Isolation and identification of 3-(2'-methyl-4'-amino-5'-pyrimidylmethyl)-4-methylthiazole-5-acetic acid (thiamine acetic acid) and 2-methyl-4-amino-5-formylaminomethylpyrimidine as metabolites of thiamine in the rat.

Abstract The compounds 3-(2'-methyl-4'-amino-5'-pyrimidylmethyl)-4-methylthiazole-5-acetic acid (thiamine acetic acid) and 2-methyl-4-amino-5-formylaminomethylpyrimidine have been identified as metabolites of thiamine in the rat. Evidence is presented which indicates that 5-(2-hydroxyethyl)-4-methylthiazole is also an important metabolite of thiamine in the rat.

Examination of the urine of rats administered thiamine labeled with 14C in either the pyrimidine or thiaaole moiety reveals the presence of some 25 to 30 metabolites of the vitamin (I, 2). A similar number of metabolites can be seen in the urine of men administered pyrimidine-and thiaeole-labeled thiamine (3, 4). Only two of these metabolites have been identified.
In addition, the compound 5-(2-hydroxyethyl)-4-methylthiazole, which cannot be detected in urine of rats at physiological levels of intakes of thiamine, has been shown to be an important metabolite of thiamine.
* This investigation was supported by United States Public Health Service Grants AM-10297 and 5TOl AM-05441.
Most of this material was taken from a thesis submitted by W. H. Amos, Jr. to the graduate faculty of Vanderbilt University in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Biochemistry.
A preliminary account was given at the Fiftyfourth Annual Meeting of the Federation of American Societies for Experimental Biology, April 1970.
This latter compound can be isolated by extracting the reaction mixture with chloroform and vacuum-distilling the chloroform extract. The boiling point of 5-(2-hydroxyethyl)-4-methylthiazole is 135" at 7 mm. The compound 4-methylthiazole-5-acetic acid was synthesized by the method of Cerecedo and Tolpin (10).
The compound 2-methyl-4-amino-5-formylaminomethylpyrimidine was synthesized according to the method of Matsukawa and Yurugi (11). The yield was low, however, and the following method was developed.
One-half gram (3.6 mmoles) of 2methyl-4-amino-5-aminomethylpyrimidine, obtained by cleavage of thiamine with liquid ammonia, was dissolved in 5 ml of 90% formic acid, and the mixture was refluxed for 10 min. The reaction mixture was cooled and acetone (5 ml) was added.
The column was eluted by downward flow of a linear gradient of distilled water to 0.35 M pyridine acetate in 2000 ml; 12-ml fractions were collected at a flow rate of 120 ml per hour. The radioactivity in a 0.5.ml aliquot of each fraction was determined by gas flow counting. chromatography on Sephadex G-10 (Pharmacia), and thin layer chromatography on microcrystalline cellulose (Avicell, FMC Corporation, Newark, Delaware).

Animals and Treatment-Female
Sprague-Dawley rats were used in these studies. Animals were housed in cages that allowed for the separate collection of urine and feces. Food and water were supplied ad l&turn.
In the experiments in which expired gases were collected, the animals were maintained in a cage especially constructed for this purpose.
Those animals receiving physiological intakes of thiamine (60 pg per day) were maintained on a thiamine-deficient diet (Nutritional Biochemicals), supplemented by daily intubations of thiamine. The animals maintained on higher levels of thiamine were fed a stock laboratory ration and given thiamine at the appropriate level either by intubation or by placing the thiamine in the drinking water. Urine was collected as previously described (1). Xpectrophotometric Determinations-In the determinations of the concentration of thiamine acetic acid and 5-(2-hydroxyethyl)-4-methylthiazole in solution, the optical densities of the solutions of thiamine acetic acid were determined at 247 nm and of 5-(2hydroxyethyl)-4-methylthiazole at 260 nm. In the cases in which the specific activities of these compounds were also to be calculated, the radioactivity of an aliquot of the solution was determined by liquid scintillation counting and the specific activity was calculated.
Chromatographic Procedures-In a typical experiment, 500 ml of urine from rats administered 14C-labeled thiamine were concentrated to 40 ml under vacuum at 40" and centrifuged to remove the salts that had precipitated.
The concentrate was applied to a column of Amberlite CG-50 (2.5 x 70 cm) which had been prepared as described previously (1). The column was developed by downward flow of a linear gradient of distilled water to 0.35 M pyridine acetate in 2000 ml. Fractions of approximately 12 ml were collected at a flow rate of 120 ml per hour.
The radioactivity of 0.5-ml aliquots of each fraction was determined by planchet counting. Fig. 1 shows t,he typical pattern of radioactivity eluted from such a column. The metabolites of thiamine in the feces were extracted as described previously (3). The ion exchange chromatography of these metabolites was carried out in the manner described for urine.
The metabolites of thiamine isolated from urine by ion exchange chromatography were sometimes chromatographed on columns (1.5 x 105 cm) of Sephadex G-10 that had been equilibrated with 0.01 N HCl.
These columns were developed with 0.01 N HCI, and 5-ml fractions were collected at a flow rate of 60 ml per hour.
The radioactivity in these fractions was determined by planchet counting.
Thin-layer chromatography was performed on 0.5-mm layers of microcrystalline cellulose.

AND DISCUSSION
If thiamine were metabolized in the rat by a "thiaminase" enzyme, the products would be 5-(2-hydroxyethyl)-4-methylthiazole and some form of the pyrimidine moiety.
Although 5-@hydroxyethyl)-4-methylthiazole has been reported to be present in the urine of rats administered 1 mg of thiamine by a single intraperitoneal injection (12), we have not been able to detect its presence under similar conditions.
The pooled urine was subjected to ion exchange chromatography as described in Fig. 1. Peak B, which contains 4-methylthiazole-5acetic acid, was reduced in volume and set aside at 4" overnight. Under these conditions, the relatively insoluble 4-methylthiazole-5-acetic acid crystallizes from aqueous solution.
The infrared spectrum of the crystallized product from Peak B was found to be identical with that of authentic 4-methylthiazole+acetic acid. Thin layer chromatography of an aliquot of Peak C1 with Solvent System D revealed the presence of unmetabolized 5-(2-hydroxyethyl)-4-methylthiazole.
Peak C1 was then subjected to liquid-Issue of November 10, 1970 W. H. Amos, Jr., and R. A. Neal 5645 liquid extraction with diethyl ether for 12 hours. Thin layer chromatography revealed that all (8 mg) of the 5-(2-hydroxyethyl)-4-methylthiazole had migrated into the ether layer. The infrared spectrum of the material in the ether layer proved to be identical with that of authentic 5-(2-hydroxyethyl)-4-methylthiazole.
The amount of 4-methylthiazole-5-acetic acid isolated from the urine in this experiment (320 mg) indicated that the unlabeled 5-(2hydroxyethyl)-4-methylthiazole rather than thiamine was the major precursor of this compound under the conditions of the experiment.
One of the intermediates in the metabolism of thiamine by a soil microorganism is thiamine acetic acid (7). This is the compound in which the hydroxyethyl side chain of thiamine is oxidized to an acid. Thiamine acetic acid is the major source of 4-methylthiazole-5-acetic acid in this microorganism (7). Accordingly, the urine of rats was examined for thiamine acetic acid. Thiamine acetic acid is eluted from an Amberlite CG-50 column in the position of Peak C1 (Fig. 1). This peak of radioactivity from the urine of rats administered thiazole-2-'*C-thiamine was reduced in volume, and a portion was chromatographed on a thin layer of cellulose along with authentic thiamine acetic acid. The thin layer plate was developed with Solvent System A. The subsequent autoradiogram showed the presence of eight radioactive components, one of which had an RF similar to that of thiamine acetic acid. The cellulose containing this compound was scraped from the plate, and the compound was eluted from the cellulose with distilled water.
The eluate was reduced in volume, and equal amounts were spotted, together with unlabeled thiamine acetic acid, on five plates coated with cellulose.
The plates were developed with Solvent Systems A, B, C, E, and F. Subsequent autoradiograms showed a radioactive compound that migrated in a manner similar to that of thiamine acetic acid in all five solvent systems. The RF values of thiamine acetic acid and the unknown compound in the five solvent systems are shown in Table I. Thiamine acetic acid, like thiamine, forms a highly fluorescent compound when treated with alkaline ferricyanide.
When these five plates were sprayed  Fig. 1. The mixture was repeatedly recrystallized from hot water.
After each crystallization, a portion of the crystals was removed and their specific activity was determined as described under "Experimental Procedures." with a solution of 1% ferricyanide in 1 M sodium hydroxide, a fluorescent band exactly corresponding to the radioactive compound appeared.
These data strongly suggested the presence of thiamine acetic acid in the urine.
To confirm its presence, unlabeled thiamine acetic acid (30 mg) was added to a sample of lyophilized material corresponding to Peak C1, Fig. 1. The mixture, which was dissolved in a minimum amount of boiling water, was set aside at 4" overnight.
A portion of the crystals of thiamine acetic acid was collected, and the specific activity was determined as described under "Experimental Procedures." The remaining crystals were redissolved, and the crystallizations and determination of specific activity were repeated four more times. The results of these repeated recrystallizations are shown in Table II. Note that the specific activity of the thiamine acetic acid decreased to a constant value after the first recrystallization and remained constant throughout the remaining crystallizations.
These data confirmed the presence of thiamine acetic acid in rat urine.
Thiazole-labeled thiamine acetic acid was administered to rats at a level of 70 pg per rat per day. Subsequent examination of the urine revealed that about 10% had been metabolized to 4methylthiazole-5-acetic acid. The remainder of the radioactivity in urine was accounted for by unmetabolized thiamine acetic acid. Thus, analogous to the microorganism (7), there would appear to be two precursors for 4-methylthiazole+acetic acid in the rat, these being 5-(2-hydroxyethyl)-4-methylthiazole and thiamine acetic acid. Because thiamine acetic acid and 5-(2hydroxyethyl)-4-methylthiazole have a common precursor and because the rate of metabolism of each metabolite is not &ecu.-  2. The most abundant ions from the mass spectra of (a) authentic 2.methyl-4-amino-5-formylaminomethylpyrimidine and the compound isolated from the bacterial culture and (b) the unknown isolated from rat urine.
All spectra were obtained using the direct sample introduction probe.
rately known, it has not been possible to determine the relative contribution of these two compounds to the 4-methylthiazole-5acetic acid excreted in the urine.
Isolation and Identification of 6-Methyl-Q-aminod-formylaminomethylpyrimidine as a Metabolite of Thiamine-About 60% of the radioactivity in Peak CZ, Fig. 1, from rats maintained on physiological levels (60 pg per day) of either thiazole-or pyrimidine-labeled thiamine was found to be accounted for by an ultraviolet-absorbing compound with an RF of about 0.55 in Solvent System A. Thus, the compound contained all or portions of both the pyrimidine and thiazole moieties. This same compound was also seen in animals administered thiamine acetic acid. This indicated that the hydroxyethyl side chain of the thiazole moiety was either missing or oxidized to an acid. The compound did not give a positive thiochrome reaction, and it was not markedly affected when dissolved in a solution of sodium bisulfite. These data indicated that the structure of thiamine had been extensively modified.
At this point, it was noted that an unidenti6ed compound which was a minor metabolic product of the metabolism of thiamine by a microorganism (14) and which was eluted from Amberlite CG-50 in a peak equivalent to Peak Cs, Fig. 1, had the same RF as the unknown in Solvent Systems A, B, C, E, and F.
A sufficient amount of this bacterial metabolite was available so that it could be purified by crystallization and examined by mass spectrometry.
The mass spectrum of the bacterial metabolite is shown in Fig. 2a. What appears to be the molecular ion of the compound occurs at m/e 166. The most prominent ion in the spectrum is m/e 137. There is also a prominent ion at m/e 122. The ion at m/e 137 could correspond to a loss of CHO from the molecular ion, and the ion at m/e 122 is typical of the intact pyrimidine moiety of thiamine (15). These data indicated that the compound might be 2-methyl-4-amino-5-formylaminomethylpyrimidine.
Consequently, this compound was synthesized as described under "Experimental Procedures." The infrared and mass spectra of the synthesized compound proved to be identical with those of the bacterial metabolite. An attempt was made to obtain a mass spectrum of the metabolite from rat urine.
Oral administration of large amounts of labeled thiamine to rats (40 mg per day) led to the accumulation of relatively large amounts of the unknown compound in urine. The metabolite in these urines was partially purified by chromatography on columns of Amberlite CG-50.
It was further purified by chromatography of the equivalent of Peak C,, Fig. 1, on a column of Sephadex G-10 as described under "Experimental Procedures." The radioactive peak from the Sephadex G-10 column was lyophilized to dryness.
The residue remaining after lyophilization was dissolved in a small amount of hot 0.1 N HCl and cooled; acetone was added until the solution became cloudy, and the material was set aside at 4" overnight.
Approximately 5 mg of crystals were collected by centrifugation and washed with a small amount of acetone. Although the mass spectrum ( Fig.  2b) of this partially purified material resembled that of the synthetic 5-formylaminomethylpyrimidine (Fig. Za), it was not an exact duplication.
One reason for the discrepancy was the presence of higher molecular weight impurities in the sample of the unknown.
Repeated attempts to further purify this material with the use of Amberlite CG-50, Sephadex G-10, and thin layer chromatography were not successful. Although the mass peaks at m/e 122, 96, and 69 are more abundant relative to the parent peak than in synthetic 5-formylaminomethylpyrimidine, the major peaks, both in the mass at which they appeared and in their relative heights, indicated that the metabolite from rat urine was the 5-formylaminomethylpyrimidine.
In order to obtain further data as to the identity of the unknown, 50 pg of synthetic 2-methyl-4-amino-5-formylaminomethylpyrimidine and about 7,000 dpm of the unknown, which had a specific activity of approximately 10,000 dpm per pg, were applied to thin layers of cellulose as spots 1 cm in diameter.  Table III. In each solvent system, the area of radioactivity of the mixture coincided exactly with the area of ultraviolet absorption.
In addition, the area of radioactivity of the mixture was continuous with the area of radioactivity of the unknown, and the area of ultraviolet absorption of the mixture was continuous with the area of ultraviolet absorption of the synthetic 2-methyMamino-5-formylaminomethylpyrimidine.
These results confirm the hypothesis that the unknown compound is 2-methyl-4-amino-5-formylaminomethylpyrimidine. Fig. 3 summarizes what is known concerning the metabolism of thiamine (1) in the rat. There appear to be two precursors for 4-methylthiazole&acetic acid (IV). One of the precursors, 5-(2.hydroxyethyl)4methylthiazole (III), is probably formed by the action of a "thiaminase"-like enzyme which also yields 2-methyl-4-amino-5-hydroxymethylpyrimidine (V) The 5-(2hydroxyethyl)-4-methylthiazole is then oxidized to 4-methylthiaeole-5-acetic acid by what is probably a two-step reaction. The second precursor is thiamine acetic acid (II) formed by the oxidation of thiamine.
In a microorganism (8), thiamine is oxidized to thiamine acetic acid by a single enzyme without the release of the intermediate aldehyde.
Whether the oxidation is catalyzed by one or two enzymes in rats remains to be determined.
The 5-hydroxymethylpyrimidine, which is postulated to be formed in the two "thiaminase" reactions, does not accumulate in the urine of animals on physiological intakes of thiamine.
This compound, however, can be isolated from the urine of animals administered large oral doses of thiamine.1 In addition, when the 5-hydroxymethylpyrimidine is administered to rats, it is metabolized to 2-methyl4-amino-5-pyrimidinecarboxylic acid (VI) (3). These data indicate that the 5hydroxymethylpyrimidine is an intermediate in the metabolism of thiamine. The oxidation of the 5-hydroxymethylpyrimidine to the 5-pyrimidinecarboxylic acid is probably a two-enzyme process.
It is not possible on the basis of the present data to speculate on the metabolic route from thiamine to 2-methyl&amino-5-formylaminomethylpyrimidine (VII) and to sulfate. The metabolic route by which carbon 2 of the thiazole moiety is released as CO* is also not yet established, although it appears that the 5-formylaminomethylpyrimidine may be one of the important precursors. In adult rats on physiological intakes of thiamine (60 pg/day), 4-methylthiazole-5-acetic acid accounts for about, 35% of the metabolites of the thiazole moiety of thiamine other than thiamine itself excreted in the urine. The relative amounts of this compound derived from thiamine acetic acid and 5-(2-hydroxyethyl)I-methylthiazole are yet to be determined. The compound 2-methyl-4-amino-5-formylaminomethylpyrimidine accounts for about 11% and thiamine acetic acid for about 4% of the metabolites of both the pyrimidine and thiazole moieties of thiamine excreted in the urine.
About 15 to 20% of the pyrimidine moiety of thiamine is excreted in the urine in the form of 2-methyl-4-amino-5-pyrimidinecarboxylic acid (3). Another 10% of the carbon 2 of the thiazole moiety is metabolized to COZ in rats maintained on physiological levels of thiamine (16). Thus, about 60% of that portion of the thiazole moiety containing carbon 2 and about 35% of the pyrimidine moiety can be accounted for by the metabolites shown in Fig. 3 Vol. 245,No. 21 percentages are distributed among some 20 to 25 as yet unidentified metabolites.
That all of these metabolites are of mammalian origin rather than products of the intestinal microflora is supported by several findings. Neal and Pearson (3) demonstrated the formation of 2-methyl-4-amino&pyrimidinecarboxylic acid from pyrimidine-%14C-thiamine by germ-free rats. Suzuoki et al. (6) also used germ-free rats to demonstrate that 4-methylthiazole-5-acetic acid was not a product of the microbial degradation of thiazole-2-l%thiamine.
It has also been found that the pattern of metabolites in the urine of animals maintained on a diet containing 1% succinylsulfathiazole is both qualitatively and quantitatively similar to that in the urine of animals not receiving the drug. 2 We have examined human urine for the presence of thiamine acetic acid and of 2-methyl-4-amino-5-formylaminomethylpyrimidine.
In both cases, a small amount of the labeled metabolite was added to approximately 1 liter of human urine, and the urine was subjected to the same ion exchange chromatography and thin layer chromatography procedures as described for rat urine. On the basis of the amounts of these compounds reisolated by these procedures, it appears certain that both metabolites are also present in human urine.
Thus, the qualitative pattern of thiamine metabolism in humans appears similar to that in rats.