Preparation and properties of retinal-oxidizing enzyme from rat intestinal mucosa.

Abstract An enzyme which converts retinal to retinoic acid was purified from rat intestinal mucosa approximately 160-fold via a combination of acetone precipitation, ammonium sulfate precipitation, and DEAE-cellulose ion exchange chromatography. The purified preparation appeared to be homogeneous in the ultracentrifuge, on ion exchange chromatography, and on polyacrylamide gel electrophoresis. The molecular weight of the enzyme was approximately 80,000 and its isoelectric point was in the neighborhood of 6.7. The enzyme preparation contained approximately 2 moles of iron per mole of enzyme. The absorption spectrum of the enzyme indicated a maximum at 280 mµ and a much smaller band around 410 mµ. Purified retinal-oxidizing enzyme from rat intestinal mucosa stoichiometrically and irreversibly converted retinal to retinoic acid. The latter was identified as the product of the reaction by ultraviolet absorption spectrum, thin layer chromatography, and gas chromatography. The product was approximately 96% pure as determined by its E1%1 cm value. The enzymatic reaction exhibited Michaelis-Menten kinetics with a Km of approximately 3.0 x 10-4 m. The rate of the reaction was constant for 120 min, directly proportional to the enzyme concentration, and maximal when GSH, NAD, FAD, and Fe2+ were added. The pH optimum was approximately 7.7. The reaction proceeded well under both aerobic and anaerobic conditions. The enzyme acted upon both the all-trans and 13-cis forms of retinal. Reduced nicotinamide adenine dinucleotide noncompetitively inhibited the reaction and was found to exhibit a control over the utilization of retinal in vitro. Thiols and metal ions stimulated the reaction, while thiol inhibitors and chelators inhibited the reaction. It is suggested that retinal-oxidizing enzyme is a metalloprotein which requires sulfhydryl groups for maximal enzymic activity. In the presence of H218O, the oxidation of retinal by retinaloxidizing enzyme appeared to resemble a dehydrogenase rather than an oxygenase or oxidase.

The purified preparation appeared to be homogeneous in the ultracentrifuge, on ion exchange chromatography, and on polyacrylamide gel electrophoresis. The molecular weight of the enzyme was approximately 80,000 and its isoelectric point was in the neighborhood of 6.7. The enzyme preparation contained approximately 2 moles of iron per mole of enzyme.
The absorption spectrum of the enzyme indicated a maximum at 280 rnp and a much smaller band around 410 ml.
Purified retinal-oxidizing enzyme from rat intestinal mucosa stoichiometrically and irreversibly converted retinal to retinoic acid. The latter was identified as the product of the reaction by ultraviolet absorption spectrum, thin layer chromatography, and gas chromatography. The product was approximately 96% pure as determined by its EyO, value.
The enzymatic reaction exhibited Michaelis-Menten kinetics with a K, of approximately 3.0 X low4 M. The rate of the reaction was constant for 120 min, directly proportional to the enzyme concentration, and maximal when GSH, NAD, FAD, and Fez+ were added.
The pH optimum was approximately 7.7. The reaction proceeded well under both aerobic and anaerobic conditions. The enzyme acted upon both the all-frans and 13-cis forms of retinal.
Reduced nicotinamide adenine dinucleotide noncompetitively inhibited the reaction and was found to exhibit a control over the utilization of retinal in vitro. Thiols  In the presence of HzlSO, the oxidation of retinal by retinaloxidizing enzyme appeared to resemble a dehydrogenase rather than an oxygenase or oxidase.
The discovery of Arens and van Dorp (l), and others (2, 3), of the growth-promoting activity of retinoic acid in retinal-deficient rats has suggested the involvement of the acid in the pathway of retinol metabolism.
Retinal, a product of the enzymatic oxidation of retinol, has been oxidized in vitro to retinoic acid by liver aldehyde dehydrogenase, liver aldehyde oxidase, and milk xanthine oxidase (4). Recently, retinoic acid has been detected in the blood, liver, intestine, and other tissues of the rat after administration of retinal (5, 6). The formation of retinoic acid from retinol (7), retinal (5,8), and p-carotene (8) has been shown in several laboratories.
On the other hand, other investigators (9) have indicated that under normal conditions the major pathway of retinal absorption involves its reduction to retinol, while a small amount of retinal is oxidized to the acid. Thus, the role of retinoic acid as an obligatory metabolite of retinol has not yet been established.
An enzyme responsible for the oxidation of retinal to retinoic acid has been partially purified from rat liver via ammonium sulfate precipitation (10). Recently, Crain,Lotspeich,and Krause (8) have shown the enzymatic conversion of retinal to retinoic acid by an enzyme system from rat intestinal mucosa. This report presents a method for isolating the retinal-oxidizing enzyme from rat intestine and describes some of its properties. No. 2 Sober (11). Sephadex G-200 was obtained from Pharmacia, and retinoic acid were recoverable. The procedure was validated H&*0 was purchased from Bio-Rad.
All other chemicals were by isolating and identifying the product and by comparing the obtained from commercial sources and were of the highest amount of product formed to the amount of substrate utilized. available purity.
(The data are presented under "Results.") Preparation of Reductase-A crude preparation from rat intestine, termed 45 to 70% (NH&SO4 fraction, was used as a source of retinal reductase. This enzyme fraction was prepared by essentially the same procedure as described by Mahadevan, Murthy, and Ganguly (10) for a similar enzyme in liver. This fraction possessed an activity of 1.15 units/O.2 ml (0.9 mg of protein).
Preparation of Crude Enzyme Extract-The method of preparing the soluble fraction from rat intestinal mucosa was a modification of the procedure described by Crain et al. (8). Male albino rats of the Wistar strain, fasted overnight and weighing 200 to 250 g., were lightly anesthesized with ether and killed by direct cardiac puncture.
The proximal half of the small intestine was removed, flushed out with two lo-ml portions of 0.9% NaCl solution, and cut lengthwise.
The mucosa was scraped off and homogenized in 0.1 M sodium phosphate buffer, pH 7.7 (8 volumes per g of mucosa), in a Dounce homogenizer. The homogenate was centrifuged at 100,000 X g (40,000 rpm) for 30 min in a Spinco model L ultracentrifuge.
The supernatant was filtered through cheesecloth, collected, and stored frozen until used. This fraction was termed the soluble fraction and constituted the crude enzyme preparation.
Determination of p-carotene oxygenase activity was similar to the spectrophotometric method of Brackkan,Myklestad,and Njaa (13). Enzymatic activity was expressed as micromoles of retinal produced per 2 hours per ml of enzyme preparation.
Xanthine Oxidase Assay-Reaction mixtures consisted of 0.01 ml of 4.5 X 10F2 M xanthine solution, 2.70 ml of 0.1 M sodium phosphate buffer (pH 7.7), and 0.30 ml of enzyme preparation.
The phosphate buffer and enzyme were pipetted into two l-cm quartz cells. To the first cell (blank) was added 0.01 ml of 0.05 M NaOH.
The reaction was initiated by placing 0.01 ml of xanthine solution into the second cell. Absorbance readings were taken in a Beckman DU spectrophotometer at 290 rnp at I-min intervals.
Activity was expressed as the increase in absorbance at 290 rnp per min per ml of enzyme preparation.

RO-Enzyme1
Assay-A modification of Futterman's method (12) was used for the determination of RO-enzyme activity. A 4.00-ml reaction mixture, consisting of 0.05 ml of retinal dispersion (2.5 pmoles), 2.45 ml of 0.1 M sodium phosphate buffer (pH 7.7), 0.50 ml of the enzyme preparation, and either 1.00 ml of a solution containing 2 pmoles each of NAD+, FAD, and FeClz or 1.00 ml of distilled water was incubated at 37" in a Dubnoff shaker for 2 hours under OZ. Aliquots of 0.5 ml were withdrawn at 0 and 2 hours and extracted five times with 2 ml of an ether-ethanol mixture (9: 1, v/v). The extracts were then pooled, and the ether was removed under a stream of nitrogen. Retinal and retinoic acid were separated by chromatography on DEAE-cellulose in the following manner. A column (1 X 4 cm) was prepared from a slurry of DEAE-cellulose in ethanol, converted to the hydroxyl form by treatment with 10 ml of ammoniacal ethanol (ethanol, 240 ml; concentrated ammonia, 10 ml), and washed with 10 ml of ethanol.
The pooled extracts of retinal and retinoic acid were added to the column and eluted with ethanol.
Retinal was recovered in the first 10 ml of the effluent and quantitatively determined by comparison of its absorbance at 380 rnE.1, in a Beckman DU spectrophotometer, to a standard synthetic sample. The retinoic acid retained by the column was subsequently eluted with 10 ml of acidified ethanol (ethanol, 300 ml, concentrated HCl, 1.5 ml). The acid was quantitatively determined by comparing the absorbance of the isolated compound at 350 rnp to that of a synthetic sample. RO-Enzyme activity was defined as the micromoles of retinoic acid produced per 2 hours per ml of enzyme preparation (0.68 mg of protein).
Reductase Assay-The enzymatic reduction of retinal to retinol was followed spectrophotometrically in a similar manner as described for the RO-enzyme assay, except that the amount of retinol in the isolated products was measured by comparing the increase in optical density at 325 rnp to a standard sample of retinol.
Other Procedures-Protein concentration of enzyme samples was determined by the method of Lowry et al. (14), non-heme iron content according to Ramsay (15), and carbohydrate by the phenol-sulfuric acid method (16). Enzyme samples were dialyzed, with agitation, in Cellophane tubings twice against 50 volumes of distilled water at 04" for periods of 1.5 hours.
Polyacrylamide disc electrophoresis was performed on 7.5% gels according to Davis (17).
Electrophoresis was carried out at pH 9.5 for 60 min at 2 to 5 ma per tube. Tracking dye was 0.005% bromophenol blue. Proteins were stained with 0.5% Buffalo blue black. Destaining was carried out for 45 min in 7% acetic acid at 12.5 ma per tube.
Analytical ultracentrifugation was carried out in a Spinco model E ultracentrifuge with a 12-mm sector cell and an A-ND head. Centrifugations were performed at 40,000 rpm and 20". Thin layer chromatography was performed on glass plates The assay procedure was standardized by extracting and chromatographing known amounts of commercial retinal and retinoic acid. Results indicated that the method was reproducible, and that approximately 97% of retinal and 93% of (20 x 20 cm) previously coated with Silica Gel G (E. Merck AG, Darmstadt, Germany). Standards of retinoic acid, retinal, retinol, and methyl retinoate, along with unknown reaction products, were spotted and developed via ascending chromatography in a solvent system consisting of chloroform-methanolwater (65:20:2).
Lipid spots were detected with 1~ vapor. Methylated retinoic acid was subjected to gas chromatog-1 The abbreviation used is: RO-enzyme, retinal-oxidizing enzyme.
raphy with a Research Specialties Company model 600 series gas chromatograph, equipped with an Hz flame detector, Gas-Chrom P support, and a 20% diethyleneglycolsuccinate column measuring 5 feet x 0.25 inch. The temperature of the column was 130", and the flow rate was 32 ml per min. Retinoic acid was methylated by diazomethane. The latter compound was prepared according to Fieser and Fieser (18).
Anaerobic studies on the enzymatic reaction were conducted by incubating the reaction mixture in a Thunberg vessel with oxygen removed under vacuum.
Oxygen content of reaction vessels was measured polarographically with the use of a vibrating Teflon-covered platinum electrode patterned after Kahn (19). The electrode response characteristics were evaluated with known concentrations of ferricyanide ion.
Isotope experiments in Hz180 were carried out under aerobic conditions.
The oxidation, which was allowed to proceed for 3 hours, resulted in approximately 70% conversion to retinoic acid. The product was extracted into diethyl ether, evaporated to dryness under Nz, and oxidized to COZ with Hg(CN)2 (20). The latter procedure minimizes oxygen exchanges which might occur during the oxidation.
Carbon dioxide was collected in an apparatus described by Lee (20) and analyzed for l*O content by determining mass 46:44 ratio with a mass spectrometer.

PuriIcation of Retinal-oxidizing Enzyme
Acetone Fractionation-The soluble fraction was cooled with stirring in a -4" alcohol bath.
Acetone, chilled to -5O", was added slowly below the liquid surface of the soluble fraction to bring the suspension to an acetone concentration of 50%. The suspension remained at -4" for an additional 10 min, and then was centrifuged at -3" for 10 min at 7500 rpm (6800 x g). The supernatant was discarded and the precipitate was immediately dried over a stream of Ns to remove traces of acetone. The sample was then dissolved in 0.1 M sodium phosphate buffer, pH 7.7. This preparation is termed the 50% acetone fraction.
Dialysis-The 50% (NH&S04 fraction was dialyzed twice against 0.001 M phosphate buffer, pH 7.7, for a total of 3 hours in order to prepare the enzyme fraction for a DEAE-cellulose ion exchange separation.
Longer periods of dialysis against buffer resulted in a considerable loss of enzymatic activity.
The effluent was collected in 5-ml fractions with a Gilson fraction collector, scanned for protein via absorbance at 280 rnp in a Beckman DU spectrophotometer, and assayed for RO-enzyme activity.
The results of a typical separation of the 50% saturated fraction on DEAE-cellulose are shown in Fig. 1A. The conditions used for the column separation are described in the legend accompanying the figure.
The RO-enzyme peak area (tubes 20 to 28) was labeled DEAEr fraction and stored in a frozen state. This enzyme fraction was quite unstable2 dur-2 Retinal-oxidizing enzyme was somewhat labile during purification procedures.
Our results suggest that the purification steps should be carried out as rapidly as possible and the enzyme should be used immediately after preparation (unpublished data). ing extensive freezing and thawings; however, very little activity was lost when samples were frozen and thawed only once or twice. Thus, whenever the enzyme preparation was used for assaying purposes, the samples were frozen and thawed only twice.
Ammonium sulfate Concentration and Dialysis-The DEAEr fraction was concentrated by the addition of ammonium sulfate (90 to 95% saturation) and centrifuged for 30 min at 7500 rpm (9000 x g), and the resulting precipitate was dissolved in a small volume of 0.001 M sodium phosphate buffer, pH 7.7. The sample was then prepared for a second DEAE-cellulose separation by dialyzing against 0.001 M phosphate buffer, pH 7.7.
Second DEAE-cellulose Chromatography-The dialyzed, concentrated DEAEr fraction was chromatographed on DEAE-cellulose under the same conditions as the first DEAE-cellulose column except that elution was carried out with a linear pH gradient (50 ml of 0.01 M sodium phosphate buffer, pH 7.7, and 50 ml of 0.01 M sodium phosphate buffer, 4.0). The results of a typical separation are shown in Fig. 1B. The fraction containing RO-enzyme activity (tubes 9 to 15) was termed DEAErr.
Purity of Enzyme-The relative homogeneity of the DEAlSI fraction was tested by ultracentrifugation, ion exchange chromatography, and polyacrylamide gel electrophoresis. Table I illustrates a summary of the specific activities, degree of pur& cation, and yields of the various enzyme fractions.
The final In each case, a single protein peak was obtained which was closely superimposable on a single activity peak. The enzyme exhibited a homogeneous sedimentation pattern in the ultracentrifuge and possessed an sZo+, of 3.6. Molecular Weight-The molecular weight of the retinal-oxidizing enzyme was estimated by thin layer and column chromatography on Sephadex G-200 according to Andrews (22). Fig.  2B shows a plot of the log molecular weight against V,, the elution volume, and Fig. 2C is a plot of the log molecular weight against the distance of migration.
The molecular weight was approximately 75,000 to 80,000. Absorption Spectrum-The absorption spectrum of the purified enzyme preparation, measured in a Beckman DU spectrophotometer, is presented in Fig. 3. The preparation exhibited a L,, at 280 rnp and an Azso:A260 ratio of 1.35. Lipid, Carbohydrate, and Iron Analyses-Lipid material contained in the various enzyme fractions was quantitatively estimated by extracting various enzyme fractions with ether-ethanol (9 : 1) and determining the amount of ether-extractable solid remaining after evaporation of the extracts under Ns. The results indicated that ether-extractable material ( <0.025% w/v) was present only in the soluble and 50% acetone fractions, while the 50% (NH&Sob, DEAEr, and DEAErr fractions were devoid of any ether-extractable material. Carbohydrate analysis via the phenol-sulfuric acid method (16) indicated that the soluble, 50% acetone, and 50% (NH&S04 fractions contained approximately 154, 73, and 19 pg per ml of sugars, respectively.
The DEAEr and DEAErr fractions did not contain any detectable carbohydrates.
Non-heme iron, estimated by the dipyridyl method (15), was present in all enzyme fractions.
The purified fraction (DEAErr) contained about 0.153% iron by weight.
This value corresponds to a iMmin of approximately 36,500, and, assuming that the enzyme possesses a molecular weight of 80,000 (Fig. 2, B  ments for maximal RO-enzyme activity are presented in Table II. Maximum rate of retinal oxidation occurred when GSH, FAD, NAD+, and Fe2+ were added to the reaction mixture. Ioknti$cation of Reaction Product-Several criteria were used to identify retinoic acid as the oxidation product of retinal. The isolated product and synthetic retinoic acid had similar absorption spectra with a X,,, at 350 rnp. The E\t,,, value at 350 rnp in ethanol for the isolated product was 1409. This represents approximately 96% pure retinoic acid.     pmoles/2 hours per reaction mixture (4.68 x 10e8 M per set).

Effects of Time and Enzyme
Concentrations- Fig.  4 illustrates Effect of pH-The effect of pH on enzymic activity was the effect of time and enzyme concentration on the conversion studied in varying pH values of sodium phosphate buffer (0.1 M). of retinal to retinoic acid. In these experiments, the reaction The pH curve was rather broad with an optimum at approximixture (8.00 ml) consisted of 0.10 ml of retinal dispersion mately pH 7.7. (5.0 pmoles), 6.90 ml of 0.1 M phosphate buffer (pH 7.7), and E$ect of Dialysis-Dialysis of enzyme preparations against 1.00 ml of the enzyme preparation.
The rate of the reaction distilled water for several hours resulted in approximately 29% was constant for 120 min and proportional to the concentra-loss of enzymic activity. This loss of activity was restored by tion of enzyme.
the addition of 0.5 pmole of each of the following to the reaction E$ect of Substrate-Enzyme activity was determined at vessels: GSH, FeC12, FAD, and NAD+.  Efect of Other Reagents-Several reagents were used to affect the rate of enzymatic oxidation of retinal. The results are presented in Table IV. Reagents were added in l.O-ml volume to the reaction mixture and the RO-enzyme activity was determined immediately after addition. Reduced glutathione stimulated and p-chloromercuribenzoate inhibited the enzymatic reaction.
This suggests that the enzyme requires sulfhydryl groups for activity.
Likewise, metal ions, especially ferrous ions, appear to be required by the enzyme since the rate of the reaction was inhibited by the chelator, a,ar'-dipyridyl, and stimu- lated by the addition of ferrous ions. Atabrine slightly inhibited the enzymic reaction and FAD reversed the inhibition.
Substrate Specijkity of RO-Enzyme-The ability of RO-enzyme to oxidize various compounds was tested and the results are presented in Table V. All-trans retinal was more actively oxidized than the 13-&s form.
Purified RO-enzyme did not attack p-carotene, all-truns retinol, all-trans retinoic acid, and xanthine.
On the other hand, an impure sample of RO-enzyme, e.g. the 50% (NH&SO4 fraction, did oxidize p-carotene and xanthine.
Anaerobic Oxidation of Retinal-RO-Enzyme activity was determined under anaerobic conditions. Samples were incubated at 37", with agitation, in Thunberg tubes from which the oxygen was previously evacuated.
Oxygen content was monitored polarographically.
The RO-enzyme activity of a typical anaerobic sample is included in Table II. Polarographic analyses of the anaerobic mixture showed that the oxygen content was negligible.
(Aerobic incubations yielded a Z&LA response while anaerobic incubations showed a response of less than 0.3 PA.) Enzymatic Incorporation of I80 from H2180-The enzymatic oxidation of retinal was performed in a H.$gO medium as described under "Experimental Procedure." Incorporation of l*O into retinoic acid from Hz180 was confirmed by extracting the product into diethyl ether, oxidizing to COz, and determining the mass 46:44 ratio with a mass spectrometer.3 The main peaks observed upon mass analyses were COZ and HCN, the latter resulting from the Hg(CN)2 treatment.
This value corresponds to 1.21 atom % 180. Since the natural abundance of 180 in CO2 is approximately 0.2yo, then the corrected value for the 180 content is 1.0 atom y. excess. 6. Effect of NADH on the reaction velocity at different substrate concentrations.
vO/v< represents the ratio of the uninhibited rate to the inhibited rate. l , first experiment with a substrate concentration of 6.25 X lo2 PM. A, second experiment with a substrate concentration of 3.12 X 102 pM. a Enzyme and 0.5 pmole of cofactor were previously incubated for 10 min before treatment.
b Enzyme samples were five times dialyzed against 50 volumes of distilled water for 1.5-hr intervals at O-5". Inhibition by NADH-Since it was observed that NADH* inhibited the enzymatic oxidation of retinal (Table II), the inhibition was studied at various NADH concentrations and at two substrate concentrations.
Data are plotted in Fig. 6 in a manner described by Johnson,Eyring,and Williams (24), i.e. a plot of Ve/Vi against NADH concentration.
The results indicated that alterations in substrate concentration did not significantly affect the inhibition by NADH.
This finding suggests that the inhibition is noncompetitive. 4 p-Diphosphopyridine nucleotide (reduced form), purchased from Sigma, is known to form "inhibitors" when stock solutions are kept for long periods of time, frozen or otherwise.
Inhibitor formation can occur without a decrease in optical density at 340 mp (23). In order to minimize this problem, all solutions were prepared in phosphate buffer immediately before use. a The amount of retinol produced was measured spectrophotometrically.

E$ect of NADf and NADH on Enzymatic Oxidation
and Reduction of Retinal-The results of direct additions of NAD+ and NADH to the reaction mixture and preliminary incubations with the enzyme preparation are presented in Table VI. The effect of dialysis on the previously incubated samples is also shown.
The data suggest that the inhibition by NADH was not reversed by the direct addition of NAD+, instead, the removal of the former via dialysis was necessary to effect a reversal by NAD+.
On the other hand, the stimulation by NAD+ was reversed by the direct addition of NADH.
In addition, an equal molar mixture of NAD+ and NADH resulted in depression of the enzymatic activity.
When a source of retinal reductase activity is added to the ROenzyme preparation without adding either NAD or NADH, the major product formed is retinoic acid (Table VII).
Maximum production of retinol occurred with the addition of NADH or an equal molar combination of NAD+ and NADH, while minimum production of retinol occurred in the presence of NAD+.

DISCUSSION
Retinal-oxidizing enzyme can be purified from rat intestine via a combination of acetone treatment, ammonium sulfate precipitation, and DEAE-cellulose chromatography.
The enzyme activity is eluted from DEAE-cellulose at an NaCl concentration of approximately 0.25 to 0.30 M and at a pH of 6.8 to 6.6. Assuming that the elution via the second DEAE-cellulose separation (Fig. 1B) is the result of pH changes, then we can predict that the isoelectric point of the enzyme is approximately 6.6 to 6.8. The purified enzyme fraction is a soluble protein with a molecular weight of 75,000 to 80,000 and contains 2 moles of non-heme iron per mole of enzyme.
The enzyme appeared homogeneous on the basis of ultracentrifugation, ion exchange chromatography, and disc electrophoresis. Attempts to elute the enzyme from the polyacrylamide gels were unsuccessful.
Thus, the single protein band observed in electrophoresis was not conclusively identified as RO-enzyme.
It is possible that the enzyme was inactivated during electrophoresis since the purified enzyme appears somewhat labile during freezings and thawings.
The absorption spectrum of the purified enzyme indicated that the X,,, was approximately 280 rnp and that an additional absorbance occurred in the visible region (410 mp). This latter finding suggests that other light-absorbing materials are present in the purified preparation.
In addition, the preparation contains a very slight trace of yellowish color. The &go:&0 ratio was relatively low (1.35) and could indicate the presence of some molecule such as TPN or DPN in the enzyme preparation.
Investigations into the kinetics of the enzymic reaction indicate that the conversion of retinal to retinoic acid is stoichiometric and exhibits Michaelis-Menten kinetics.
The enzyme possesses a fairly strong affinity for retinal as indicated by its K, value.6 The reaction rate is constant over a relatively long period of time (129 min) and is proportional to the enzyme concentration. The optimum pH for the enzymic reaction is approximately 7.7.
Retinoic acid was identified as the reaction product by several criteria.
First, it exhibited an absorption spectrum identical with the spectrum of commercial retinoic acid. The EiY& value at 350 rnp for the product indicated that it was 96% pure retinoic acid. Second, on thin layer chromatography the product and its methylated derivative had the same RF values as their corresponding standards.
Finally, the methylated derivative and methyl retinoate exhibited the same pattern and retention time on gas chromatography.
Although undialyzed preparations of RO-enzyme, without external additives, can catalyze the oxidation of retinal, significant increases in the reactions occur when such cofactors as NAD, GSH, FAD, and Fez+ are added to the preparation (Table II). Maximum activity occurs when all cofactors are present.
In addition, the same cofactors can restore the RO-enzyme activity lost during dialysis.
Therefore, it appears that the enzyme requires several cofactors that act in concert to bring about the oxidation of retinal.
This statement is also supported by our observation that RO-enzyme preparation exhibit a low AB~:AQ+x ratio and a small absorption peak at 410 rnp, thereby suggesting that other nonprotein, light-absorbing substances are associated with the purified enzyme preparation.
It is probable that these substances may be cofactors which are strongly bound to the enzyme since they were not removed during purification.
Although we have not been successful in showing the oxidant of the reaction, oxygen does not appear to be a likely candidate since the conversion of retinal to retinoic acid takes place under anaerobic conditions (Table II).
Apparently, various electron carriers are involved in the oxidation process; however, the terminal acceptor for these electrons has not yet been identified.
Liver retinal oxidase (10) and several other enzymes (26) involved in oxidation reactions have been shown to be flavoproteins which require metal ions and sulfhydryl groups for activity.
Evidence reported in this paper suggests that RO-enzyme may require certain metal ions, particularly ferrous ions, inasmuch as the ferrous ion stimulates RO-enzyme activity and the metal chelator a,&-dipyridyl inhibits the reaction (Table IV). In addition, . RO-enzyme preparations contain 2 moles of iron per mole of enzyme. Cyanide, which inhibits several metalloflavoenzymes (27), also inhibits RO-enzyme activity (Table IV).
Ferric ions, 5 The authors must emphasize that the K,,, value for ROE may merely represent an apparent value since retinal is water-insoluble and dispersed in the incubation mixture from an acetone solution cw.
on the other hand, inhibit RO-enzyme activity.
The precise relationship between ferrous, ferric ions, FAD, and RO-enzyme activity is unknown.
Retinal-oxidizing enzyme may also require intact sulfhydryl groups for activity.
Sulfhydryl reagents, such as GSH and ascorbate, stimulate RO-enzyme activity while sulfhydryl inhibitors, such as 2-iodoacetamide and p-chloromercuribenzoate, inhibit the reaction.
Data from anaerobic incubations, HZ'*0 incubations, and polarographic analysis of the reaction mixture suggest that the enzymatic reaction does not use atmospheric oxygen; instead, it utilizes the oxygen from water.
The amount of dissolved oxygen in 4.0 ml of water at standard pressure is approximately 0.97 pmoles (28). Under anaerobic conditions, the 4.0.ml RO-enzyme reaction mixture would contain considerably less oxygen.
Polarographic analysis of the reaction mixture under aerobic and anaerobic conditions showed a decrease in microamperes from 2.0 to 0.3 PA, respectively. This decrease would correspond to a significant decline in oxygen concentration.
Consequently, one would not expect to find an adequate amount of oxygen available to effect the conversion of 1 to 2 pmoles of retinal to retinoic acid. Furthermore, enzymatic oxidation of retinal in the presence of H&*0, under aerobic conditions, resulted in a significant incorporation of 180 into retinoic acid. Hence, the reaction utilizes the oxygen of water, thereby resembling a dehydrogenase rather than an oxidase or oxygenase reaction.
Retinal-oxidizing enzyme preparations oxidize both the alltrans and 13-cis forms of retinal; however, the former compound was the better substrate (Table V). The enzyme would not convert retinoic acid into retinal in the presence or absence of NADH, thus indicating that the RO-enzyme reaction is irreversible. In addition, retinal reductase, /?-carotene oxygenase, and xanthine oxidase activities were absent in the RO-enzyme preparation. Previous reports differ on the effect of NAD+ and NADH on RO-enzyme activity. In this present study, we find that NAD+ increases RO-enzyme activity while NADH depresses the rate of this reaction.
Our data indicate that NADH acts as a noncompetitive inhibitor of this enzyme (Fig. 6). We found no indication that NADH interfered with the assay method6 or that NADH had been oxidized to NAD+ by the enzyme preparation.
(The latter statement stems from the observation that the absorbance of a mixture of enzyme and NADH at 340 rnp did not change during 2 hours of incubation.) In experiments in which one cofactor was previously incubated with enzyme (Table VI), and the other cofactor was added later, the results indicate that NADH can reverse the effect of NAD+, while the opposite relationship does not hold true. Both effects can be partially reversed by dialysis.
Since it has been shown by others (25) that retinal reductase from rat intestine requires NADH for the conversion of retinal to retinol, an experiment was carried out to study the effect of NAD+ and NADH on a solution containing both the reductase and the retinal-oxidizing enzyme. The results presented in 6 The effect of NADH on the substrate (retinal) and product (retinoic acid) was tested. Incubations of known amounts of either substrate or product or both with NADH did not alter the efficiency of the assay technique.