate GNMT remains a primary focus for future research. However, it is evident from these studies that although post-translational control of the enzyme (8,9) may contribute to this regulation, a significant portion is at the level of transcription and/or translation, as indicated by the retinoid-mediated increase in abundance of GNMT. It appears that activation of the enzyme by all three retinoids is directly proportional to the induction of GNMT protein, a clear indication that the induction of GNMT synthesis by retinoids plays an important role in the increase in its activity. Moreover, our cross-linking studies suggest that induction of GNMT by retinoids results in the protein existing primarily in its enzymatically active homotetrameric state. Newly synthesized immunoreactive protein was not present to any great extent in its monomeric or dimeric form, the latter required for GNMT to function as a polycyclic aromatic hydrocarbon-binding protein (14,23).
Our findings also demonstrate that up-regulation of GNMT results in biologically important alterations in methyl group metabolism. This may not be surprising because GNMT represents 1–3% of total cytosolic protein in the liver (3). The loss of methyl groups due to the activation of GNMT by retinoids appeared to compromise SAM-dependent transmethylation reactions, namely, the methylation of DNA. Rats treated with ATRA not only exhibited the greatest activation of GNMT, they also exhibited the greatest decrease in endogenous methylation status of hepatic DNA, indicating that a functional deficiency of methyl groups may exist. The potential for retinoid treatment to compromise SAM-dependent transmethylation is also supported by the observation that rats treated with ATRA exhibited a significant reduction in creatinine synthesis (McMullen et al., unpublished observations). Because creatinine synthesis is one of the major depots for methyl groups from SAM, this further stresses the physiologic importance of retinoid-mediated alterations in GNMT.
Retinoid-mediated hypomethylation of DNA has significant implications due to the link between DNA methylation and a number of important processes, such as gene expression and development (24–27). During the course of these and other studies, it was evident that rats within a treatment group exhibited a large degree of variability with respect to methylation status of DNA. Our treatment period (10 d) was fairly short compared with other reports on hypomethylation of DNA that typically employ treatment periods ranging from 1 to 4 wk using methyl- and/or folate-deficient diets (15,27,28). We are currently examining the time course of retinoid-mediated changes in methyl group metabolism, which we expect will aid in explaining this variability. Although only ATRA-treated rats exhibited a significant increase in hypomethylation of DNA, it is important to note that we found a positive correlation (r = 0.667, P = 0.0013) between GNMT activity and DNA hypomethylation (data not shown).
In contrast to SAM-dependent transmethylation, retinoid-treated rats failed to exhibit significant changes in hepatic concentrations of total glutathione, indicating that the transsulfuration pathway remains partially intact, at least with respect to maintaining normal glutathione levels. Earlier work demonstrated that increased methionine catabolism by dietary CRA resulted in a significant increase in hepatic taurine concentrations that was achieved in part at the expense of reduced inorganic sulfate excretion and diminished hepatic glutathione levels (10,29).
In summary, it is clear that retinoids, as vitamin A or as a retinoic acid derivative, represent a potent group of compounds capable of perturbing methyl group metabolism. Although the doses utilized in these studies are pharmacologic in magnitude, preliminary work suggests that modulation of GNMT can be achieved using more physiologic levels (∼1 μmol/kg body) as well. Thus, future work will examine this issue by performing both dose-response and time-course studies. Moreover, the potential effect of retinoid administration on perturbation of the folate-dependent one-carbon pool and subsequent remethylation of homocysteine will be a component of these future studies. This research direction is supported by previous studies demonstrating that vitamin A status influences the one-carbon pool (30,31) as well as a recent report that plasma homocysteine levels were elevated in patients receiving CRA therapy (32).
LITERATURE CITED
Abbreviations
- ATRA
all-trans-retinoic acid
- CRA
13-cis-retinoic acid
- DMSO
dimethyl sulfoxide
- DSS
disuccinimidyl suberate
- GNMT
glycine N-methyltransferase
- MTHFR
5,10-methylenetetrahydrofolate reductase
- RP
retinyl palmitate
- SAH
S-adenosylhomocysteine
- SAM
S-adenosylmethionine
- THF
tetrahydrofolate





