Metabolism of Gallic Acid and Syringic Acid by Pseudomonas putida*

Abstract Cell-free extracts of Pseudomonas putida, grown with syringic acid as carbon source, catalyzed the oxidation of 1 mole of gallate (3,4,5-trihydroxybenzoate) by 1 mole of oxygen to give 2 moles of pyruvate and 1 mole of carbon dioxide. Oxaloacetate, formed as an intermediate in reactions, could be reduced quantitatively when NADH and sufficient malate dehydrogenase were added, despite the presence in extracts of oxaloacetate decarboxylase. With these conditions of assay it was shown that oxaloacetate was the precursor of half of the pyruvate formed from gallate. Similarly, when 1 mole of gallate was first oxidized by protocatechuate 4,5-oxygenase purified from Pseudomonas testosteroni and was then degraded by cell-free extracts, 1 mole of oxaloacetate and 1 mole of pyruvate were again formed. Although gallate oxygenase proved too labile to isolate, it was concluded from these results that the enzyme gave the same ring fission product as that obtained by the action of protocatechuate 4,5-oxygenase on gallate, namely 4-carboxy-2-hydroxy-cis,cis-muconic acid. This compound appeared to undergo enzymic hydration to give 4-carboxy-4-hydroxy-2-oxoadipate which was then cleaved to oxaloacetate and pyruvate. Cell-free extracts oxidized 3-O-methylgallate, but protocatechuate (3,4-dihydroxybenzoate), 5-carboxy-3,4-dihydroxy benzoate and 3,4-dihydroxy-5-methylbenzoate were not attacked. Syringic acid was oxidized only in the presence of substrate amounts of NADH, when 2 moles of oxygen were taken up and 1 mole of pyruvate appeared; additional pyruvate was formed by a much slower reaction. There was no requirement for added NADH when concentrated cell extracts were used to oxidize syringate. These and other observations indicate that the C-5 methoxyl group of syringate was oxidized by an O-demethylase. The benzene nucleus of the resulting 3-O-methylgallate was then cleaved between C-3 and C-4 to give, as final products, pyruvate and a compound presumed to be a monomethyl ester of oxaloacetic acid. The substrate specificity of syringate O-demethylase was studied in experiments with intact cells.

4,5-oxygenase purified from Pseudomonas testosteroni and was then degraded by cell-free extracts, 1 mole of oxaloacetate and 1 mole of pyruvate were again formed. Although gallate oxygenase proved too labile to isolate, it was concluded from these results that the enzyme gave the same ring fission product as that obtained by the action of protocatechuate 4,5-oxygenase on gallate, namely 4-carboxy-2-hydroxy-cis,cis-muconic acid. This compound appeared to undergo enzymic hydration to give 4-carboxy-4-hydroxy-2oxoadipate which was then cleaved to oxaloacetate and pyruvate.
Syringic acid was oxidized only in the presence of substrate amounts of NADH, when 2 moles of oxygen were taken up and 1 mole of pyruvate appeared; additional pyruvate was formed by a much slower reaction. There was no requirement for added NADH when concentrated cell extracts were used to oxidize syringate. These and other observations indicate that the C-5 methoxyl group of syringate was oxidized by an 0-demethylase.
The benzene nucleus of the resulting 3-0-methylgallate was then cleaved between C-3 and C-4 to give, as final products, pyruvate and a compound presumed to be a monomethyl ester of oxaloacetic acid. The substrate specificity of syringate 0-demethylase was studied in experiments with intact cells. *  Gallic acid occurs in the free state in tea and many other plants (1) and is also encountered in the form of esters, notably in gallotannins (2). A biosynthetic pathway that involves the dehydrogenation of 5-dehydroshikimate has been proposed for gallic acid (3), but little information is available concerning its enzymatic degradation.
Beveridge and Hugo (4) isolated cu-ketoglutarate as its 2,4-dinitrophenylhydrazone from cultures of Pseudomonas conuexa metabolizing gallic acid, and they suggested that this compound might be formed directly after fission of the benzene nucleus.
However, no experiments were performed with cell-free extracts, and the possibility was not excluded that the oc-ketoglutarate obt.ained might arise from the operation of the tricarboxylic acid cycle. Micro-organisms capable of utilizing gallate as their sole source of carbon for growth are not readily isolated (4) ; moreover, solutions of this substrate rapidly darken in color when shaken in air at neutral or alkaline pH. Accordingly, we isolated for use in this investigation a strain of Pseudomonas @ida which oxidized gallate readily to completion after growth at the expense of syringic acid, namely the 3,5-dimethylether of gallic acid. Syringic acid is stable in aqueous solution and is also a natural product which has been extracted from certain lignins when exposed to the action of wood-rotting fungi (5).

MATERIALS AND METHODS
Organism-The strain utilized in these experiments was isolated from soil in St Preparation of Cell-free Extracts-Cells were grown in batches of 40 liters in a fermentor provided with stirring and vigorous forced aeration.
The growth medium was inoculated with two cultures, each of 50 ml, which had grown overnight in 250-ml Erlenmeyer flasks placed on a gyrotary shaker. After growth in the fermentor for 14 hours, a second addition was made of 20 g of sodium syringate and the culture was harvested 4 hours later; yield, 70 g wet weight of cells.
Cells were washed and then crushed without abrasive in the bacterial press of Hughes (8)) and cell-free extracts were prepared in 0.1 M KHz-NasH-phosphate buffer of pH 7.0 as described by Dagley et al. (9). The protein contents of extracts were determined by a modification of the biuret method (10). Ability to oxidize gallic acid was lost when extracts were maintained at 60" for 7 min, cooled in an ice bath, and the precipitate removed by centrifuging.
The clear supernatant', referred to as a heat-treated extract, still cont,ained active enzymes that metabolized the ring fission product of gallic acid to oxaloacetate and pyruvate.
These enzymes, and also gallate oxygenase, were effectively absent from crude extracts of cells grown with succinate as carbon source.
Chemic&-Gallic acid, of analytical reagent grade, was from Mallinckrodt Chemical Works, St. Louis, and syringic acid was from Aldrich Chemical Company.
Solutions of 4-hydroxy-4-methyl-2oxoglutarate and 4-carboxy-4.hydroxy-2-oxoadipate were prepared from their lactones by mild alkaline hydrolysis with N KOH (13) made either by Warburg respirometry with air as the gas phase (16) or by means of the oxygen electrode as described by Dagley et al. (9).
Chromatography-The procedures used to identify the 2,4-dinitrophenylhydrazone of pyruvate by means of thin layer and paper chromatography were those described by Dagley and Gibson (16).

Oxidation of Substrates by Intact Cells-Suspensions
of intact cells, grown at the expense of syringic acid (III, Scheme I), oxidized gallic acid (I), and 3-O-methylgallic acid (II) at almost equal rates, although the final uptake of oxygen for II was less than that for I ( Fig. 1). Syringic acid was oxidized more slowly, but when uptake ceased the amount of oxygen consumed was approximately the same as for gnllic acid. With vanillic acid (IV) and 3,5-dimethoxybenzoic acid (V) the oxygen consumed was about 75% of that required for converting one methoxyl group of each substrate into a hydroxyl group, carbon dioxide, and water.
Oxidation of Substrates by Cell-free Extracts-Orrygen was rapidly consumed when reaction mixtures of 3 ml, containing 54 mg of extract protein and 5 pmoles of either gallate, S-O-methylgallate, or syringate, were shaken in a Warburg respirometer. However, this technique proved to be of limited value for investigating stoichiometry, since extracts at this concentration showed substantial oxygen uptake in the absence of added substrates.
On dilution, the gallate oxygenase activities of extracts diminished and a proportion of the gallate remained unchanged when amounts of substrate convenient for respirometry were shaken with diluted extracts.
Measurements of oxygen used in ring fission and 0-demethylation were therefore made by means of the oxygen electrode, which requires much lower concentrations of substrates and permits the use of dilute extracts.
Under these conditions 0.4 Imole of oxygen was consumed by 0.4 pmole of gallic and 3-0-methylgallic acids; and there was no attack on 3,4-dihydroxy-5-methylbenzoic and 5-carboxy-3, $-dihydroxybenzoic acids in which the hydroxyl group at C-5 (or C-3) of gallic acid is replaced by methyl and carboxyl, respectively (Fig. aa). Syringic acid was not oxidized unless NADH was present; 0.4 pmole of oxygen was then taken up by 0.2 pmole of substrate (Fig. Zb). By contrast, no increase in oxygen consumption was observed for 3-O-methylgallic acid when NADH was present (Fig. 2~). There was no requirement for NADH when syringate was oxidized in the Warburg respirometer.
It is probable that, under these conditions, NADH required for 0-demethylation (18)  required to initiate oxidation (Fig. 2b). Six reaction mixtures were therefore set up, each of which contained in a final volume of 3 ml, 0.25 pmole of syringic acid, 115 pmole of Tris-HCl buffer of pH 7.5, 2.8 mg of cell extract protein, and 0.5 pmole of NADH. The mixtures were then divided into three pairs. The first pair were incubated for 15 min at 36", the second for 30 min, and the third for 45 min. 811 reactions were stopped by heating in a water bath for 2 min at loo", denatured protein was removed by centrifuging, and 2-ml aliquots were then withdrawn for spectrophotometric determination of pyruvate using lact,ate dehydrogenase. In each determination, a cuvette containing the aliquot from one reaction mixture of a pair received no NADH, and this solution served as a blank when oxidation of KADII was measured in the second aliquot.
From the mixtures incubated for 15 min, 30 min, and 45 min, the amounts of pyruvate formed from 1 pmole of syringic acid were, respectively, 1.16, 1.24, and 1.36 pmoles. In Fig. 2b, the oxidation of syringic acid was complete within 3 min. Assuming that 1 pmole of pyruvate was formed from 1 pmole of substrat,e in this period, as was the case for 3-O-methylgallate, it appears that further reactions occurred which gave rise to an increase in the pyruvate concentration of about 1% per min.
Formation of Oxaloacetate from Galtate-Carbon dioxide was evolved when gallate was oxidized by crude cell-free extracts. When the reaction was conducted in a Warburg respirometer that contained 12 mg of extract protein, gas exchange ccascd when 1 pmole of carbon dioxide had been evolved and 1 pmole of oxygen consumed.
In Scheme 2 it is proposed that 1 mole of carbon dioxide is liberated from the oxaloacetate formed from 1 mole of gallate, a reaction readily catalyzed by cell extracts.
This decarboxylase activity was an obstacle in dcmonstrat'ing directly the conversion of gallate into oxaloacetate, since fractionation procedurcs that removed the decarboxylasc invariably resulted in simultaneous loss of gallnte oxygenase.
However, when a high concentration of malate dehydrogenase was added to reaction mixtures, oxaloacetnte was reduced faster than it was dccnrboxylatcd, and it became possible to observe a stoichiometric oxidation of added NADH.
Thus, 2 pmoles of SADHI pmole of gallate were oxidized when cell extract was added to a reaction mixture containing lactate dehydrogenase.
When this enzyme was replaced by malate dehydrogenase, only 1 pmole of NAl)H was oxidized; but a further 1 pmole of K\';\DH reacted when lactate dehydrogenasc was then added (Fig. 3). A similar experiment showed that 4-carbory-4-hydroxy-20r;oadipate (VIII, Scheme 2) was also cleaved by cell extracts to give equimolar amounts of oxaloacetate and pyruvatc.
Although, as previously observed, 1 pmole of NADII was oxidized rapidly when lactate dchydrogenase was incubated with 3~0~metl~ylgallate and cell extract, a different result was obtained for malate dehydrogennse.
When this enzyme was present with 3-0-methylgallate in the incubation mixture, the oxidation of KhDH proceeded initially at only onefifth of the rate shown for gallate in Fig. 3 and virtually ceased when about 0.5 pmole of NADII had been oxidized for 1 pmole of 3-0-methylgallatc present.

Enzymic
Ring Fission of Gatlic Acid-The loss of gallate oxygenase activity during fractionation was also an obstacle to the use of cell extracts for preparing the ring fission product of gallic acid. Inactivation may have been due to the lnbilc nature of the protein.
Alternatively, since activity was abolished completely when 1 mM of or,&-dipyridyl was added to reaction mixtures, it appeared that this enzyme, like certain other dioxygenases (19) is Fe++-dependent.
However, when attempts were 6441 made to activate the enzyme by treating with Fe++ ions, it was found necessary t,o remove the excess since ferrous sulfate reacted rapidly with the substrate, b vallic acid, to give a dark green complex. It was also difficult to achieve this removal effectively without, simultaneous inactivation, since Fe++ ions appeared to be very loosely bound to the enzyme.
An alternative experimental approach became available when it was found that gallate was attacked by protocatechuate 4,5-oxygenase (EC 1.13.1.8). First, this enzyme, which has been purified from Pseudomonas testosteroni (9), was used to accumulate a ring fission compound from gallic acid. Second, this same compound was shown to be metabolized rapidly and quantitatively by extracts of Pseudomonas p&da to give, like gallic acid, equimolar amounts of oxaloacetate and pgruvate.
It was therefore concluded that protocatechuate 4,5-oxygenase and gallate oxygenase cleaved the benzene nucleus of gallic acid to give the same ring fission compomld.
These experiments were conducted as follows, using a preparation of protocatechuate 4,5-oxygenase (24 mg of protein per mg) which was kindly provided by Dr. J. M. Wood, University-of Illinois.
An extract of syringate-grown Pseudomonas pulida u-as also used which, after heat treatment, was devoid of gallate oxygenase but retained activity towards other metabolites in the proposed reaction sequence (Scheme 2). X cuvette contained, in a final volume of 3 ml, 123 pmoles of Tris-HCl buffer of pH 8.0, 0.41 mg of NADH, 1.2 mg of protocatechuate 4,5-oxygenase, and 5 ~1 of lactate dehydrogenase. No oxidation of K\',41>H was observed when 0.2 pmole of gnllic acid was added.
After incubating for 1.5 min, during which period the absorbance remained unchanged, an addition of 0.05 ml of heat-treated extract (0.42 mg of protein) was made and this was followed immediately by a decrease in absorbance at 340 nm. When lactate dehydrogenase was omitted, and replaced by 5 ~1 of malate dehydrogenase, oxidation of NAl)H was again dependent upon the addition of heat-treated estract.
With lactate dehydrogenase the decrease 03- A second cuvette contained 34.5 units of lactate dehydrogenase and received 3 mg of cell extract protein at 3. Oxidation of NADH by cell extract alone is shown by the broken lines.
in absorbance corresponded to reduction of 0.38 pmole of pyruvate. In the second experiment with malate dehydrogenase, 0.19 pmole of oxaloacetate reacted; and when lactate dehydrogenase was then added, a further decrease in absorbance indicated the presence of 0.19 pmole of pyruvate.
Therefore, from 1 mole of gallate, oxidized to the ring fission product of protocatechuate 4,5-oxygenase, approximately 2 moles of pyruvate were formed by the heat-treated extract of Pseudomonas pulida; and of this yield, 1 mole of pyruvate arose from oxaloacetate which could be trapped by reduction to malate.
These results support the proposed structure (VII, Scheme 2) for the product of ring fission of gallate.
Further evidence was obtained by observing the changes in ultraviolet absorption when 50 ~1 of protocatechuate 4,5-oxygenase were added to 0.5 pmole of gallic acid in 3 ml of 0.1 M phosphate buffer, pH 7.0. An immediate decrease in absorbance at 260 nm, the wave length of maximum absorbance for gallic acid at pH 7.0, was accompanied by an increase at 310 nm; these changes were complete after 2 min. On addition of sufficient NaOH to bring the pH to 12.6, the peak at 310 nm was shifted to 355 nm with increased absorbance (Fig. 4); the shift was freely reversible on acidification.

Ultraviolet
absorption spectra of gallic acid, and the product formed by ring fission of gallate using protocatechuate 4,5-oxygenase.
The concentration of gallic acid, and of the ring fission product prepared as described in the text, was 0.17 $1~. 1, gallic acid at pH 7.0; 2, ring fission product at pH 7.0; 3, ring fission product at pH 12.6. 6442 1 pmole of gallate into 2 pmoles of pyruvate with the consumption of 1 pmole of oxygen.
Such extracts contained a powerful decarboxylase which rapidly converted oxaloacetate into pyruvate; but it was possible to demonstrate the formation of 1 pmole of oxaloacetate from 1 pmole of gallate by adding NADH, together with sufficient malate dehydrogenase to reduce oxaloacetate faster than it was decarboxylated.
Evidence of the involvement of 4-carboxy-Z-hydroxy-cis, cis-muconic acid (VII in Scheme 2) was less direct, and stemmed from the finding that gallate was oxidized by protocatechuate 4,5-oxygenase, purified from another organism (Pseudomonas testosteroni).
From the known properties of this enzyme (9,21) it is highly probable that the benzene nucleus of gallate is cleaved between C-3 and C-4 (or C-4 and C-5) to give VII, although fission between C-2 and C-3, to give an aldehyde acid, might be a formal possibility.
However, this second alternative is discounted by the spectroscopic properties of the ring fission product, which were those of a substituted 2-hydroxymuconic acid, such as VII.
When gallate was incubated, first with protocatechuate 4,5-oxygenase and then with an extract of syringate-grown Pseudomonas putida that lacked gallate oxygenase, it was found that approximately 1 pmole each of oxaloacetate and pyruvate were formed from 1 pmole of gallate.
Although, therefore, gallate and protocatechuate oxygenases arc separate enzymes, differing at least insofar as gallate oxygenase does not attack protocatechuate, it appears that they oxidize gallate to the same ring fission product, VII. 4-Carboxy-4-hydroxy-2-ketoadipic acid (VIII) was rapidly cleaved to oxaloacetate and pyruvate by an aldolase present in extracts of syringate-grown Pseudomonas; this enzyme has now been purified (22). Insofar as it involves enzymic hydration and retroaldol cleavage, the reaction sequence of Scheme 2 resembles those for the meta-fission pathways studied previously (23); but it differs as follows.
The compounds formed by oxidative ring fission of catechol, 3-methylcatechol or 4-methylcatechol all give rise to an oxoenoic acid, either by hydrolytic fission (24) or by oxidation followed by decarbosylation (25). The osoenoic acid so formed is then hydrated.
In Scheme 2, however, direct hydration of the ring fission product is shown.
Oxygen reacted readily with 3-O-methylgallate, although the rate was somewhat less than that for gallate (Fig. Za). It appears, for the following reasons, that the benzene nucleus was cleaved between C-4 and C-3, which bears the methoxyl group, rather than between C-4 and C-5 which bears hydroxyl.
(a) One of the products formed from 3-0-methylgallate (pyruvate) was rapidly reduced by NADH when lactate dehydrogenase was added; but the other product differed from oxaloacetic acid insofar as NADH was slowly, and apparently incompletely, oxidized in the presence of malatc dehydrogenase.
Neither lactate dehydrogenase nor malate dehydrogenase catalyzed a reaction between 3-0-met,hylgallate and NADH until cell extract was added. (b) The ultraviolet spectrum of the ring fission product of 3-O-COOH COOH methylgallate, like that of gallate (Fig. 4) showed an absorption peak at 310 nm which disappeared on the addition of dilute sodium hydroxide with the appearance of a new peak at 355 nm. This alkali shift may be attributed to ionization of the enol group at C-2 (VII, Scheme 2) and would not occur if C-2 carried a methoxyl group; however, it would not be prevented by esterification of the C-6 carboxyl. Protocatechuate 4,5-oxygenase, and not gallate oxygenase, was used to prepare solutions of ring fission product; but this compound was rapidly metabolized to pyruvate by heat-treated extracts of syringate-grown Pseudomonas pufida. The oxidation of syringic acid by cell extracts was not initiated until NADH was added (Fig. 2b). This requirement for NADH indicates that ring fission is preceded by 0-demet,hylation according to the following equation: Syringatc + O2 + NADH + H+ + 3-0-methylgallate + HCHO + NAD+ Ribbons (18) showed that an analogous equation described 0-demethylation of vanillate, veratrate and m-methoxybenzoate by extracts of Pseudomonas aeruginosa.
We found that extracts catalyzed an NAD+-linked oxidation of formaldehyde; and the ability of concentrated cell extracts to oxidize syringate without additions of NADf may be accounted for by the coupling of demethylation with the reaction: Formaldehyde + HzO + NAD+ + formate + NADII + H+ Of the 2 pmoles of oxygen taken up by 1 pmole of syringate in the oxygen electrode experiments (Fig. 2b), 1 pmole of oxygen would be required in 0-demethylation and 1 pmole for ring fission of the 3-O-methylgallate formed; that is, only one methoxyl group was oxidized.
It was also found (Fig. 1) that only one of the two meta-placed methoxyl groups of 3,5-dimethoxybenzoate was oxidized by intact cells. Degradation of 3-0-methylgallate would account for the rapid formation of 1 pmole of pyruvnte observed when syringate was incubated with cell extracts, and the slow appearance of more pyruvate on prolonged incubation (about 0.01 pmole per min) may bc due to the slow release of oxaloacetate from its monomethyl ester, followed by decarboxylation. We have no information as to the significance of this suggested hydrolysis in the metabolism of syringate by the organism, although it may be noted that when rapid oxidation ceased at 156 min ( Fig. 1) the total amount of oxygen consumed by intact cells was significantly less for 3-0-methylgallate than for gallate, despite the fact that 1 mole of the former substrate provides more carbon than the latter.
This observation suggests that methylat.ion has rendered a position of the gallic acid molecule less susceptible to rapid degradation.
We found that an NAD+-linked methanol dehydrogenase was present in cell extracts, which might be used to metabolize methanol eventually released from a methyl ester. COOH SCHEME 2 644.3