Nicotinic Acid Metabolism

Previous reports from this laboratory have described the isolation of 6-hydroxynicotinic acid and 6-0x0-1 ,4,5,6-tetrahydronicotinic acid from the reaction media of a clostridial species grown on nicotinic acid (1). Bn enzyme that catalyzes the reversible oxidation of nicotinic acid to HNA’ has been purified from extracts of this microorganism (2). Since fortified crude extract of this Clostridium can metabolize HNA as well as nicotinic acid (l), this degradation was studied to define further the fermentation of nicotinic acid in this microorganism and the role of 6-0x0-1 ,4,5,6-tetrahydronicotinic acid. This paper reports the partial purification of an enzyme which catalyzes the reversible reduction of HNA to the 4-5-dihydroderivative, 6-0x0-1,4,5,6-tetrahydronicotinic acid. This enzyme requires reduced ferredoxin or reduced methyl viologen dye as the electron donor for the reduction of HNA.

In the presence of pyruvate, other intermediates accumulated; two of these were identified, one as 1,4,5,6tetrahydro-6-oxonicotinic acid, and the other as a!-methyleneglutaric acid.
The pathways for nicotinic acid degradation may be considered as involving successive oxidative and reductive steps to 1,4,5,6-tetrahydro-6-oxonicotinic acid, then through some unidentified steps to a-methyleneglutaric acid, which is subsequently converted to propionic and acetic acids and carbon dioxide.
Pyruvate and lactate were shown not to be obligatory intermediates in the production of propionate. This microorganism has a high content of Blz coenzyme, but its function in nicotinic acid fermentation is not yet understood.
The isolation of a clostridium that ferments nicotinic acid to propionate, acetate, CO,, and ammonia has already been described (1). Based on isotopic studies with nicotinic acid labeled at various specific positions, two possible schemes of nicotinic acid degradation were proposed. This paper describes the isolation and structural proof of two intermediates in the fermentation. In addition, pool experiments in which pyruvate was used to trap isotope derived from the variously labeled nicotinic acids have been performed. These show that pyruvate is not a free intermediate in t.he pathway to propionate.
Finally a Blz coenzyme has been found to be present in the organism in large amounts.
This suggests that a Blz coenzyme is involved in some step of the fermentation.

EXPERIMENTAL PROCEDURE
The organism was grown on the media described previously in 300-liter fermentation tanks under a nitrogen atmosphere (1). * Fellow of the National Foundation.
Present address, National Institutes of Arthritis and Metabolic Diseases, National Institutes of Health, Bethesda, Maryland 20014.
Yields were about 1 g of cells (wet weight) per liter. The cells were harvested with a Sharples centrifuge, lyophilized, and stored in a desiccator at -15".
Cell-free extracts were prepared by suspending 6 g of dried cells in 60 ml of 0.03 M potassium phosphate buffer, pH 7.3, containing 0.03% NazS .7Hz0, and passing this through a French pressure cell. The particulate material was sediment.ed at 16,000 x g for 1 hour and discarded.
The supernatant fraction could be stored in aliquots in a liquid nitrogen freezer for 2 months without appreciable loss of activity.
6-Hydroxynicotinic acid was measured by its absorption at 290 rnp at pH 7.3. It could also be distinguished from nicotinic acid by its failure to react with cyanogen bromide (2). Pyruvate was isolated from reaction mixtures as the 2,4-dinitrophenylhydrazone, and its specific activity was determined on the chromatographically purified material (3). Propionate, acetate, and lactate were isolated by chromatography on Celite (4). Carbon dioxide was collected in NaOH and was precipitated and plated as barium carbonate.
Protein was measured by sulfosalicylic acid precipitation (5). The biuret reaction gave falsely high values which were probably due to a substance present in the crude bacterial extracts.
Other details are described in a previous paper (1). 6-Hydroxynicotinic acid was purchased from Light Chemical Company, sodium pyruvate from Calbiochem, and a-methylglutaric acid from Aldrich. y-Hydroxy-y-methyl-cr-ketoglutaric acid was a gift from Dr. A. Marcus, Beltsville, Maryland.
Elemental analyses were made by Micro-Tech Laboratories, Skokie, Illinois.
All melting points were determined on a Kofler hot stage apparatus.
Ultraviolet absorption spectra were measured with a Cary recording spectrophotometer, model 14M; infrared spectra were determined by Mrs. K. Warren on a Beckman IR-7 spectrometer; nuclear magnetic resonance spectra were measured with a Varian Associates model A-60 spectrometer. Vapor phase chromatography was performed on two different columns: (a) a g-foot column of 0.75y0 SE-30 on Gas-Chrom P and (b) a 6-foot column of 1% QF-1 on Gas-Chrom P.
The reaction was stopped with 0.25 ml of 10% perchloric acid, and the precipitate was centrifuged and discarded.
The supernatant was passed through a Dowex 50-H+ column (0.8 X 5 cm) which was washed with water, and 5.0-ml fractions were collected.

Isolation of Intermediates A and B
When nicotinic acid was decomposed in the presence of large amounts of pyruvate, intermediates other than 6-hydroxynicotinic acid were found to accumulate.
To facilitate the isolation and characterization of these intermediates, a large scale experiment was performed as follows.
A mixture containing 0.2 M 7-r4C-nicotinic acid (0.21 PC per mmole), 0.2 M sodium pyruvate; 0.1 M potassium phosphate (pH 7.0), 0.015% NazS. 7Hz0, and 96 g of dried cells in a total volume of 2,400 ml was incubated for 2+ hours at 30" under a helium atmosphere. The reaction was stopped by the addition of 60% perchloric acid with stirring to pH 2.0. The insoluble material was sedimented at 15,000 x g for 30 min, and the precipitate was discarded.
The supernatant was brought to pH 7.0 with KOH.
The precipitate of KCIOI was removed by centrifugation.
The liquid phase was passed through a Dowex 50-H+ column (5 x 45 cm) to remove nicotinic acid. The effluent solution was acidified with 6 N H&SO4 to pH less than 2.0 and was continuously extracted with ether for 24 hours. The ether-soluble material was con- centrated to a small volume under a stream of helium, diluted to 10 liters with water, adjusted to pH 5.0 with 2 N NH40H, and applied to a Dowex 1-formate column (6 X 15 cm). The column was washed with water, followed by NH4-formate buffer, pH 4.0, of increasing molar&y (0.05 to 0.53 M) (Fig. 2). Fractions of 15 ml each were collected, and 0%ml aliquots were counted in a liquid scintillation spectrometer.
The optical density at 270 rnp was also monitored.
As can be seen from Fig. 2, this chromatographic procedure resolved the ether extract into major radioactive components, hereafter referred to as Intermediates A, B, and C. Fractions containing Intermediates A were pooled, as were those containing Fraction B, and lyophilized.
The material was redissolved in water, adjusted to pH 2.0 or less with HzS04, and extracted continuously into ether for 24 hours. The ether-soluble material was concentrated to a small volume, diluted with 100 ml of HzO, and again lyophilized.
This material was taken up in the appropriate organic solvent and crystallized as described below.
The yields of recrystallized A and B were 76 and 17 mg, respectively.
The methyl ester of Intermediate A was obtained by treating a crude sample, 28 mg, in 2 ml of methanol with a large excess of The infrared spectrum (chloroform) showed absorption at 1704 cm-' (strong) (conjugate ester C=O), 1656 cm-l (strong) (amide C=O), and 3408, 3250, and 3140 cm-i (nonbonded and bonded N-H) (Fig. 5). The nuclear magnetic resonance spectrum was determined in CDC& ( Fig. 7) with tetramethylsilane as internal reference.
A sample of the methyl ester, 14 mg, in 2 ml of acetic acid and The infrared spectrum (chloroform) showed absorptions at 1740 cm-i (strong) (ester C=O), 1667 cm+ (amide C=O), and 3405 and 3240 cm-1 (nonbonded and bonded N-H) (Fig. 6). This methyl ester was identical in melting point and infrared spectrum with an authentic sample of methyl 5-ketonipecotate prepared as follows.
A mixture of 0.5 g of 6-hydroxynicotinic acid in 10 ml of 0.75 M NH40H and 0.2 g of Rh-Al203 (5%) catalyst was shaken with hydrogen at 25 p.s.i. and room temperature for 4 hours. The catalyst was removed by filtration, and the filtrate was concentrated under reduced pressure to 2 ml; acidification gave a precipitate of 5-ketonipecotic acid. The dried acid was suspended in 10 ml of methanol and treated with an ethereal solution of diazomethane.
After removal of solvent, the residue was sublimed in a vacuum to furnish methyl 5ketonipecotate, m.p. 107-109".
The chromatographic behavior of Intermediate A, its infrared absorption between 1600 and 1800 cm-l, and its reactivity toward diazomethane are consistent with the properties of a carboxylic acid. Examination of the infrared spectrum of the methyl ester of A reveals (a) a secondary amide group (Y,,, 1656 cm-l and 3408 cm-l, Fig. 3) which is present in a cyclic structure since it lacks the absorption of an amide II band at about 1550 cm+ (6), and (b) an (Y ,&unsaturated ester, (vmax 1704 cm-l, Fig. 4), shifting to 1740 cm+ upon catalytic hydrogenation (Fig. 6). The elemental analysis of Intermediate A, C6HTN03, indicated that it contains 2 more hydrogen atoms than 6-hydroxynicotinic acid; the presence of an cY,/3-unsaturated ester and the la&am function suggest the possible structure to be Ia or Ib. Consideration of the ultraviolet absorption spectrum and group of the ester, the broad peak at 6 = 8.08 ppm to N-H, analysis of the nuclear magnetic resonance spectrum of the and the doublet at 6 = 7.31 ppm to an olefinic proton adjacent methyl ester of A led conclusively to the assignment of Structure to a N-or 0-function. That t.his olefinic proton is definite13 Ia for Intermediate A. In the nuclear magnetic resonance adjacent to the NH function was demonstrated by observing spectrum of the methyl ester of A in CD'& (Fig. 7), the comples the spectrum in the presence of trace of DzO upon which the multiplet at 6 = 2.63 ppm can be attributed to -CH&H2-at N-H absorption at 8.08 ppm disappeared while the doublet at positions 4 and 5, the singlet at 6 = 3.78 ppm to the -CH, 7.31 ppm collapsed into a singlet. None of these spectral features  ,  I  I  I  I  I  I  I  I  I  I  I  I  I  I  I   The material from tubes 230 to 290 (Fig. 2) was purified by recrystallization from ethyl acetate-cyclohexane. The pure compound, m.p. 130-130.5", had no ultraviolet absorption above 225 rnp, decolorized neutral permanganate instantly, and showed a broad absorption band in the infrared spectrum (KBr pellet) at 1710 to 1730 cm-1 (acid C=O).
Calculated: C 50.12, H 5.60 Found : C 50.00, H 5.60 Quantitative hydrogenation was conducted in a microhydrogenator (7) with palladium on charcoal (5%) as catalyst; 1.646 mg of Intermediate B absorbed 0.25 ml of Hz (an equimolar quantity to a compound with the formula C6HS04). The catalyst, was removed by filtration, and the filtrate was evaporated to dryness under nitrogen.
The residue, m.p. 66-69", was taken up in 0.5 ml of methanol and treated with excess diazomethane in ether. After standing for 1 hour at room temperature, t,he solution was concentrat.ed, and samples were analyzed by vapor phase chromatography in two different columns. In both cases a single peak was obt.ained; it had a ret'ention time (12.6 min for Column a at 73" and 2.4 min for Column b at 86" under 10 p.s.i. of carrier gas) identical with that of an authentic sample of dimethyl cr-methylglutarate.
The above characteristics indicated that Intermediate B is probably a-methyleneglutaric acid. This identity was firmly established by comparison of Intermediate B with an authentic sample of cu-methyleneglutaric acid prepared according to Buchman, Reims, and Schlatter (8). r;tilization of Intermediates Jletabolism of 6-llydroxynicotinic &id-Incubation of 6hydroxynicotinic acid with cell-free extracts that complet.ely ut.ilized nicotinic acid did not result in any disappearance of the 6.hydroxynicotinic acid (Table I). When pyruvate was added, however, 6-hydroxynicotinic acid was completely utilized and was not converted to nicotinic acid. In the conversion of nicotinic acid to 6-hydroxynicotinic acid, electrons are lost from the molecule. These ultimately appear in the fermentation products since the over-all ferment.ation can be represented by a balanced equation (9). It is likely that pyruvate supplies the electrons that. are lacking in 6-hydroxynicotinic acid and allows its metabolism in this system, which ordinarily carries out the tightly coupled metabolism of nicotinic acid. Hydrogen gas, another possible source of electrons, was unable to act like pyruvate. It is not known if it is ut.ilized by these extracts. Extracts-Previous studies showed t.hat the ferment,ation of nicotinate by washed cell suspensions of the clostridium involves a mechanism in which one half of the COz and one half of the propionate carboxyl carbon are derived from carbon atom 7 (carboxyl group of nicotinate) ; the other half of each fermentation product is derived from carbon atom 6 of nicotinate (I).
The data of Table II   From considerations of the structures of the three intermediates that have been isolated, it is supposed t.hat the decomposition of nicotinic acid involves successive oxidative and reductive steps to form 1,4,5,6-tetrahydro-6-oxonicotinate, followed by a series of unidentified steps to form ol-methyleneglutarate; this is illustrated in Fig. 8. On the basis of previous studies with variously labeled nicotinate substrates (l), it seems likely that the further conversion of a-methyleneglutarate involves the formation of two identical C3 derivatives or of two dissimilar CI compounds that are in equilibrium with each other.1 Ultimately the two C1 compounds are converted to propionate, acetate, and COZ. This is illustrated by the lower part of Fig. 8.
In view of the fact that various CB compounds (lactate, pyru- pyruvate.
The results are summarized in Table III. It can be seen that with either labeled substrate the specific radioactivity of propionate formed was very much greater than the specific activities of the reisolated pyruvate.
This result precludes an obligatory role of free pyruvate in the conversion of either C-6 or C-7 of nicotinate to propionate.
The fact that the specific activity of the propionate is considerably less than that of the added nicotinate is attributed to the dilution of labeled propionate with unlabeled propionate present in the crude extract and possibly also to the formation of propionate from the pool of unlabeled pyruvate. Table III shows that lactate isolated from Experiment B had a relatively low specific isotope content; therefore lactate cannot be an obligatory intermediate in the conversion of C-7 of nicotinate to propionate.
In considering possible mechanisms for the conversion of cymethyleneglutarate to propionate, a mechanism involving y-hydroxy-y-methyl-a-ketoglutarate appeared reasonable since the latter compound occurs in plants, where it has been shown to undergo enzymatic cleavage to 2 moles of pyruvate (10). However, neither y-hydroxy-y-methyl-c-ketoglutarate nor its lactone is metabolized by cell-free extracts.
Whereas this appears to rule out these compounds as normal intermediates per se, the possibility that a derivative (viz. the CoA derivative) is involved cannot be excluded.

Coenzymes
During isolation of the intermediates of nicotinate metabolism, it was noted that the organism contains rather large amounts of a reddish compound that was partially purified and was identified as a vitamin Blz derivative.
To isolate the compound, 1 g of dried cells was extracted with 80% ethanol at 70" for 30 min. The B,z coenzyme was purified by phenol extraction followed by ion exchange chromatography (11). The absorption spectrum is similar to that of a benzimidazole cobamide type of coenzyme (Fig. 9). On exposure to tungsten light, a characteristic peak appears at 350 rnp, and on the addition of cyanide, a characteristic peak at 367 rnp is seen. Based on the absorption spectrum, the content of coenzyme is 0.084 I.cmole per g of dried cells.
resolve the system for a B~z coenzyme does not eliminate its participation, since it may be involved in a step that is not ratelimiting in our assay, or it may be tightly bound to an enzyme as in the methylmalonyl-CoA-succinyl CoA mutase studied by Mazumder, Sasakawa, and Ochoa (14).