- Email: [email protected]
Phenotypes and circadian rhythm in utilization of formate in purine nucleotide biosynthesis de novo in adult humans
Life Sciences 88 (2011) 688–692
Contents lists available at ScienceDirect
Life Sciences j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / l i f e s c i e
Phenotypes and circadian rhythm in utilization of formate in purine nucleotide biosynthesis de novo in adult humans Joseph E. Baggott a, Gregory S. Gorman b, Sarah L. Morgan a,c,⁎ a b c
Department of Nutrition Sciences, University of Alabama at Birmingham, Birmingham, AL 35294, USA McWhorter School of Pharmacy, Samford University, Birmingham, AL 35229, USA Department of Medicine, University of Alabama at Birmingham, Birmingham, AL 35294, USA
a r t i c l e
i n f o
Article history: Received 3 May 2010 Accepted 28 January 2011 Keywords: Purine nucleotide biosynthesis Formate Humans Phenotypes Circadian rhythm
a b s t r a c t Aims: Folate coenzymes and dependent enzymes introduce one carbon units at positions 2 (C2) and 8 (C8) of the purine ring during de novo biosynthesis. Formate is one source of one-carbon units. Although much is known about lower organisms, little data exists describing formate utilization for purine biosynthesis in humans. Main methods: Mass-spectrometric analysis of urinary uric acid, the final purine catabolite, following 1.0 g oral doses of sodium [13C] formate was performed and detected 13C enrichment at C2 and C8 separately. Key findings: Three phenotypes were suggested. One incorporates 13C 0.72 to 2.0% into C2 versus only 0 to 0.07% into C8. Another incorporates only 0 to 0.05% 13C into C2 or C8. A third phenotype incorporates 13C into C8 (0.15%) but C2 incorporation (0.44%) is still greater. In subjects who incorporated 13C formate into C2, peak enrichment occurred in voids from 8–12 h (24 h clock) suggesting a circadian rhythm. Significance: Evidence that mammalian liver introduces C8 and that C2 is introduced in a non-hepatic site would explain our results. Our data are not similar to those in non-mammalian organisms or cells in culture and are not consistent with the hypothesis that formate from folate-dependent metabolism in mitochondria is a major one carbon source for purine biosynthesis. Timing of peak 13C enrichment at C2 corresponds to maximal DNA synthesis in human bone marrow. Phenotypes may explain the efficacy (or lack of) of certain anticancer and immunosuppressive drugs. © 2011 Elsevier Inc. All rights reserved.
Introduction During purine nucleotide biosynthesis de novo (PNB), two folatecoenzyme-dependent transformylases, aminoimidazolecarboxamide ribotide (AICAR) and glycinamide ribotide (GAR), introduce a carbon at the formyl-oxidation state into positions 2 (C2) and 8 (C8) of the purine ring, respectively as shown in Fig. 1 (Garrett and Grisham, 2005). Carbon at C8 is incorporated first followed by carbon at C2 in the PNB pathway. Formate, one source of these carbons (Reaction 1, Fig. 1), occurs in human plasma at ≥20 μM concentrations, and oral doses of formate maximally increase plasma formate concentrations in only 30–60 min and rapidly expand the in vivo formate pool (Hanzlik et al., 2005). Peaks in labeled formate incorporation into urinary uric acid, the final product of purine catabolism, appear in 1 to 3 days after dosing of labeled formate in humans (Stahelin et al., 1970; Baggott et al., 2007b) and the label was found almost exclusively in C2 and C8 of purines when labeled formate was given to laboratory ⁎ Corresponding author at: Department of Medicine, University of Alabama at Birmingham, Birmingham, AL 35294, USA. Tel.: + 1 205 934 3235; fax: + 1 205 996 2072. E-mail address: [email protected] (S.L. Morgan). 0024-3205/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.lfs.2011.02.007
animals (Baggott et al., 2007b). Isotope exchange of the C2 and C8 positions of purines does not occur in vivo (Abrams, 1956; Bennett and Karlsson, 1957). Therefore, 13C-labeled formate incorporation into C2 and C8 of uric acid can be used to measure how and the extent to which this compound is used in PNB. In cultured cells, formate is incorporated equally into the C2 and C8 positions (Jeong and Schirch, 1996; Kastanos et al., 1997; Fu et al., 2001). A similar pattern is found in rodents and chickens (Drysdale et al., 1951; Marsh, 1951). In the literature, consistent data on humans is lacking. Therefore, this study was conducted to detect patterns of C2 and C8 incorporation (phenotypes) and a circadian rhythm in the utilization of formate for PNB. Using a liquid-chromatography mass spectrometric method (LC/MS/MS) (Gorman et al., 2003), we measured 13C enrichment of urinary uric acid in six human subjects who were given an oral dose of [13C] sodium formate. We found distinct phenotypes in 13C-enrichment pattern at the C2 and C8 positions of uric acid and a circadian rhythm in 13C-enrichment at C2. These results are discussed with respect to the PNB process in adult humans. Phenotypes in formate utilization in PNB may be useful in explaining efficacy (or lack of) of cancer chemotherapeutic and immunosuppressive drugs (e.g. methotrexate) that interfere with PNB. We also discuss our data with respect to the hypothesis: that
J.E. Baggott et al. / Life Sciences 88 (2011) 688–692
689
Fig. 1. Diagrammatic metabolic pathway of 10-formyltetrahydrofolate (10-HCO-H4folate) formed from tetrahydrofolate (H4folate) and formate, 2-carbon of glycine (Gly), the 3-carbon of serine (Ser) and imidazole-ring 2-carbon of histidine (His). Key enzymes are: 1) 10-HCO-H4folate synthetase; 2) glycine-cleavage system (GCS); and 3) serine hydroxymethyltransferase (SHMT). 10-HCO-H4folate is incorporated into carbon 2 (C2) and 8 (C8) of the purine ring catalyzed by AICAR and GAR transformylases, respectively.
formate generated by folate-dependent metabolism in mitochondria is a major source of carbon for PNB (called here mitochondrialformate hypothesis) (Christensen and MacKenzie, 2006; Pasternak et al., 1994). Materials and methods Protocol The study was approved by the University of Alabama at Birmingham, Institutional Review Board for Human Use. Six adult males (five Caucasian and one Hispanic) with no history of serious diseases participated in this study. First, without any dosing of 13C-compound, they collected urine at each void for 24 h, measured the volume and saved each aliquot to establish baseline % 13C at the C2 and C8 positions of urinary uric acid. This procedure was repeated after an oral dose of 1.0 g of [13C] sodium formate (99% 13C purity, Cambridge Isotope Lab., Andover, MA) in 100 ml of water. The oral dose was ingested from 12:00 to 17:00 (24-h clock). One subject (B.C.) collected his urine for 48 h after the dose. LC/MS/MS assay An LC/MS/MS method was used to measure the 13C-enrichment at C2 and C8 and % 13C was calculated as previously described (Baggott et al., 2007a,b; Gorman et al., 2003). The 13C-enrichment at C2 and C8 was measured by subtracting the % 13C in time-of-day-matched baseline voids from % 13C after dosing. The amount of uric acid in each void is proportional to the area-under-the-curve of the extracted ion transition m/z 167 → 124, which is uric acid containing only 12C, 1H, 14N and 16O (Baggott et al., 2007a,b; Gorman et al., 2003). These values and urine volumes were used to determine the relative amounts of uric acid excreted in each void and in 24 h. Statistics The paired t-test was used to detect % 13C-enrichment greater than zero in all subjects with the exception of % 13C-enrichment in C2 for subject D.M. whose data was not normally distributed. The Wilcoxon paired-sample test was, therefore, used in this particular instance.
Only C2/C8 ratios with both enrichments positive were used to determine the median. The 24-h data from three subjects in the previous report (Baggott et al., 2007b) were combined with data from four subjects in the current study (D.M., J.M., C.B. and M.D.) in order to evaluate diurnal variation in the utilization of [13C] formate for C2 during PNB. Mean % 13 C-enrichments in up to seven voids (from seven subjects) in 4h time periods in a 24-h day were calculated. When a subject had two voids in one 4-h time period, values of % 13C-enrichment were averaged. Peak % 13C-enrichments at C2 were compared to non-peak ones using the t-test.
Results Enrichment of C2 and C8 by [13C] formate In voids after the [13C] formate dose, mean % 13C-enrichment at the C2 position varied from −0.17 to 2.00%, whereas that at C8 varied from −0.06 to 0.15% (Table 1). Each void contained N5% of the total 24-h urinary uric acid excretion. The theoretical lower limit for 13Cenrichment is zero: however, negative enrichments can be obtained because of instrumental error and errors inherent in small differences between large numbers and the fact that the natural abundance of 13C in food (baseline value) varies with its source (Baggott et al., 2007b). The standard deviation of the % 13C enrichment in control experiments (i.e. two baseline experiments subtracted from each other) has been reported as ±0.11% and ±0.08% for C2 and C8, respectively (Baggott et al., 2007a). Therefore, some statistically significant mean % 13 C enrichments at C8 in Table 1 may be doubtful. Median C2/C8 ratios were always greater than 1.0 even in subject B.C. who incorporated insignificant amounts of 13C into C2 (Table 1). Median C2/C8 ratios varied from 1.6 to 19 among subjects. The % 13C-enrichments did not correlate with the body weights of the subjects or the relative amounts of uric acid excreted in 24 h (Table 1). Fig. 2 shows variation in % 13C-enrichments at C2 and C8 of uric acid in each void after the dose. In general, in subjects with substantial % 13 C-enrichment at C2, the peak values occurred in voids after the first to the third voids. The % enrichment at C2 was not correlated with the amount of uric acid in the void and each void contained N5% of the total 24-h urinary uric acid excretion. The highest % 13C-enrichments at C8 occurred in subject M.D.; however, these did not parallel the
690
J.E. Baggott et al. / Life Sciences 88 (2011) 688–692
Table 1 13 C-Enrichment at the C2 and C8 positions of urinary uric acid after an oral dose of [13C] formate in adult males. Subject
D.M. J.M. C.B. B.C. J.S. M.D. a b c d e
Body weight, height, BMI (kg/cm; kg/m2)
Number of voids
85/189; 96/170; 98/195; 65/175; 135/185; 84/188;
9 6 7 9 9 10
23.8 33.2 25.8 21.2 39.4 23.8
Mean % enrichment ± SD C2
C8
0.72 ± 2.23a 2.0 ± 1.5c 0.79 ± 0.97c 0.02 ± 0.05 − 0.17 ± 0.09 0.44 ± 0.41c
0.00 ± 0.05 − 0.06 ± 0.02 0.07 ± 0.06c 0.05 ± 0.06c − 0.01 ± 0.05 0.15 ± 0.17c
Median C2/C8 ratio (range)
Relative amount of urinary uric acid/24 h
19b N/Ad 12 (2.4–21) 1.6 (0.1–12) N/Ad 5.1(1.1–52)
2.3 1.1 1.1 1.3, 1.3e 2.5 1.0
Significantly greater than zero by Wilcoxon paired-sample test (P b 0.05). Only one value could be calculated. Significantly greater than zero by paired t-test (P b 0.05). Enrichments at C2 or C8 were all negative numbers. B.C. collected urine for 48 h.
highest or lowest % 13C-enrichment at C2 in this subject. In general, enrichment at C2 was not correlated with that at C8 in our subjects. Phenotypes Although the data is from only 6 subjects, there are apparently patterns among these 6 subjects (Table 1, Fig. 2). In one pattern, 92– 100% of the total 13C formate incorporated was incorporated into the C2 position (subjects D.M., J.M. and C.B.). In another pattern, almost no [13C] formate was incorporated into the C2 and C8 positions (subjects B.C. and J.S.). In these latter two subjects, low % enrichment at C2 apparently did not increase % enrichment at C8. A possible third phenotype (subject M.D.) incorporated some [13C] formate into C8; however C2 incorporation was ~ 3-fold greater. Circadian rhythm The mean % 13C-enrichments at C2 were plotted against six time periods (Fig. 3). The highest mean occurred in the 8–12 h period. The mean (±SEM) from combined data of two 4-h periods (8 to 16 h, i.e., peak values) was 1.53 (±0.45)% and was ~ 3-fold higher than the mean of combined data from the 16 to 8 h periods of 0.54 (±0.11)%, and the difference was significant (p b 0.025). The data in Fig. 3 probably do not indicate simple clearance of the formate dose through uric acid since circadian rhythms with 3 peaks were previously observed over 3 days after the dose (Baggott et al., 2007b).
Discussion The variations in total incorporation of formate into purines are similar to previous reports. Data from Buchanan and Rollins (1962) indicate a 6.6-fold difference in the specific activity in urinary uric acid after an oral [14C] formate dose in patients with or without gout. Unexpectedly the specific activity in patients with gout was relatively low. Similar to our results, the extent of uric acid labeling did not correlate with the amount excreted. Stahelin et al. (1970) observed a 6.4-fold difference in % of [14C] formate excreted in 11 days as uric acid. Unexpectedly the folate deficient subjects had greater [14C] formate incorporation. Baggott et al. (2007b) reported that adult males incorporated more 13 C into the C2 than C8 position of urinary uric acid after a [13C] formate dose. Previous subjects A and B are similar to subjects D.M., J.M. and C.B. while previous subject C is similar to subject M.D. If PNB in mammalian liver stops at the AICAR step, this could explain why C2 is preferentially labeled in some subjects. It is likely that rabbit liver supplies an advanced intermediate in PNB to bone marrow (and other non-hepatic sites) (Lajtha and Vane, 1958). This could explain why mammalian liver does not metabolize AICAR to IMP. Several lines of evidence support this conclusion. Rat hepatocytes incubated with aminoimidazolecarboxamide-riboside (AICA-riboside) accumulate AICAR; which is not metabolized to purine nucleotides (Vincent et al., 1991; Corton et al., 1995). Therefore, PNB is not completed. Nicotinamide treatment of mice increases liver incorporation of [14C] formate and 2-
Fig. 2. The % enrichment of C2 (open bars) and C8 (closed bars) of uric acid in each void after [13C] formate dosing in six subjects. The X axis increases with time after dosing but time intervals between voids are different. Urine collection was for 24 h except for subject B.C. who collected for 48 h. Numbers next to bars are % enrichments that were out of range.
J.E. Baggott et al. / Life Sciences 88 (2011) 688–692
Fig. 3. The mean % enrichment of C2 (SEM bars) of urinary uric acid from [13C] formate are plotted in six time periods in the day for subjects (D.M., J.M., C.B. and M.D.) in the present study and combined with subjects (A, B and C) in a previous study (Baggott et al., 2007b). When more than one void in a time period occurred in a subject, the average was used. The number of voids plus averaged voids was 3, 6, 6, 7, 6 and 7 (from left to right).
[14C] glycine into adenosine nucleotides however increased PNB induced by nicotinamide does not occur in liver slices, again PNB is not completed (Shuster et al., 1958). Rat liver slices do not biosynthesize labeled uric acid or allantoin from U-[14C] serine although other labeled metabolites are detected (Matsuda et al., 1973). Mouse liver slice has a low capacity to biosynthesize purines from 14C formate compared to other organs (Allsop and Watts, 1990). Flux rates of purine metabolism in isolated rat hepatocytes using purine tracers indicated low PNB activity (Schwendel et al., 1997). The above evidence indicates that PNB in mammalian liver stops at AICAR or at an earlier step. Since substrate cycling of AICAR to AICA-riboside (which crosses cell membranes) occurs in isolated rat hepatocytes (Vincent et al., 1996), and AICA, an AICAR metabolite, is normally found in human urine (McGreer et al., 1961), PNB in human liver probably stops at AICAR; therefore, its riboside is in circulation. AICAR to IMP metabolism should occur in rat liver since the specific activity of AICAR transformylase is greater than that of GAR transformylase (Deacon et al., 1985). However, AICAR transformylase is likely to be profoundly inhibited by adenosine nucleotides (Wall et al., 2000) that are many orders of magnitude greater in concentration than AICAR inside these cells (Schwendel et al., 1997). Hepatic AICAR, is therefore exported to the blood as AICA-riboside. A non-hepatic site could metabolize AICAR to IMP. One site for AICAR to IMP metabolism using formate would be erythrocytes (and bone marrow precursors). Erythrocytes contain high specific activities of 10-formyltetrahydrofolate (10-HCO-H4folate) synthetase (Reaction 3, Fig. 1) and AICAR transformylase and a low specific activity of serine hydroxymethyl transferase (SHMT) (Reaction 2, Fig. 1) (Bertino et al., 1962; Wagner and Levitch, 1973). Thus, erythrocytes prefer to utilize formate. Erythrocytes metabolize AICAriboside to IMP in the presence of formate in vitro (Wagner and Levitch, 1973). In patients with purine-nucleotide-overproduction syndromes, AICAR accumulates in erythrocytes (Sidi and Mitchell, 1985). AICAR is probably furnished to erythrocytes and bone marrow at rates exceeding their capacity to metabolize it and AICAR accumulation indicates a normal metabolic process that is overwhelmed. An extreme case of this is found in a patient with no activity of erythrocyte AICAR transformylase who accumulates large concentrations of AICAR in erythrocytes (Marie et al., 2004). Completion of PNB from AICAR could occur in erythrocytes and bone marrow with formate utilization and this would explain substantial 13C enrichment at C2 in some subjects. If carbon 3 of serine primarily supplies a one carbon unit to GAR transformylase this would explain why C8 is generally not enriched by
691
13 C formate. Benkovic and colleagues (Caperelli et al., 1980; Smith et al., 1980) have shown that GAR transformylase purifies as a complex with SHMT (Reaction 3, Fig. 1) and the trifunctional folate metabolizing enzyme (TFM) in chicken liver and AICAR transformylase activity is low in the final purification. This enzyme complex could furnish one-carbon units from 3-serine to GAR transformylase (Fig. 1) and dilute [13C] formate. A GAR transformylase–SHMT–TFM complex may also exist in human liver. An oral dose of 2-[13C] glycine enriches C5 plus C8 but not C2 of uric acid in humans (Baggott et al., 2007b). The metabolism of 2-glycine to 10-HCO-H4folate requires the glycine-cleavage system (GCS) that is primarily found in the mitochondria of liver (Reaction 2, Fig. 1) (Kikuchi, 1973). Both GCS and SHMT are found in liver mitochondria which, due to carbon unit exchange, results in the formation of 3-[13C] serine from one 2-[13C] glycine and one unlabeled glycine (Kikuchi, 1973). The above exchange reaction occurs in vivo in rat liver (Fern and Garlick, 1974). Since 3-[13C] serine can cross mitochondrial membranes (Christensen and MacKenzie, 2006), it could furnish its labeled carbon to a cytoplasmic liver GAR transformylase–SHMT–TFM complex that would enrich C8. If [13C] formate is formed in the liver from 2-[13C] glycine or 3-[13C] serine it is formed in small quantities. If this were not true, substantial C2 enrichment by 2-[13C] glycine dosing should have occurred because the in vivo [13C] formate pool was increased. The above evidence suggests that human liver prefers to utilize the 3-serine for GAR transformylase which is consistent with the existence of a GAR transformylase–SHMT–TFM complex. One subject (M.D.) and one previous subject did enrich C8 following a [13C] formate dose (Baggott et al., 2007b). It is possible that their hepatic GAR transformylase–SHMT–TFM complex was not presented with enough serine to dilute [13C] formate. Some subjects may have reduced capacity in erythrocytes to incorporate formate into C2 of purine and this could explain the low C2 labeling in some subjects. Wagner and Levitch (1973) observed a 100fold variation in the capacity of human erythrocytes to metabolize AICA-riboside and formate to IMP in just 16 subjects. They attributed this to variation to erythrocyte tetrahydrofolate (H4folate) levels (Fig. 1). Obviously individuals with a low erythrocyte metabolism capacity still have an active PNB. An extreme example of this is an individual with no erythrocyte AICAR transformylase, but with normal uric acid production (Marie et al., 2004). In this individual, a site other than erythrocytes must complete PNB utilizing liver derived AICA-riboside. Subjects B.C. and J.S. are completing PNB in another non-hepatic tissue. One candidate would be cardiac and skeletal muscles. Dog muscles metabolize infused AICA-riboside to purine nucleotides (Sabina et al., 1982). However, when formate was added to the infusate, there was no increase in purine-nucleotides indicating that formate is not utilized. The low activity of 10-HCO-H4folate synthetase (Reaction 1, Fig. 1) in these muscles would preclude formate utilization (Cheek and Appling, 1989; Whiteley, 1960). Serine, a formate alternative, is a possibility since cardiac and skeletal muscles contain relatively high SHMT activities (Reaction 3, Fig. 1) (Whiteley, 1960). Subjects B.C. and J.S. could be utilizing the 3-serine for both C2 (in muscle) and C8 (in liver). Smaaland et al. (1991) reported that peak DNA synthesis in human bone marrow occurs at the 8–16-h period in healthy male volunteers. This corresponds to peak time of 13C-enrichments at C2 (Fig. 3). AICAR transformylase activity is ~ 7-fold higher than GAR transformylase activity in human bone marrow (Deacon et al., 1985) and is consistent with a circadian rhythm in 13C-enrichments at C2. Mouse bone marrow shows a circadian rhythm in thymidylate synthase with a peak corresponding to peak DNA synthesis (Lincoln et al., 2000). Thus bone marrow may be responsible for the substantial incorporation of [13C] formate into C2. Our results are not consistent with the mitochondrial-formate hypothesis. Many non folate dependent metabolic sources of formate
692
J.E. Baggott et al. / Life Sciences 88 (2011) 688–692
(methylthioadenosine, tryptophan, etc.) exist; it is unlikely that this is the only important source (Hanzlik et al., 2005). Methylthioadenosine is considered a major formate source in humans by some (Deacon et al., 1990). This hypothesis cannot easily explain subjects B.C. and J.S. who utilize little formate for PNB and other subjects that have low C8 enrichment. Results presented here and previously (Baggott et al., 2007b) are different from results in cultured cells and chicken, where formate and 2-glycine are incorporated equally into C2 and C8 (Jeong and Schirch, 1996; Kastanos et al., 1997; Fu et al., 2001; Marsh, 1951). In these experiments, there are no organ-to-organ interactions or in the case of chickens, they are a uricotelic animal. In rodents, the C2/C8 ratios from labeled formate are reported to be 1.0 (Drysdale et al., 1951), and 1.7 to 2.0 (Shuster and Goldin, 1959). SHMT activity (Reaction 3, Fig. 1) is higher in monkey liver compared to rat liver (Block et al., 1985). Primate liver may use 3-serine for GAR transformylase (C8) to a greater extent and could explain higher C2/C8 ratios. Conclusion There are possibly 3 phenotypes in PNB from formate in humans. The genotypes producing these PNB phenotypes may be complex. Our data could be explained if in liver PNB stops at AICAR formation and AICAR to IMP metabolism is completed at extra hepatic sites, which may be erythrocytes and bone marrow. The circadian rhythm of C2 enrichment from formate may suggest a genotype that utilizes bone marrow. The PNB pathway is essential since there are only human genotypes which overproduce purines indicating that underproduction genotypes are lethal. PNB may be a robust process allowing different genotype portfolios to produce adequate purines. Phenotypes in PNB from formate may be helpful in predicting who will respond favorably to chemotherapeutic and immunosuppressant drugs which interfere with PNB. Conflict of interest statement The authors have no conflicts of interest to report.
References Abrams R. Stability of the adenine ring structure in the rat. Biochim Biophys Acta 1956;21:439–40. Allsop J, Watts RWE. Purine synthesis de novo and salvage in hypoxanthine phosphorisyltransferase-deficient mice. Enzyme 1990;43:155–9. Baggott JE, Gorman GS, Morgan SL, Tamura T. 13C-enrichment at carbons 8 and 2 of uric acid after 13C-labeled folate dose in man. Biochem Biophys Res Commun 2007a;361:307–10. Baggott JE, Gorman GS, Tamura T. 13C Enrichment of carbons 2 and 8 of purine by folatedependent reactions after [13C] formate and [2-13C] glycine dosing in adult humans. Metabolism 2007b;56:708–15. Bennett EL, Karlsson H. Metabolism of adenine in mice. J Biol Chem 1957;229:39–50. Bertino JR, Simmons B, Donohue DM. Purification and properties of the formateactivating enzyme from erythrocyte. J Biol Chem 1962;237:1314–8. Block KA, Eells JT, Noker PE, Hawtrey CA, Tephly TR. Role of hepatic tetrahydrofolate in the species difference in methanol toxicity. Proc Natl Acad Sci USA 1985;82: 3854–8. Buchanan DL, Rollins JM. Lack of correlation between gout and the incorporation of isotopic formate into uric acid. Yale J Biol Med 1962;34:31–9. Caperelli CA, Benkovic PA, Chettur G, Benkovic SJ. Purification of a complex catalyzing folate cofactor synthesis and transformylation in de novo purine biosynthesis. J Biol Chem 1980;255:1885–90. Cheek WD, Appling DR. Purification, immunoassay and tissue distribution of rat C1tetrahydrofolate synthase. Arch Biochem Biophys 1989;270:504–12. Christensen KE, Mackenzie RE. Mitochondrial one-carbon metabolism is adapted to the specific needs of yeast, plants, and mammals. BioEssays 2006;28:595–605. Corton JM, Gillespie JG, Hawley SA, Hardie DG. 5-Aminoimidazole-4-carboxamide ribonucleotide. A specific method for activating AMP-activated protein kinase in intact cells? Eur J Biochem 1995;229:558–65.
Deacon R, Chanarin I, Lumb M, Perry J. Role of folate dependent transformylases in synthesis of purine in bone marrow of man and in bone marrow and liver of rats. J Clin Pathol 1985;38:1349–52. Deacon R, Bottiglieri T, Chanarin I, Lumb M, Perry J. Methylithioadenosine serves as a single carbon source to the folate co-enzyme pool in rat bone marrow cells. Biochim Biophys Acta 1990;1034:342–6. Drysdale GR, Plaut GWE, Lardy HA. The relationship of folic acid to formate metabolism in the rat: formate incorporation into purines. Journal of Biological Chemistry 1951;193:533–88. Fern EB, Garlick PJ. The specific radioactivity of the tissue free amino acid pool as a basis for measuring the rate of protein synthesis in the rat in vivo. Biochem J 1974;142: 413–9. Fu T-F, Rife JP, Schirch V. The role of serine hydroxymethyltransferase isozymes in onecarbon metabolism in MCF-7 cells as determined by C13 NMR. Arch Biochem Biophys 2001;398:42–50. Garrett RH, Grisham CH. The synthesis and degradation of nucleotides. Biochemistry. 3 rd ed. Belmont, CA: Thomson Brooks/Cole; 2005. chapter 26. Gorman GS, Tamura T, Baggott JE. Mass spectrometric method for detecting carbon 13 enrichment introduced by folate coenzymes in uric acid. Anal Biochem 2003;321: 188–91. Hanzlik RP, Fowler SC, Eells JT. Absorption and elimination of formate following oral administration of calcium formate in female human subjects. Drug Metab Dispos 2005;33:282–6. Jeong S-S, Schirch V. Role of cystolic serine hydroxymethyltransferase in one-carbon metabolism in Neurospora crassa. Arch Biochem Biophys 1996;335:333–41. Kastanos EK, Woldman YY, Appling DR. Role of mitochondrial and cytoplasmic serine hydroxymethyltransferase isozymes in de novo purine synthesis in Saccharomyces cerevisiae. Biochemistry 1997;36:14956–64. Kikuchi G. The glycine cleavage system: composition, reaction mechanism, and physiological significance. Mol Cell Biochem 1973;1:169–87. Lajtha LG, Vane JR. Dependence of bone marrow cells on the liver for purine supply. Nature 1958;182:191–2. Lincoln DW, Hrushesky WJM, Wood PA. Circadian organization of thymidylate synthase activity in normal tissues: a possible basis for 5-fluouracil chronotherapeutic advantage. Int J Cancer 2000;88:479–85. Marie S, Heron B, Bitoun P, Timmerman T, Van den Berghe G, Vincent MF. AICAribosiduria: a novel, neurologically devastating inborn error of purine biosynthesis caused by mutation of ATIC. Am J Hum Genet 2004;74:1276–81. Marsh WH. On the biosynthesis of purine in the bird; role of formate. J Biol Chem 1951;190:633–41. Matsuda Y, Kuroda Y, Kobayashi K, Katunuma N. Comparative studies on glutamine, serine, and glycine metabolism in ureotelic and uriotelic animals. J Biochem (Tokyo) 1973;73:291–8. McGreer PL, McGreer EG, Griffin MC. Excretion of 4-amino-5-imidazolecarboxamide in human urine. Can J Biochem Physiol 1961;39:591–603. Pasternak LB, Laude Jr DD, Appling DR. 13C NMR analysis of intercompartmental flow of one-carbon units into choline and purines in Saccharomyces cerevisiae. Biochemistry 1994;33:74–82. Sabina RL, Kernstine KH, Boyd RL, Holmes EW, Swain JL. Metabolism of 5-amino-4imidazolecarboxamide riboside in cardiac and skeletal muscle. J Biol Chem 1982;257:10178–83. Schwendel A, Grune T, Holzhutter HG, Siems WG. Models for the regulation of purine metabolism in rat hepatocytes: evaluation of tracer kinetic experiments. Am J Physiol 1997;273:G239–46. Shuster L, Goldin A. The effect of nicotinamide on incorporation in vivo of formate-14C. J Biol Chem 1959;234:129–33. Shuster L, Langan TA, Kaplan NO, Goldin A. Significance of induced in vivo synthesis of diphosphopyridine nucleotide. Nature 1958;182:512–5. Sidi Y, Mitchell BS. Z-nucleotide accumulation in erythrocytes from Lesch–Nyhan patients. J Clin Investig 1985;76:2416–9. Smaaland R, Laerum OD, Lote K, Sletvold O, Sothern RB, Bjerknes R. DNA synthesis in human bone marrow is circadian stage dependent. Blood 1991;77:2603–11. Smith GK, Mueller WT, Wasserman GF, Taylor WD, Benkovic SJ. Characterization of the enzyme complex involving the folate-requiring enzymes of de novo purine biosynthesis. Biochemistry 1980;19:4313–21. Stahelin HB, Stokstad ELR, Winchell HS. Formate oxidation and its incorporation into uric acid in folic acid deficiency. J Nucl Med 1970;11:247–54. Vincent MF, Marangos PJ, Gruber HE, Van den Berghe G. Inhibition by AICA riboside of gluconeogenesis in isolated rat hepatocytes. Diabetes 1991;40:1259–66. Vincent MF, Bontemps F, Van den Berghe G. Substrate cycling between 5-amino-4imidazolecarboxamide riboside and its monophosphate in isolated rat hepatocytes. Biochem Pharmacol 1996;52:999-1006. Wagner C, Levitch ME. The utilization of formate by human erythrocytes. Biochim Biophys Acta 1973;304:623–33. Wall M, Shim JH, Benkovic SJ. Human AICAR transformylase: role of the 4-carboxamide of AICAR in binding and catalysis. Biochemistry 2000;39:11303–11. Whiteley HR. The distribution of the formate activating enzyme and other enzymes involving tetrahydrofolic acid in animal tissues. Comp Biochem Physiol 1960;1: 227–47.
Contents lists available at ScienceDirect
Life Sciences j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / l i f e s c i e
Phenotypes and circadian rhythm in utilization of formate in purine nucleotide biosynthesis de novo in adult humans Joseph E. Baggott a, Gregory S. Gorman b, Sarah L. Morgan a,c,⁎ a b c
Department of Nutrition Sciences, University of Alabama at Birmingham, Birmingham, AL 35294, USA McWhorter School of Pharmacy, Samford University, Birmingham, AL 35229, USA Department of Medicine, University of Alabama at Birmingham, Birmingham, AL 35294, USA
a r t i c l e
i n f o
Article history: Received 3 May 2010 Accepted 28 January 2011 Keywords: Purine nucleotide biosynthesis Formate Humans Phenotypes Circadian rhythm
a b s t r a c t Aims: Folate coenzymes and dependent enzymes introduce one carbon units at positions 2 (C2) and 8 (C8) of the purine ring during de novo biosynthesis. Formate is one source of one-carbon units. Although much is known about lower organisms, little data exists describing formate utilization for purine biosynthesis in humans. Main methods: Mass-spectrometric analysis of urinary uric acid, the final purine catabolite, following 1.0 g oral doses of sodium [13C] formate was performed and detected 13C enrichment at C2 and C8 separately. Key findings: Three phenotypes were suggested. One incorporates 13C 0.72 to 2.0% into C2 versus only 0 to 0.07% into C8. Another incorporates only 0 to 0.05% 13C into C2 or C8. A third phenotype incorporates 13C into C8 (0.15%) but C2 incorporation (0.44%) is still greater. In subjects who incorporated 13C formate into C2, peak enrichment occurred in voids from 8–12 h (24 h clock) suggesting a circadian rhythm. Significance: Evidence that mammalian liver introduces C8 and that C2 is introduced in a non-hepatic site would explain our results. Our data are not similar to those in non-mammalian organisms or cells in culture and are not consistent with the hypothesis that formate from folate-dependent metabolism in mitochondria is a major one carbon source for purine biosynthesis. Timing of peak 13C enrichment at C2 corresponds to maximal DNA synthesis in human bone marrow. Phenotypes may explain the efficacy (or lack of) of certain anticancer and immunosuppressive drugs. © 2011 Elsevier Inc. All rights reserved.
Introduction During purine nucleotide biosynthesis de novo (PNB), two folatecoenzyme-dependent transformylases, aminoimidazolecarboxamide ribotide (AICAR) and glycinamide ribotide (GAR), introduce a carbon at the formyl-oxidation state into positions 2 (C2) and 8 (C8) of the purine ring, respectively as shown in Fig. 1 (Garrett and Grisham, 2005). Carbon at C8 is incorporated first followed by carbon at C2 in the PNB pathway. Formate, one source of these carbons (Reaction 1, Fig. 1), occurs in human plasma at ≥20 μM concentrations, and oral doses of formate maximally increase plasma formate concentrations in only 30–60 min and rapidly expand the in vivo formate pool (Hanzlik et al., 2005). Peaks in labeled formate incorporation into urinary uric acid, the final product of purine catabolism, appear in 1 to 3 days after dosing of labeled formate in humans (Stahelin et al., 1970; Baggott et al., 2007b) and the label was found almost exclusively in C2 and C8 of purines when labeled formate was given to laboratory ⁎ Corresponding author at: Department of Medicine, University of Alabama at Birmingham, Birmingham, AL 35294, USA. Tel.: + 1 205 934 3235; fax: + 1 205 996 2072. E-mail address: [email protected] (S.L. Morgan). 0024-3205/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.lfs.2011.02.007
animals (Baggott et al., 2007b). Isotope exchange of the C2 and C8 positions of purines does not occur in vivo (Abrams, 1956; Bennett and Karlsson, 1957). Therefore, 13C-labeled formate incorporation into C2 and C8 of uric acid can be used to measure how and the extent to which this compound is used in PNB. In cultured cells, formate is incorporated equally into the C2 and C8 positions (Jeong and Schirch, 1996; Kastanos et al., 1997; Fu et al., 2001). A similar pattern is found in rodents and chickens (Drysdale et al., 1951; Marsh, 1951). In the literature, consistent data on humans is lacking. Therefore, this study was conducted to detect patterns of C2 and C8 incorporation (phenotypes) and a circadian rhythm in the utilization of formate for PNB. Using a liquid-chromatography mass spectrometric method (LC/MS/MS) (Gorman et al., 2003), we measured 13C enrichment of urinary uric acid in six human subjects who were given an oral dose of [13C] sodium formate. We found distinct phenotypes in 13C-enrichment pattern at the C2 and C8 positions of uric acid and a circadian rhythm in 13C-enrichment at C2. These results are discussed with respect to the PNB process in adult humans. Phenotypes in formate utilization in PNB may be useful in explaining efficacy (or lack of) of cancer chemotherapeutic and immunosuppressive drugs (e.g. methotrexate) that interfere with PNB. We also discuss our data with respect to the hypothesis: that
J.E. Baggott et al. / Life Sciences 88 (2011) 688–692
689
Fig. 1. Diagrammatic metabolic pathway of 10-formyltetrahydrofolate (10-HCO-H4folate) formed from tetrahydrofolate (H4folate) and formate, 2-carbon of glycine (Gly), the 3-carbon of serine (Ser) and imidazole-ring 2-carbon of histidine (His). Key enzymes are: 1) 10-HCO-H4folate synthetase; 2) glycine-cleavage system (GCS); and 3) serine hydroxymethyltransferase (SHMT). 10-HCO-H4folate is incorporated into carbon 2 (C2) and 8 (C8) of the purine ring catalyzed by AICAR and GAR transformylases, respectively.
formate generated by folate-dependent metabolism in mitochondria is a major source of carbon for PNB (called here mitochondrialformate hypothesis) (Christensen and MacKenzie, 2006; Pasternak et al., 1994). Materials and methods Protocol The study was approved by the University of Alabama at Birmingham, Institutional Review Board for Human Use. Six adult males (five Caucasian and one Hispanic) with no history of serious diseases participated in this study. First, without any dosing of 13C-compound, they collected urine at each void for 24 h, measured the volume and saved each aliquot to establish baseline % 13C at the C2 and C8 positions of urinary uric acid. This procedure was repeated after an oral dose of 1.0 g of [13C] sodium formate (99% 13C purity, Cambridge Isotope Lab., Andover, MA) in 100 ml of water. The oral dose was ingested from 12:00 to 17:00 (24-h clock). One subject (B.C.) collected his urine for 48 h after the dose. LC/MS/MS assay An LC/MS/MS method was used to measure the 13C-enrichment at C2 and C8 and % 13C was calculated as previously described (Baggott et al., 2007a,b; Gorman et al., 2003). The 13C-enrichment at C2 and C8 was measured by subtracting the % 13C in time-of-day-matched baseline voids from % 13C after dosing. The amount of uric acid in each void is proportional to the area-under-the-curve of the extracted ion transition m/z 167 → 124, which is uric acid containing only 12C, 1H, 14N and 16O (Baggott et al., 2007a,b; Gorman et al., 2003). These values and urine volumes were used to determine the relative amounts of uric acid excreted in each void and in 24 h. Statistics The paired t-test was used to detect % 13C-enrichment greater than zero in all subjects with the exception of % 13C-enrichment in C2 for subject D.M. whose data was not normally distributed. The Wilcoxon paired-sample test was, therefore, used in this particular instance.
Only C2/C8 ratios with both enrichments positive were used to determine the median. The 24-h data from three subjects in the previous report (Baggott et al., 2007b) were combined with data from four subjects in the current study (D.M., J.M., C.B. and M.D.) in order to evaluate diurnal variation in the utilization of [13C] formate for C2 during PNB. Mean % 13 C-enrichments in up to seven voids (from seven subjects) in 4h time periods in a 24-h day were calculated. When a subject had two voids in one 4-h time period, values of % 13C-enrichment were averaged. Peak % 13C-enrichments at C2 were compared to non-peak ones using the t-test.
Results Enrichment of C2 and C8 by [13C] formate In voids after the [13C] formate dose, mean % 13C-enrichment at the C2 position varied from −0.17 to 2.00%, whereas that at C8 varied from −0.06 to 0.15% (Table 1). Each void contained N5% of the total 24-h urinary uric acid excretion. The theoretical lower limit for 13Cenrichment is zero: however, negative enrichments can be obtained because of instrumental error and errors inherent in small differences between large numbers and the fact that the natural abundance of 13C in food (baseline value) varies with its source (Baggott et al., 2007b). The standard deviation of the % 13C enrichment in control experiments (i.e. two baseline experiments subtracted from each other) has been reported as ±0.11% and ±0.08% for C2 and C8, respectively (Baggott et al., 2007a). Therefore, some statistically significant mean % 13 C enrichments at C8 in Table 1 may be doubtful. Median C2/C8 ratios were always greater than 1.0 even in subject B.C. who incorporated insignificant amounts of 13C into C2 (Table 1). Median C2/C8 ratios varied from 1.6 to 19 among subjects. The % 13C-enrichments did not correlate with the body weights of the subjects or the relative amounts of uric acid excreted in 24 h (Table 1). Fig. 2 shows variation in % 13C-enrichments at C2 and C8 of uric acid in each void after the dose. In general, in subjects with substantial % 13 C-enrichment at C2, the peak values occurred in voids after the first to the third voids. The % enrichment at C2 was not correlated with the amount of uric acid in the void and each void contained N5% of the total 24-h urinary uric acid excretion. The highest % 13C-enrichments at C8 occurred in subject M.D.; however, these did not parallel the
690
J.E. Baggott et al. / Life Sciences 88 (2011) 688–692
Table 1 13 C-Enrichment at the C2 and C8 positions of urinary uric acid after an oral dose of [13C] formate in adult males. Subject
D.M. J.M. C.B. B.C. J.S. M.D. a b c d e
Body weight, height, BMI (kg/cm; kg/m2)
Number of voids
85/189; 96/170; 98/195; 65/175; 135/185; 84/188;
9 6 7 9 9 10
23.8 33.2 25.8 21.2 39.4 23.8
Mean % enrichment ± SD C2
C8
0.72 ± 2.23a 2.0 ± 1.5c 0.79 ± 0.97c 0.02 ± 0.05 − 0.17 ± 0.09 0.44 ± 0.41c
0.00 ± 0.05 − 0.06 ± 0.02 0.07 ± 0.06c 0.05 ± 0.06c − 0.01 ± 0.05 0.15 ± 0.17c
Median C2/C8 ratio (range)
Relative amount of urinary uric acid/24 h
19b N/Ad 12 (2.4–21) 1.6 (0.1–12) N/Ad 5.1(1.1–52)
2.3 1.1 1.1 1.3, 1.3e 2.5 1.0
Significantly greater than zero by Wilcoxon paired-sample test (P b 0.05). Only one value could be calculated. Significantly greater than zero by paired t-test (P b 0.05). Enrichments at C2 or C8 were all negative numbers. B.C. collected urine for 48 h.
highest or lowest % 13C-enrichment at C2 in this subject. In general, enrichment at C2 was not correlated with that at C8 in our subjects. Phenotypes Although the data is from only 6 subjects, there are apparently patterns among these 6 subjects (Table 1, Fig. 2). In one pattern, 92– 100% of the total 13C formate incorporated was incorporated into the C2 position (subjects D.M., J.M. and C.B.). In another pattern, almost no [13C] formate was incorporated into the C2 and C8 positions (subjects B.C. and J.S.). In these latter two subjects, low % enrichment at C2 apparently did not increase % enrichment at C8. A possible third phenotype (subject M.D.) incorporated some [13C] formate into C8; however C2 incorporation was ~ 3-fold greater. Circadian rhythm The mean % 13C-enrichments at C2 were plotted against six time periods (Fig. 3). The highest mean occurred in the 8–12 h period. The mean (±SEM) from combined data of two 4-h periods (8 to 16 h, i.e., peak values) was 1.53 (±0.45)% and was ~ 3-fold higher than the mean of combined data from the 16 to 8 h periods of 0.54 (±0.11)%, and the difference was significant (p b 0.025). The data in Fig. 3 probably do not indicate simple clearance of the formate dose through uric acid since circadian rhythms with 3 peaks were previously observed over 3 days after the dose (Baggott et al., 2007b).
Discussion The variations in total incorporation of formate into purines are similar to previous reports. Data from Buchanan and Rollins (1962) indicate a 6.6-fold difference in the specific activity in urinary uric acid after an oral [14C] formate dose in patients with or without gout. Unexpectedly the specific activity in patients with gout was relatively low. Similar to our results, the extent of uric acid labeling did not correlate with the amount excreted. Stahelin et al. (1970) observed a 6.4-fold difference in % of [14C] formate excreted in 11 days as uric acid. Unexpectedly the folate deficient subjects had greater [14C] formate incorporation. Baggott et al. (2007b) reported that adult males incorporated more 13 C into the C2 than C8 position of urinary uric acid after a [13C] formate dose. Previous subjects A and B are similar to subjects D.M., J.M. and C.B. while previous subject C is similar to subject M.D. If PNB in mammalian liver stops at the AICAR step, this could explain why C2 is preferentially labeled in some subjects. It is likely that rabbit liver supplies an advanced intermediate in PNB to bone marrow (and other non-hepatic sites) (Lajtha and Vane, 1958). This could explain why mammalian liver does not metabolize AICAR to IMP. Several lines of evidence support this conclusion. Rat hepatocytes incubated with aminoimidazolecarboxamide-riboside (AICA-riboside) accumulate AICAR; which is not metabolized to purine nucleotides (Vincent et al., 1991; Corton et al., 1995). Therefore, PNB is not completed. Nicotinamide treatment of mice increases liver incorporation of [14C] formate and 2-
Fig. 2. The % enrichment of C2 (open bars) and C8 (closed bars) of uric acid in each void after [13C] formate dosing in six subjects. The X axis increases with time after dosing but time intervals between voids are different. Urine collection was for 24 h except for subject B.C. who collected for 48 h. Numbers next to bars are % enrichments that were out of range.
J.E. Baggott et al. / Life Sciences 88 (2011) 688–692
Fig. 3. The mean % enrichment of C2 (SEM bars) of urinary uric acid from [13C] formate are plotted in six time periods in the day for subjects (D.M., J.M., C.B. and M.D.) in the present study and combined with subjects (A, B and C) in a previous study (Baggott et al., 2007b). When more than one void in a time period occurred in a subject, the average was used. The number of voids plus averaged voids was 3, 6, 6, 7, 6 and 7 (from left to right).
[14C] glycine into adenosine nucleotides however increased PNB induced by nicotinamide does not occur in liver slices, again PNB is not completed (Shuster et al., 1958). Rat liver slices do not biosynthesize labeled uric acid or allantoin from U-[14C] serine although other labeled metabolites are detected (Matsuda et al., 1973). Mouse liver slice has a low capacity to biosynthesize purines from 14C formate compared to other organs (Allsop and Watts, 1990). Flux rates of purine metabolism in isolated rat hepatocytes using purine tracers indicated low PNB activity (Schwendel et al., 1997). The above evidence indicates that PNB in mammalian liver stops at AICAR or at an earlier step. Since substrate cycling of AICAR to AICA-riboside (which crosses cell membranes) occurs in isolated rat hepatocytes (Vincent et al., 1996), and AICA, an AICAR metabolite, is normally found in human urine (McGreer et al., 1961), PNB in human liver probably stops at AICAR; therefore, its riboside is in circulation. AICAR to IMP metabolism should occur in rat liver since the specific activity of AICAR transformylase is greater than that of GAR transformylase (Deacon et al., 1985). However, AICAR transformylase is likely to be profoundly inhibited by adenosine nucleotides (Wall et al., 2000) that are many orders of magnitude greater in concentration than AICAR inside these cells (Schwendel et al., 1997). Hepatic AICAR, is therefore exported to the blood as AICA-riboside. A non-hepatic site could metabolize AICAR to IMP. One site for AICAR to IMP metabolism using formate would be erythrocytes (and bone marrow precursors). Erythrocytes contain high specific activities of 10-formyltetrahydrofolate (10-HCO-H4folate) synthetase (Reaction 3, Fig. 1) and AICAR transformylase and a low specific activity of serine hydroxymethyl transferase (SHMT) (Reaction 2, Fig. 1) (Bertino et al., 1962; Wagner and Levitch, 1973). Thus, erythrocytes prefer to utilize formate. Erythrocytes metabolize AICAriboside to IMP in the presence of formate in vitro (Wagner and Levitch, 1973). In patients with purine-nucleotide-overproduction syndromes, AICAR accumulates in erythrocytes (Sidi and Mitchell, 1985). AICAR is probably furnished to erythrocytes and bone marrow at rates exceeding their capacity to metabolize it and AICAR accumulation indicates a normal metabolic process that is overwhelmed. An extreme case of this is found in a patient with no activity of erythrocyte AICAR transformylase who accumulates large concentrations of AICAR in erythrocytes (Marie et al., 2004). Completion of PNB from AICAR could occur in erythrocytes and bone marrow with formate utilization and this would explain substantial 13C enrichment at C2 in some subjects. If carbon 3 of serine primarily supplies a one carbon unit to GAR transformylase this would explain why C8 is generally not enriched by
691
13 C formate. Benkovic and colleagues (Caperelli et al., 1980; Smith et al., 1980) have shown that GAR transformylase purifies as a complex with SHMT (Reaction 3, Fig. 1) and the trifunctional folate metabolizing enzyme (TFM) in chicken liver and AICAR transformylase activity is low in the final purification. This enzyme complex could furnish one-carbon units from 3-serine to GAR transformylase (Fig. 1) and dilute [13C] formate. A GAR transformylase–SHMT–TFM complex may also exist in human liver. An oral dose of 2-[13C] glycine enriches C5 plus C8 but not C2 of uric acid in humans (Baggott et al., 2007b). The metabolism of 2-glycine to 10-HCO-H4folate requires the glycine-cleavage system (GCS) that is primarily found in the mitochondria of liver (Reaction 2, Fig. 1) (Kikuchi, 1973). Both GCS and SHMT are found in liver mitochondria which, due to carbon unit exchange, results in the formation of 3-[13C] serine from one 2-[13C] glycine and one unlabeled glycine (Kikuchi, 1973). The above exchange reaction occurs in vivo in rat liver (Fern and Garlick, 1974). Since 3-[13C] serine can cross mitochondrial membranes (Christensen and MacKenzie, 2006), it could furnish its labeled carbon to a cytoplasmic liver GAR transformylase–SHMT–TFM complex that would enrich C8. If [13C] formate is formed in the liver from 2-[13C] glycine or 3-[13C] serine it is formed in small quantities. If this were not true, substantial C2 enrichment by 2-[13C] glycine dosing should have occurred because the in vivo [13C] formate pool was increased. The above evidence suggests that human liver prefers to utilize the 3-serine for GAR transformylase which is consistent with the existence of a GAR transformylase–SHMT–TFM complex. One subject (M.D.) and one previous subject did enrich C8 following a [13C] formate dose (Baggott et al., 2007b). It is possible that their hepatic GAR transformylase–SHMT–TFM complex was not presented with enough serine to dilute [13C] formate. Some subjects may have reduced capacity in erythrocytes to incorporate formate into C2 of purine and this could explain the low C2 labeling in some subjects. Wagner and Levitch (1973) observed a 100fold variation in the capacity of human erythrocytes to metabolize AICA-riboside and formate to IMP in just 16 subjects. They attributed this to variation to erythrocyte tetrahydrofolate (H4folate) levels (Fig. 1). Obviously individuals with a low erythrocyte metabolism capacity still have an active PNB. An extreme example of this is an individual with no erythrocyte AICAR transformylase, but with normal uric acid production (Marie et al., 2004). In this individual, a site other than erythrocytes must complete PNB utilizing liver derived AICA-riboside. Subjects B.C. and J.S. are completing PNB in another non-hepatic tissue. One candidate would be cardiac and skeletal muscles. Dog muscles metabolize infused AICA-riboside to purine nucleotides (Sabina et al., 1982). However, when formate was added to the infusate, there was no increase in purine-nucleotides indicating that formate is not utilized. The low activity of 10-HCO-H4folate synthetase (Reaction 1, Fig. 1) in these muscles would preclude formate utilization (Cheek and Appling, 1989; Whiteley, 1960). Serine, a formate alternative, is a possibility since cardiac and skeletal muscles contain relatively high SHMT activities (Reaction 3, Fig. 1) (Whiteley, 1960). Subjects B.C. and J.S. could be utilizing the 3-serine for both C2 (in muscle) and C8 (in liver). Smaaland et al. (1991) reported that peak DNA synthesis in human bone marrow occurs at the 8–16-h period in healthy male volunteers. This corresponds to peak time of 13C-enrichments at C2 (Fig. 3). AICAR transformylase activity is ~ 7-fold higher than GAR transformylase activity in human bone marrow (Deacon et al., 1985) and is consistent with a circadian rhythm in 13C-enrichments at C2. Mouse bone marrow shows a circadian rhythm in thymidylate synthase with a peak corresponding to peak DNA synthesis (Lincoln et al., 2000). Thus bone marrow may be responsible for the substantial incorporation of [13C] formate into C2. Our results are not consistent with the mitochondrial-formate hypothesis. Many non folate dependent metabolic sources of formate
692
J.E. Baggott et al. / Life Sciences 88 (2011) 688–692
(methylthioadenosine, tryptophan, etc.) exist; it is unlikely that this is the only important source (Hanzlik et al., 2005). Methylthioadenosine is considered a major formate source in humans by some (Deacon et al., 1990). This hypothesis cannot easily explain subjects B.C. and J.S. who utilize little formate for PNB and other subjects that have low C8 enrichment. Results presented here and previously (Baggott et al., 2007b) are different from results in cultured cells and chicken, where formate and 2-glycine are incorporated equally into C2 and C8 (Jeong and Schirch, 1996; Kastanos et al., 1997; Fu et al., 2001; Marsh, 1951). In these experiments, there are no organ-to-organ interactions or in the case of chickens, they are a uricotelic animal. In rodents, the C2/C8 ratios from labeled formate are reported to be 1.0 (Drysdale et al., 1951), and 1.7 to 2.0 (Shuster and Goldin, 1959). SHMT activity (Reaction 3, Fig. 1) is higher in monkey liver compared to rat liver (Block et al., 1985). Primate liver may use 3-serine for GAR transformylase (C8) to a greater extent and could explain higher C2/C8 ratios. Conclusion There are possibly 3 phenotypes in PNB from formate in humans. The genotypes producing these PNB phenotypes may be complex. Our data could be explained if in liver PNB stops at AICAR formation and AICAR to IMP metabolism is completed at extra hepatic sites, which may be erythrocytes and bone marrow. The circadian rhythm of C2 enrichment from formate may suggest a genotype that utilizes bone marrow. The PNB pathway is essential since there are only human genotypes which overproduce purines indicating that underproduction genotypes are lethal. PNB may be a robust process allowing different genotype portfolios to produce adequate purines. Phenotypes in PNB from formate may be helpful in predicting who will respond favorably to chemotherapeutic and immunosuppressant drugs which interfere with PNB. Conflict of interest statement The authors have no conflicts of interest to report.
References Abrams R. Stability of the adenine ring structure in the rat. Biochim Biophys Acta 1956;21:439–40. Allsop J, Watts RWE. Purine synthesis de novo and salvage in hypoxanthine phosphorisyltransferase-deficient mice. Enzyme 1990;43:155–9. Baggott JE, Gorman GS, Morgan SL, Tamura T. 13C-enrichment at carbons 8 and 2 of uric acid after 13C-labeled folate dose in man. Biochem Biophys Res Commun 2007a;361:307–10. Baggott JE, Gorman GS, Tamura T. 13C Enrichment of carbons 2 and 8 of purine by folatedependent reactions after [13C] formate and [2-13C] glycine dosing in adult humans. Metabolism 2007b;56:708–15. Bennett EL, Karlsson H. Metabolism of adenine in mice. J Biol Chem 1957;229:39–50. Bertino JR, Simmons B, Donohue DM. Purification and properties of the formateactivating enzyme from erythrocyte. J Biol Chem 1962;237:1314–8. Block KA, Eells JT, Noker PE, Hawtrey CA, Tephly TR. Role of hepatic tetrahydrofolate in the species difference in methanol toxicity. Proc Natl Acad Sci USA 1985;82: 3854–8. Buchanan DL, Rollins JM. Lack of correlation between gout and the incorporation of isotopic formate into uric acid. Yale J Biol Med 1962;34:31–9. Caperelli CA, Benkovic PA, Chettur G, Benkovic SJ. Purification of a complex catalyzing folate cofactor synthesis and transformylation in de novo purine biosynthesis. J Biol Chem 1980;255:1885–90. Cheek WD, Appling DR. Purification, immunoassay and tissue distribution of rat C1tetrahydrofolate synthase. Arch Biochem Biophys 1989;270:504–12. Christensen KE, Mackenzie RE. Mitochondrial one-carbon metabolism is adapted to the specific needs of yeast, plants, and mammals. BioEssays 2006;28:595–605. Corton JM, Gillespie JG, Hawley SA, Hardie DG. 5-Aminoimidazole-4-carboxamide ribonucleotide. A specific method for activating AMP-activated protein kinase in intact cells? Eur J Biochem 1995;229:558–65.
Deacon R, Chanarin I, Lumb M, Perry J. Role of folate dependent transformylases in synthesis of purine in bone marrow of man and in bone marrow and liver of rats. J Clin Pathol 1985;38:1349–52. Deacon R, Bottiglieri T, Chanarin I, Lumb M, Perry J. Methylithioadenosine serves as a single carbon source to the folate co-enzyme pool in rat bone marrow cells. Biochim Biophys Acta 1990;1034:342–6. Drysdale GR, Plaut GWE, Lardy HA. The relationship of folic acid to formate metabolism in the rat: formate incorporation into purines. Journal of Biological Chemistry 1951;193:533–88. Fern EB, Garlick PJ. The specific radioactivity of the tissue free amino acid pool as a basis for measuring the rate of protein synthesis in the rat in vivo. Biochem J 1974;142: 413–9. Fu T-F, Rife JP, Schirch V. The role of serine hydroxymethyltransferase isozymes in onecarbon metabolism in MCF-7 cells as determined by C13 NMR. Arch Biochem Biophys 2001;398:42–50. Garrett RH, Grisham CH. The synthesis and degradation of nucleotides. Biochemistry. 3 rd ed. Belmont, CA: Thomson Brooks/Cole; 2005. chapter 26. Gorman GS, Tamura T, Baggott JE. Mass spectrometric method for detecting carbon 13 enrichment introduced by folate coenzymes in uric acid. Anal Biochem 2003;321: 188–91. Hanzlik RP, Fowler SC, Eells JT. Absorption and elimination of formate following oral administration of calcium formate in female human subjects. Drug Metab Dispos 2005;33:282–6. Jeong S-S, Schirch V. Role of cystolic serine hydroxymethyltransferase in one-carbon metabolism in Neurospora crassa. Arch Biochem Biophys 1996;335:333–41. Kastanos EK, Woldman YY, Appling DR. Role of mitochondrial and cytoplasmic serine hydroxymethyltransferase isozymes in de novo purine synthesis in Saccharomyces cerevisiae. Biochemistry 1997;36:14956–64. Kikuchi G. The glycine cleavage system: composition, reaction mechanism, and physiological significance. Mol Cell Biochem 1973;1:169–87. Lajtha LG, Vane JR. Dependence of bone marrow cells on the liver for purine supply. Nature 1958;182:191–2. Lincoln DW, Hrushesky WJM, Wood PA. Circadian organization of thymidylate synthase activity in normal tissues: a possible basis for 5-fluouracil chronotherapeutic advantage. Int J Cancer 2000;88:479–85. Marie S, Heron B, Bitoun P, Timmerman T, Van den Berghe G, Vincent MF. AICAribosiduria: a novel, neurologically devastating inborn error of purine biosynthesis caused by mutation of ATIC. Am J Hum Genet 2004;74:1276–81. Marsh WH. On the biosynthesis of purine in the bird; role of formate. J Biol Chem 1951;190:633–41. Matsuda Y, Kuroda Y, Kobayashi K, Katunuma N. Comparative studies on glutamine, serine, and glycine metabolism in ureotelic and uriotelic animals. J Biochem (Tokyo) 1973;73:291–8. McGreer PL, McGreer EG, Griffin MC. Excretion of 4-amino-5-imidazolecarboxamide in human urine. Can J Biochem Physiol 1961;39:591–603. Pasternak LB, Laude Jr DD, Appling DR. 13C NMR analysis of intercompartmental flow of one-carbon units into choline and purines in Saccharomyces cerevisiae. Biochemistry 1994;33:74–82. Sabina RL, Kernstine KH, Boyd RL, Holmes EW, Swain JL. Metabolism of 5-amino-4imidazolecarboxamide riboside in cardiac and skeletal muscle. J Biol Chem 1982;257:10178–83. Schwendel A, Grune T, Holzhutter HG, Siems WG. Models for the regulation of purine metabolism in rat hepatocytes: evaluation of tracer kinetic experiments. Am J Physiol 1997;273:G239–46. Shuster L, Goldin A. The effect of nicotinamide on incorporation in vivo of formate-14C. J Biol Chem 1959;234:129–33. Shuster L, Langan TA, Kaplan NO, Goldin A. Significance of induced in vivo synthesis of diphosphopyridine nucleotide. Nature 1958;182:512–5. Sidi Y, Mitchell BS. Z-nucleotide accumulation in erythrocytes from Lesch–Nyhan patients. J Clin Investig 1985;76:2416–9. Smaaland R, Laerum OD, Lote K, Sletvold O, Sothern RB, Bjerknes R. DNA synthesis in human bone marrow is circadian stage dependent. Blood 1991;77:2603–11. Smith GK, Mueller WT, Wasserman GF, Taylor WD, Benkovic SJ. Characterization of the enzyme complex involving the folate-requiring enzymes of de novo purine biosynthesis. Biochemistry 1980;19:4313–21. Stahelin HB, Stokstad ELR, Winchell HS. Formate oxidation and its incorporation into uric acid in folic acid deficiency. J Nucl Med 1970;11:247–54. Vincent MF, Marangos PJ, Gruber HE, Van den Berghe G. Inhibition by AICA riboside of gluconeogenesis in isolated rat hepatocytes. Diabetes 1991;40:1259–66. Vincent MF, Bontemps F, Van den Berghe G. Substrate cycling between 5-amino-4imidazolecarboxamide riboside and its monophosphate in isolated rat hepatocytes. Biochem Pharmacol 1996;52:999-1006. Wagner C, Levitch ME. The utilization of formate by human erythrocytes. Biochim Biophys Acta 1973;304:623–33. Wall M, Shim JH, Benkovic SJ. Human AICAR transformylase: role of the 4-carboxamide of AICAR in binding and catalysis. Biochemistry 2000;39:11303–11. Whiteley HR. The distribution of the formate activating enzyme and other enzymes involving tetrahydrofolic acid in animal tissues. Comp Biochem Physiol 1960;1: 227–47.