Studies of a Flavoprotein, Salicylate Hydroxylase

Salicylate hydroxylase has been shown to react with certain substrate analogues, effecting an “uncoupling” of oxygen reduction from hydroxylation (WHITE-STEVENS, R. H., AND KAMIN, H. (1972) J. Biol. Chem. 247, 2358-2370; (1970) Biochem. Biophys. Res. Commun. 38, 882). The mechnism of action of this enzyme has now been examined by spectrophotometric and stopped flow techniques. Compounds which facilitate DPNH binding and oxidation perturb the absorption spectrum of the enzyme. From this perturbation, an apparent K, commensurate with catalytic K, values for benzoate and salicylate can be obtained. With salicylate, one salicylate is bound per flavin. Anaerobic reduction with limiting DPNH yields only fully reduced and fully oxidized flavin; semiquinone is formed only with photoreduction in the presence of EDTA. The two flavins act independently. As observed in the stopped flow, formation of enzyme-substrate complex is very rapid; the presence of salicylate facilitates DPNH binding and determines the rate of enzyme reduction to E-FADH2. The catalytic V,,,, for DPNH oxidation with several organic compounds appears to be reflected specifically in the rate constants for enzyme reduction. The reoxidation of reduced enzyme by 02 is independent of the nature or even the presence of aromatic compound and is strictly proportional to O2 tension over a wide [02] range. Depending upon the aromatic molecule, either the rate of reduction of enzyme, the rate of oxidation, or a combination of both can be rate-limiting for catalysis. The hydroxylating species, designated “E-FAD-H202,” must be short lived and can either (a) stoichiometrically hydroxylate a suitable substrate such as salicylate, (b) decompose rapidly and quantitatively to HzOz if the substrate is totally unsuitable (i.e. benzoate), or (c) hydroxylate with erratic formation of H202, for substrates which act in a manner intermediate between

(I -hydroxylation, l-decarboxylating)), has been induced and isolated from a soil bacterium grown on salicylate as sole carbon source.
When benzoate is substituted for salicylate, DPNH is oxidized with the same Max as with salicylate but with higher Km for both benzoate and DPNH.
With salicylate, the reaction products are catechol and HZO; with benzoate, the benzoate is unchanged, but H2U2 is formed stoichiometrically with DPNH oxidized. Both salicylate and benzoate facilitate DPNH binding. Benzoate binds at the salicylate site, competitively inhibiting salicylate hydroxylation, and permitting DPNH binding and oxidation. But since benzoate, a "pseudosubstrate," cannot be hydroxylated, the oxygen utilized decomposes to H202, and oxygen reduction is considered as C'uncoupled" from hydroxylation. A search for possible active intermediates in the reduction of O2 to Hz02 by the uncoupled reaction has failed to yield evidence for any oxygen radical species.
An examination of various substituted benzoates and salicylates has revealed a range of behavior intermediate between the "substrate" (salicylate) and pseudosubstrate (benzoate) modes. These compounds are hydroxylated, (in several cases more rapidly than salicylate), but some of the oxygen utilized is diverted to H&.
These compounds are bound to enzyme with a K, higher than the inducer, salicylate, and also facilitate DPNH binding, but less effectively than does salicylate. This enzyme, as isolated from a strain of &udomonas putida, has been described and extensively studied by Yamamoto et al. (I), Katagiri et al. (2,3), and Takemori et ai. (4,5). Its mechanism of action has been studied by spectrophotometric (3,4), Auorometric (6)) and stopped flow (5) techniques.
The enzyme contains one FAT) and one polypept,ide chain per 57,200 molecular weight (l), and exhibits a strong specificity for substrates bea,ring hydrosyl and carboxyl substituents at ortho positions.
Recent publications from this laboratory (7-9) have described a salicylate hydroxylase, which has different physical and kinetic properties, from a different (and as yet unidentified) soil microorganism.
Our enzyme consists of 2 moles of FAD and two subunits per 91,000 molecular weight, and has the distinctive property of catalvzing a benzoat,e-stimulated oxidation of ZIPNH with the same catalytic VnRX as salicylate hydrosylation, but without detectable alteration of the benzoate molecule (7, S), The characteristics of this reaction suggest that, in the presence of benzoate, oxygen is activated or reduced in a "normal" fashion, but hydroxylation cannot proceed. The product of oxygen reduction in this case is H& We have called benzoate a 'Q)seudosubstrate" and have termed t,he benzoate effect "uncoupling of oxygen activation from hydroxylation" (7)(8)(9). Since the l)ublication of our original report (7) on the benzoate effect, two laboratories have reported analogous effects in other flavoprotein hydroxylases (10, 11). A partial "uncoupling" effect had been observed previously by Storm and Kaufman (12) with the pteridine-con taining phenylalanine hydrosvlase f ram ra 6 liver. They found that structural changes in e%her substrate  or pteridine cofa#ct,or would affect the TPNH to tyrosine stoichiometry; where excess TPNH was utilized, HZ& was also detected.
The present paper documents further some of our previous reports, and describes additional experiments designed to es-amine more fully the effects of salicylate, benzoate, and other aromatic compounds.
Stopped Aow studies have been per, formed and will be reported in a subsequent palmer (13). Such studies, combined with those reported in the present paper, have permitted us to propose a, react,ion mechanism (13) which accommodates "substrates," "pseudosubstrates," and compounds I-hich displav both modes of activitv. u d It was made up as a 1% suspension in water and titrated to pH 7.0 with KOH.
The resultant mixture was quite turbid, but was stable in the cold. Hydrogen peroxide, purchased as the 30y0 Mallinckrodt analytical reagent, was diluted in water and standardized by its molar extinction of 43.6 at 240 nm. FAD, FMN, and riboflavin were from Sigma. Crystalline beef liver cat,alase of 150,000 units per ml was obtained from Worthington.
Yeast alcohol dehydrogenase and horse heart cytochrome c, type VI, were both from Sigma. Bovine superoxide dismutase was a gift from Dr. Irwin Fridovich 04).

Growth of Microorganism
The organism was maintained and subcultured in a liquid basal medium containing sodium salicylate as the sole carbon source. The medium, similar to that of Katagiri  For large scale growth, 40.liter carboys were inoculated with a l-liter inoculum and the bacteria were grown with vigorous aeration at 37". The carboys were heated with a Briskeet heating tape, and temperature was cont.rolled with a Yellow Springs Instrument Co. telethermometer. The culture medium darkened with growth, and this darkening was greater at higher pH. Cultures became alkaline with time; if the pH rose above 8.0, the yield of enzyme fell. HCl was therefore usually added as sodium salicylate was consumed.
Salicylate concentrations greater than 0.27, inhibited growth and this compound was therefore added in several increments.
Salicylate concentration was maintained between 0.02 and 0.2%) and was monitored during growth by withdrawing an aliquot of the culture medium and reading The organisms were harvested in a Sharples, Lourdes, or De Lava1 continuous flow centrifuge, and when the volume of culture in the fermentor was reduced to about 1 liter, 40 liters of additional sterile medium were pumped in. This second lot of bacteria was grown up and harvested as previously described. The resultant cell paste from 80 liters of medium was washed three to four times with 0.033 M potassium phosphate, pH 7.00, and frozen; about 300 g of bacteria were produced.
The frozen cell paste was stable for months.

Extraction and Purilfication of Enzyme
The purification procedure was a modification of that of Yamamoto et al. (1). The progress of purification is summarized in Table I. Step 1. Xonication-Frozen cell paste was thawed and homogenized with 4 volumes of 0.033 M potassium phosphate buffer, pH 7.0, in a Brinkmann Polytron homogenizer for 2 min to break up clumps of bacteria.
The suspension was then sonicated at full power in a beaker (in 800.ml batches) for 20 to 30 min, in 5-min bursts, with a Heat Systems-Ultrasonics, Inc., model W185D sonifier. Temperature was kept below 15" with an ice bath.
The sonicate was centrifuged for 1 hour at the highest speed (50,000 x g) in the Sorvall RC2B centrifuge, SS 34 rotor.
The dark brown supernatant solution was carefully decanted from the cell debris and a black, gelatinous precipitate.
Step 2. Protamine Sulfate Treatment-Protamine sulfate had been used by Hosokawa and Stanier (16) and Yamamoto et al. (1) to remove nucleic acid in the course of purification of their enzymes. We also used this technique but found that our salicylate hydroxylase was precipitable at higher protamine sulfate concentrations, permitting differential precipitation and subsequent elution of enzyme.
The sonic supernatant was adjusted to a protein concentration of 5 mg per ml. Protamine sulfate was slowly added in the cold as a 1 y0 suspension.
One hundred fifty to 300 ml per liter of enzyme (higher amounts for higher protein concentration) were usually sufficient to sediment nucleic acid without precipitating enzyme. After this precipitate was removed by centrifugation (15 min at 15,000 rpm), additional protamine sulfate solution, in 150-to 200-ml increments per liter of original sonic supernatant was added, stirred for 2 to 3 hours, and recentrifuged.
The supernatant was assayed between pro-2360 Xalicylate Hydroxylase. I. Properties Vol. 247, n-o. 8 tamine additions, and the procedure was repeated until less than phate, pH 7.0, and frozen. The yield of enzyme was about 10% of the original activity remained. 20 to 25% of that in the original sonic extract. The enzyme-protamine sulfate precipitates were suspended in 25.ml portions of 0.033 M potassium phosphate buffer, pH 7.00, containing 0.05 M NaCl. One hour of stirring at 37" in 50.ml centrifuge tubes with 3 inch magnets was sufficient to elute most of the activity.
The suspension was then centrifuged 15 min at 15,000 rpm and the yellow supernatant collected.
Reextraction of the protamine sulfate precipitate with the same buffer would increase yield without lowering specific activity.
In early preparations, an added step of hydroxylapatite chromatography (19) was included, and samples of enzyme purified through this step were used for molecular weight and electrophoresis experiments to be described.
However, this procedure did not give appreciable additional purification, and subsequent preparations omitted this step.

Enzyme Assays
Step S. Ammonium Sulfate Fractionation-The combined eluates from the protamine sulfate step were fractionated between 47 and 65% ammonium sulfate saturation, using the formula of di Jeso (18). The precipitate was allowed to stand 4 hours before centrifuging 15 min at 15,000 rpm. The precipitate from the 65% fractionation was reconstituted quickly in 0.02 M K2HPOd and dialyzed overnight against that solution.
Step 4. DEAE-cellulose Chromatography-The reconstituted ammonium sulfate fraction was applied to the top of a DEAEcellulose column, 1.7 x 17 cm, previously equilibrated with 0.02 M K~HPOI.
After washing the column with several hundred milliliters of KzHPOd, elution was begun with a linear gradient of 500 ml of 0.02 M KzHPOd and 500 ml of 0.2 M K2HPOb. Fractions of 13 to 16 ml were collected.
The peak of enzyme activity was eluted between 0.06 and 0.09 M phosphate concentration.
The appearance of activity coincided with the appearance of protein and of 450-nm absorbance.
The peak fractions were pooled, concentrated by a 45 to 65% ammonium sulfate fractionation, dialyzed against 0.02 M potassium phos-Protein concentration was determined by the microbiuret method (20) with bovine serum albumin as a standard.
The albumin in turn was standardized by its extinction at 279 nm of 0.67 for 1 mg per ml (21). For procedures such as column chromatography, protein concentration was monitored from optical densities at 280 and 260 nm. The extinctions for free FAD, cited by Yagi (22) at 280, 260, and 450 nm, were assumed for bound FAD.
Enzyme assays were performed with a Gilford 2400 or Cary 14 recording spectrophotometer and a standard assay mixture of 1 mM EDTA, 133 pM sodium salicylate, 147 pM DPNH, and 0.02 M potassium phosphate buffer, pH 7.62, in a volume of 3 ml with a l-cm path length cuvette.
Addition of FAD (6.7 pM) or omission of EDTA had no effect on the activity.
One unit of activity represented the oxidation of 1 pmole of DPNH per min measured at 340 nm and 27". Occasional assays required cells of path lengths from 0.1 to 10 cm. These were calibrated before use with standardized solutions of DPNH. Assays at very high pyridine nucleotide concentrations (10 to 20 mM), and other procedures requiring measurement of oxygen uptake, were performed in a Gilson model KM Oxygraph equipped with a Clark oxygen electrode and thermostatted at 25". Full scale deflection was set at standard oxygen tension at 25" which was assumed to be 240 PM.
At high substrate concentrations, substrate absorption would sometimes interfere at 340 nm. In these cases spectrophotometric assays were performed at 360, 365, or 370 nm with the appropriate extinction coefficients for DPNH.
It was sometimes advantageous to assay at 296 nm, the absorption maximum of salicylate.
At this' wave length a AA296 of -3400 was established for the salicylate to catechol reaction.
This extinction, combined with the AA296 of -1300 for the reaction DPNH to DPN+ gave an effective molar extinction coefficient of -4700 M-l cm-l. 1. Induction of salicylate hydroxylase by growth of bacteria on salicylate.
Bacteria were grown to population limit in 500 ml of 1% Difco yeast extract broth, harvested, and incubated in 500 ml fresh broth for 2 more hours. Freshly harvested bacteria were washed free of broth twice with 33 mM potassium phosphate, pH 7.00, and suspended in 0.2y0 sodium salicylate minimal media with vigorous aeration at 37". Of this medium, 250-ml aliquots were withdrawn approximately every half-hour. The optical density of the culture was read at 600 nm and that of the supernatant at 296 nm. The ordinate for "salicylate in medium" is not shown, but corresponds to a starting point of 2.04 g per liter and a finishing point after 8.8 hours of 1.37 g per liter. Bacterial samples were washed twice with phosphate buffer, resuspended in 5 ml of buffer, sonicated 1 min, and centrifuged at 27,000 X g, for 30 min. The specific activity in units per mg of salicylate hydroxylase found in the sonic extracts, as well as the optical densities at 600 and 296 nm of the culture medium, are plotted as a function of time.

EXPERIMENTAL PROCEDURE AND RESULTS
Selection of Organism---The salicylate hydroxylase described herein was isolated from an organism obtained from garden soil a few millimeters from a freshly creosoted telephone pole. The organism was selected by its ability to grow in medium containing 0.17, salicylate as sole carbon source. The culture was purified by selection of single colonies from plates, containing either 1% Difco yeast extract, or 0.1% sodium salicylate in minimal medium, with 2.5% Bacto-agar as a solid support.
The colonies grown on salicylate-agar, like the liquid cultures, turned dark with growth.
The organism, a gram-negative rod, has not been identified. ' It grows only aerobically, and will grow on dextrose or yeast extract broths.
The sole nitrogen source used was KNOP and, as expected, the bacterium has nitrate reductase activity.
The data of Fig. 1  and resuspended in salicylate minimal medium, salicylate hydroxylase activity appeared as depicted.
Synthesis of enzyme g 2 by the cell preceded both the consumption of salicylate in the medium and growth of the organism.
After an induction period of 9 hours, the specific activity of the enzyme reached a level of 0.27 unit per mg of protein.
This compares with a level of 0.8 to 1.5 units per mg observed in crude sonic extracts of fermentorgrown bacteria.  eluates. The enzyme also migrated as one peak in the analytical ultracentrifuge; these experiments will be described shortly.
Characterization of Enzyme: Flavoprotein-The spectrum of the purified salicylate hydroxylase shown in Fig. 3 is that of a typical flavoprotein bearing no other chromophoric prosthetic groups.
Addition of a small amount of sodium dithionite effected a reduction of the 450~nm peak to that of the fully reduced flavin.
The fluorescence of solutions of enzyme and of free FAD, both having the same absorbance at 450 nm, was measured.
At an excitation wave length of 462 nm, the fluorescence emission of free FAD at 520 nm was 13.3 times higher than that for the flavoprotein, indicating a quenching of flavin fluorescence upon binding to the protein. The supernatant from a boiled solution of the enzyme was chromatographed on paper by the method of Yagi (22). Solutions of free FAD, FMN, and enzyme supernatant were spotted in the dark onto Whatman No. 1 filter paper and developed with 5% Na2HP04 as solvent.
Samples of boiled enzyme supernatant were almost indistinguishable from samples of pure FAD, The progress of sedimentation at 25" was recorded at 8-min intervals (from left to right) after reaching a rotor speed of 56,100 rpm. evidence for catalytically significant reducible groups other than flavin. EDTA even at concentrations of 0.01 M does not inhibit. Zinc, copper, and iron sulfates, at concentrations of 1 X 10h3 M, had no effect on activity, nor did potassium cyanide, sodium azide, or ascorbate. A plot of In (radius from center of rotor to enzyme peak) versus time was linear over the 80-min period of measurement.
A partial specific volume of 0.728 was determined by the method of Schachman Photographic plates were exposed after centrifuging at 23,141 rpm for 24 and 28 hours. A plot of the natural logarithm of t,he displacement of five fringes (averaged) ver.su.s the square of the fringe radius yielded a straight line, whose slope corresponded to a molecular weight of 91,000. This molecular weight compares with a value of 57,200 obtained by Yamamoto et al. (1). The linearity of this plot over the range examined suggested the presence of but one molecular weight species.
The molecular weight of reduced and carboxymethylated protein was studied by means of gel filtration.
Salicylate hydroxylase (5 mg) was denatured and reduced in 6.0 M guanidine hydrochloride plus 0.1 M mercaptoethanol and carboxymethylated with a 10% excess of iodoacetamide at pH 8.5. The protein was dialyzed against 0.01 M acetic acid, lyophilized, and the entire reduction procedure repeated. The final product was taken up in 0.17 ml of 6 M guanidine-HCl, mixed with DNP-alanine and blue dextran as low and high molecular weight markers, and applied to the top of a 4v/, Agarose A-5M column. The column was eluted with 6.0 M guanidine-HCl by the method of Fish et al. (26), and fractions of approximately 0.9 ml were collected, weighed, and read for optical density at 280 nm. The elution curve indicated only one large symmetrical peak, whose position corresponded to a molecular weight of 43,000, in comparison with standard curves previously run on the same column (26). The symmetry and the presence of a single peak suggest the presence of only one type of subunit.
Peptide mapping has not yet been performed.
The two peak tubes from the 6 M guanidine-HCl gel filtration column were diluted to an A280 of 0.190 (about 0.2 mg per ml) with 6 M guanidine-HCl and analyzed by sedimentation equilibrium. The molecular weight was calculated as discussed previously, assuming a density of 1.141 g per cc for 6.0 M guanidine-HCl and a partial specific volume of 0.718 (0.01 cc per g less than that for native protein (27, 28)) for the denatured protein.
The physical properties of purified Ealicylate hydroxylase are summarized in Table II and are consistent with a structure of two FAD and two subunits per mole. A minimum molecular weight per flavin of 48,700 was calculated from the peak fraction in the DEAE-cellulose eluate, measuring protein by microbiuret reaction (20) and using a millimolar extinction coefficient of 11.3 for FAD (22). This value is in reasonable agreement with the subunit molecular weight previously cited. Specific activity of various preparations was proportional to flavin content. It is clear that the enzyme of this present study is physically different from the salicylate hydroxylase of Yamamoto et al. (1). Their enzyme is composed of one FAD and one subunit per 57,200 molecular weight.
Our enzyme is a molecule of 91,000 molecular weight with two FAD and two subunits. Characterization of Enzyme Activity--The requirements for salicylate hydroxylase activity are described in Table III. Salicylate hydroxylase was 76% as active with TPNH as with DPNH under those conditions.
In the absence of salicylate, pyridine nucleotide oxidase activity was low: 2 to 4a/, of normal Attempts to determine the optimum pH for the DPNH-and TPNH-dependent salicylate hydroxylase activities were complicated by anion effects (see below) of the different buffers of overlapping pH ranges which were used. In phosphate buffer, which was not inhibitory, the pH optimum for DPNH-dependent salicylate hydroxylase activity was broad with only small differences between pH 7.0 and 8.5, and sharp drops in activity either below pH 6.0 or above pH 9.0. For TPNH-dependent salicylate hydroxylase, the pH optimum was shifted to pH 6.0 to 7.5.
Product Study-The data of Fig. 5 show the enzymatic conversion of salicylate to catechol, employing a coupled assay with salicylate hydroxylase and yeast alcohol dehydrogenase.
Ethanol was added to 0.5 M concentration as a source of reducing power, and DPNH was added in only catalytic quantities. Change of the salicylate spectrum into that of catechol can be seen as the reaction progresses with time.
Comparison with an independently run spectrum of catechol at the same concentration as the salicylate used indicated that 98% of the salicylate was converted to catechol.
The existence of isosbestic points at 251 and 280 nm suggest that only two aromatic species were present during the reaction. Independent assays with separate enzymes showed that ethanol had no effect on salicylate hydroxylase activity at the levels used, and salicylate had no effect on alcohol dehydrogenase.
Stoichiometry---The product st'udy in Fig. 5  The data for the apparent K, values of pyridine nucleotides rate of reaction at 340 nm with excess DPNH was measured as in the presence of high concentrations of salicylate, and at ata function of salicylate concentration; the data so obtained are mospheric 02 tension, are presented in Fig. 7. It can be seen plotted in Fig. 6 by the method of Lineweaver and Burk (29).
that, although the apparent K, for TPNH (99 PM) is approxi-The K, so obtained was 2.7 PM. mately 6 times that for DPNH (17 PM), the V,,, values for the Concentrations of salicylate in excess of 5 mM were found to two nucleotides are virtually identical. As will be shown ( This in-upon the presence of salicylate; without aromatic substrate, the hibition is probably an anion effect, as will be described presently. DPNH apparent K, was 710 FM and TPNH apparent K, was 20 mM. Thus salicylate markedly decreases the apparent K, of pyridine nucleotide. This phenomenon had also been observed with P. putida salicylate hydroxylase (2); with other bacterial flavoprotein hydroxylases, the inducer also appears to markedly decrease pyridine nucleotide K, (10, 11,16,30).
Inhibition by Monovalent Anions-The salicylate hydroxylase of Yamamoto et al. (1) was routinely assayed in Tris-hydrochloride at pH 8.0. Our salicylate hydroxylase showed diminished activity in Tris-HCl buffer as compared to phosphate, but Trissulfate was not inhibitory.
It was established that the inhibiting factor was chloride, since NaCl would also inhibit catalysis; the curves of percentage inhibition versus molarity of Cl-were hyperbolic.
The point of 50% inhibition was estimated graphically as 0.06 M Cl-. Points at lower con-hydroxylase activity. This inhibition was roughly related to centrations represent averages of four to seven separate deter-anionic size in a pattern following the Hofmeister Lyotropic minations.
series. The data are presented in Fig. 8 anion, inhibited least whereas iodide and thiocyanate, the largest tested, inhibited most.

Nature of Anionic Inhibition-Inhibition
of salicylate hydroxylase by sodium chloride was shown to be competitive with respect to salicylate, as seen in Fig. 9. With other anions, the situation was more complex.
With thiocyanate, the lines of a Lineweaver-Burk plot (1 /v versus 1 /[salicylateJ) at different thiocyanate concentrations intersected somewhat to the left of the l/v axis; with iodide the pattern diverged even more from that expected for competitive inhibition.
A small portion of the inhibition seen in the presence of excess NaBr and NaI (but not NaCl and NaSCN) could consistently be reversed by the addition of FAD.
Thus, at 0.2 M NaI concentration, inhibition was 89% in the absence of FAD and 83y0 in t,he presence of 6.7 PM FAD; thus, at this NaI concentration, FAD provided a 55% stimulation of residual activity. These data suggest that I-and Br-may, in addition to other effects, cause some dissociation of FAD.
Salicylate hydroxylase was inhibited by cyanide (KCN) and azide (NaN3) at concentrations 10 mM or higher, and by urea: 50% inhibition was observed in 2 M urea and 86cc in 5 M urea. This inhibition has not yet been studied as a function of time.
The previously eited inhibition by excess salicylate may be related to the anionic nature of this compound.
When enzyme previously incubated for short periods of time with varying concentrations of NaCI, NaSCN, or NaI was diluted into a cuvette and assayed, inhibition of activity was reversed by the dilution.
In the presence of iodide and thiocyanate (but not chloride), prolonged (1 to 4 hours) incubations at 1 M concentrations caused 70 to 90% irreversible inactivation. E$ects of Mercurials-Enzyme previously incubated at 0" in 1 mM p-chloromercuriphenylsulfonate was slowly inactivated; this inactivation was first order with a half-time of about 36 min. The presence of 100 pM salicylate slowed the rate of inactivation 5-to B-fold, while 30 mM benzoate prolonged the half-time of inactivation 53%, over that of a control lacking aromatic substrate.
E$ects of Benzoate-In the course of studying the substrate or inhibitor specificity of salicylate hydroxylase with various organic reaction would not inhibit, but rather appeared to stimulate (by about 10%) salicylate hydroxylation activity measured at 340 nm in the standard assay. In the absence of salicylate, benzoate caused the rapid disappearance of DPNH.
A K, for benzoate of 2.0 mM was determined, as shown in Fig. 10. This K, is about 700 times that for salicylate.
In this experiment, DPNH concentration was that of the standard assay, 147 PM.
A separate Lineweaver-Burk determination of the K, for DPNH in the presence of 30 mM (15 times Km) benzoate was 164 pM, as compared to the 16.7 pM seen in the presence of salicylate and 710 pM in the absence of aromatic substrate.
Thus, benzoate decreases DPNH K,, but not as effectively as does salicylate. These effects of benzoate appear to present an additional distinction between the enzyme reported herein and that described by Yamamoto    An attempt to find a hydroxylated product of the benzoate reaction was made with the coupled assay described in Fig. 5, substituting 1 mrvr benzoate for the salicylate. No change in the benzoate spectrum could be seen after many hours, even when the concentrations of salicylate hydroxylase and alcohol dehydrogenase were increased IO-fold.
A large scale reaction was then attempted employing 1 InM concentrations of both benzoate and DPNH. Disappearance of DPNH (which ceased if the flask was not shaken) was followed by withdrawing aliquots of the mixture and assaying at 340 nm. When 98% of the DPNH had been oxidized, the reaction mixture was acidified with HCl, extracted repeatedly with ether, and the extracts were dried and evaporated.
The residue reconstituted in ethanol had a spectrum identical with that of benzoic acid. A control reaction mixture omitting salicylate hydroxylase was run simultaneously.
Ether extraction of this control produced benzoic acid in the same amounts as the experimental incubation mixture.
Benzoate thus appeared not to be changed in any way by the enzymatic reaction.
In an aerobic reaction mixture containing excess (300 pM) DPNH and 100 PM benzoate, all of the DPNH present was oxidized.
The rate was much faster than in the absence of benzoate. In a similar experiment, salicylate only permitted rapid oxidation of about 100 pM DPNH.
Thus, benzoate appeared to act catalytically whereas salicylate, as expected, served as a substrate.
The results of the three experiments described above indicated that, despite its effects on DPNH oxidation, benzoate was not itself hydroxylated.
The hypothesis was therefore entertained that benzoate mimicked salicylate by binding at the salicylate site, facilitating DPNH binding, enzyme reduction, and reaction with 02. But since benzoate could not itself be hydroxylated, the oxygen reduction could be "uncoupled" from hydroxylation. Benxoate as Competitive Inhibitor-If benzoate were bound at Reaction conditions were similar to those described in the standard assay, but with the addition of FAD to 6.7 pM. Reactions were initiated by addition of 1.5 pg of salicylate hydroxylase.
The units of the ordinate represent micromoles of salicylate disappearance per min; the observed absorbance change at 296 nm includes a molar extinction of 1300 due to DPNH oxidation as well as an extinction of 3406 for salicylate disappearance.
the salicylate site, it should serve as a competitive inhibitor for salicylate hydroxylation.
This was studied by conducting measurements at 296 nm, the salicylate peak, instead of 340 nm. Fig. 11 shows that benzoate was indeed a competitive inhibitor.
The Ki for benzoate was 3.1 mM, which compares to its K, for the stimulation of DPNH oxidation of 2 mM. The discrepancy between K, and Ki is not considered serious, since the DPNH concentration used was less than the K, value for DPNH in the presence of benzoate. The area of Fig. 11 near the ordinate was re-examined in an experiment at higher salicylate concentrations (up to 500 pM) ; the lines did indeed converge to a single Ti,,,.
Fate of Oxygen with Benzoate-The stimulation of DPNH oxidation by benzoate raised the question of the fate of the oxygen utilized.
Oxygen consumption was measured in the Gilson oxygraph as shown in Fig. 12 Addition of catalase initially to the benzoate-mediated reaction (Curve B) effected a total oxygen consumption of only half of that seen for salicylate.
Catalase had no effect upon oxygen uptake in the salicylate reaction mixture, whether added initially (Curve D) or after completion of the reaction (Curve C).
The data in Fig. 12 indicate that hydrogen peroxide is formed in the benzoate-mediated reaction, with a 1: 1 stoichiometry observed between DPNH and HzOz. In the salicylate reaction, catechol (Fig. 5) was formed, and the data of Fig. 12 indicate that the other product of oxygen reduction is water.
We have termed the benzoate-mediated DPNH oxidation an "uncoupling  Table IV. It can be seen that the V,,, for pyridine nucleotide disappearance is essentially identical with salicylate or benzoate, as well as TPNH or DPNH. The identity of V,,, found here suggests a common rate-limiting region for the reaction mechanism of the benzoate and salicylate reactions, despite the fact that salicylate is hydroxylated and benzoate is not. Stopped flow data, to be presented in the following paper (13), indicated that a combination of two steps rather than a single one was rate-limiting.
The Vmax for DPNH disappearance in the absence of aromatic substrate was only 3.50/, of the rate in the presence of benzoate or salicylate.
As previously stated, the apparent K, for DPNH is decreased 4.3-and 43-fold in the presence of benzoate and salicylate, respectively, as compared to this value in the absence of aromatic substrate.
The same rough proportions are seen with TPNH, which in all cases has a K, about an order of magnitude higher than that for DPNH.
Search for Free Radical Oxygen Intermediates-A series of experiments were designed to see if oxygen intermediates could be detected in the benzoate-mediated DPNH oxidation reaction. Such intermediates could conceivably have been enzyme-bound in the presence of salicylate, but liberated in analogous reactions with benzoate.
Both reducing and oxidizing radicals were sought, with methods utilizing both reduced and oxidized cytochrome c, superoxide dismutase, ethanol, mannitol, and the initiation of sulfite oxidation.
Superoxide (02.) is known to reduce ferricytochrome c (14). If 01. were produced in the benzoate reaction, or if it were an active intermediate in the hydroxylation of salicylate, then it might be detected by a superoxide dismutase-inhibitable reduction of cytochrome c in the former case, or in the latter, by superoxide dismutase inhibition of salicylate hydroxylation (14). All attempts to reduce cytochrome c by either salicylate or benzoate assay mixtures with salicylate hydroxylase failed, and superoxide dismutase had no effect upon salicylate-or benzoatemediated DPNH oxidation.
Thus if 02. were an active intermediate in the hydroxylation reaction, it was tightly bound and inaccessible to the dismutase.
No DPNH or TPNH cytochrome c osidoreductase activity could be observed even at high levels of salicylate hydroxylase.
Attempts were made to find oxidizing radicals, such as the hydroxyl radical (OH.) in assay mixtures. Oxidation of dithionite-reduced cytochrome c was used as a test. Enzyme-dependent oxidation of ferrocytochrome c in the presence of DPNH and benzoate was seen but this was so slow and small in extent as to be uninterpretable.
Mannitol, at 0.4 M concentration, did not interfere with the benzoate-or salicylatemediated oxidation of DPNH, nor did ethanol at concentrations up to 1.5 M. We do not feel that the results of these attempts to detect oxidizing radicals are as yet interpretable.
E$ect of Reaction Products--None of the reaction products, DPN+, catechol, or (in the case of benzoate) Hz02, inhibited salicylate hydroxylation or benzoate-mediated DPNH oxidation. HzOz, even at 1 M concentration, could not substitute for DPNH in the enzymatic hydroxylation of salicylate. Catalatic Activity of Salicylate Hydroxylase-The product of oxygen reduction in the benzoate-mediated DPNH oxidation was Hz02 (Fig. 12). The slow DPNH oxidase activity in the absence of substrate also produced HzOz, but in less than stoichiometric amounts. The lack of stoichiometry could be accounted for by a low catalatic activity of the purified enzyme, demonstrated in the oxygraph with 200 PM HzOz. It has not been established whether the low catalatic activity is inherent, and of possible mechanistic significance, or due to a trace impurity.
Substrates and Pseudosubstrates-A variety of substituted benzoates, salicylates, and aromatic compounds were examined to ascertain enzyme specificity and determine which compounds are substrates like salicylate or pseudosubstrates like benzoate. These compounds were first tested under assay conditions identical with that of salicylate.
The results are shown in Table V  in Table VI, with limiting concentrations of DPNH (200 PM), only part of the oxygen was reduced to water; the remainder was reduced to HtOz. This oxygen to Hz02 flux is designated as "percentage pseudosubstrate activity" and was ascertained as described in the legend for Table VI, by addition of catalase to the reaction mixture in the oxygraph (Fig. 12)  pound.
Correlations between structure and kinetic behavior cannot be clearly drawn for most compounds; as will be shown in the subsequent paper, the structure of the compound specifically determines the rate of flavin reduction, and this step need not be rate-limiting in catalysis.
No striking correlation can be extracted between structure and apparent K,. The data of