Biochimica et Biophysics Acta, 574 (1979) 0 Elsevier/North-Holland Biomedical Press

l-7

BBA 57402

THE OXIDATION AND DECARBOXYLATION HORSERADISH PEROXIDASE

R.M. MCKENZIE

and E.C. NELSON

ACID BY

*

Department of Biochemistry, Agricultural Experiment Stillwater, OK 74074 (U.S.A.) (Received

OF RETINOIC

Station, Oklahoma State University,

March 8th, 1979)

Key words: Retinoic acid metabolism;

Vitamin A; Peroxidase; Heme catalysis

Summary The decarboxylation of retinoic acid by horseradish peroxidase was investigated. A marked increase in the yield of products was obtained. However, the data indicated the reaction was a nonenzymatic, heme catalyzed peroxidation. Previously reported requirements for phosphate, oxygen and ferrous ion were eliminated when hydrogen peroxide was provided. Peroxide also eliminated the EDTA and cyanide induced inhibition of the phosphate dependent system. In the presence of hydrogen peroxide, horseradish peroxidase was not essential to the reaction; heme equivalent amounts of hemoglobin decarboxylated retinoic acid with equal facility. However, hemoglobin was ineffective in the absence of hydrogen peroxide. Attainment of 50-60% decarboxylation represented complete utilization of the available retinoic acid. Thus the products of the reaction can be divided into two groups, products of retinoic acid oxidation and products of an oxidative decarboxylation of retinoic acid.

Introduction The decarboxylation of retinoic acid and the products of that reaction have been of interest since the discovery of a metabolically active decarboxylation product [ 11. The reaction has been studied using liver microsomes [ 2,3] horseradish peroxidase and crude liver powders [4] as catalysts. In all of these systems retinoic acid was decarboxylated to the extent of only about 50%. Using [15-14C]- and [10,11-3H]retinoic acid, in concert with thin-layer chromatography, the decarboxylation of retinoic acid and the formation of tri* To whom correspondence

and reprint

requests

should be addressed.

2

tiated decarboxylation products was studied [4-6]. These workers concluded that the residual 14C in the reaction mixture was unused retinoic acid. The same apparent product patterns were observed when retinoic acid was incubated with liver microsomes, crude liver powders and horseradish peroxidase [4,6]. This observation, combined with the apparently similar substrate requirements [4] and the lower contaminant levels afforded by the horseradish peroxidase system led to its use as a model system. This effort resulted in the isolation and tentative identification of an alcoholic decarboxylation product [ 51. Although contamination from solvents and/or other reagents had been reduced, product yield was low. Thus, it was imperative that the yield of products be increased. Materials and Methods Horseradish peroxidase (type VI RZ 3.1) and Pipes buffer were obtained from Sigma Chemical Co. (St. Louis, MO). All-trans-retinoic acid and all-trans[15-14C]retinoic acid (approx. 19 Ci/M) were gifts from Hoffmann-La Roche, Inc. (Nutley, NJ). Decarboxylation measurements were performed in 18 X 150 mm stoppered test tubes, essentially as described by Nelson et al. [4]. The stoppers supported a small glass rod to which a glass cup was attached. This cup contained the CO* trapping solution and wick which were present from the beginning of the incubation. The standard incubation contained (final concentrations) 0.83 mM FeCl,, 8.3 PM retinoic acid (lo-20 - 10’ cpm/assay added as a methanolic solution from a 60 pug/ml stock), 8.3 pug/ml horseradish peroxidase in 83 mM potassium phosphate buffer (pH 6.4), all in a final volume of 2.4 ml. The enzyme was dissolved in 50 mM KC1 and used to initiate the reaction. The incubations were agitated gently for 60 min at 37°C in the dark. The reaction was terminated by the introduction of 0.1 ml 0.2 M NaHCO, followed immediately by the addition of 0.2 ml 2 N HCl. The NaHCO, addition improved the precision of the decarboxylation values. Where indicated, hydrogen peroxide, at a nominal concentration of 1 mM, was substituted for FeCl,. Other reagents being common, these reaction mixtures are distinguished from each other by referring to them as either the iron or the peroxide-catalyzed assay. This notation is used only for the purpose of distinguishing between these two mixtures. Results and Discussion In view of the apparent incomplete utilization of retinoic acid, the first efforts were directed at increasing the extent of decarboxylation. Previous studies of the decarboxylation of retinoic acid by horseradish peroxidase indicated a requirement for Fe*‘, phosphate and molecular oxygen [4], However, at the initiation of this study, iron-free control incubations routinely decarboxylated retinoic acid to the same extent as the complete iron assay. Thus, before proceeding, the development of an adequate control was essential. The substitution of acid treated horseradish peroxidase for untreated peroxidase was attempted. However, contrary to a prior report [4] acid pretreatment of either aqueous or phosphate buffered solutions of the enzyme had no influ-

3 TABLE THE

I

EFFECT

OF EDTA

DECARBOXYLATION Acid

peroxidase

AND

ACID

PRETREATMENT

OF RETINOIC

was prepared

from

OF HORSERADISH

ACID IN THE IRON

:

a 1

1 mixture

.4ND PEROXIDE

of 0.8 mg/ml

horseradish

PEROXIDASE

ON THE

ASSAYS peroxidase

and 0.2 N HCl.

After 15 min at room temperature the pH was adjusted to PH 6.4 with 0.1 N KOH and the final protein concentration brought to 0.2 mg/ml. The same reagents were used to prepare the material for the peroxidase blank (essentially 50 mM KCl) and the peroxidase by substitution of water for the protein solution on the one hand and by changing the order of mixing the reagents on the other. The reaction mixture contained 20 nmol all-tmns-[ 1 5-14 C] retinoic acid, 20 pg peroxidase, acid treated peroxidase or an equivalent volume of blank solution. 200 pmol sodium phosphate (PH 6.4) and either 20 pmol FeCl2 or 22 Mmol hydrogen peroxide in a final volume of 2.4 ml. The reactions were initiated by the addition of peroxidase or its equivalent. 20 Fmol KCN or 50 pmol EDTA (approx. pH 7.0) were added where indicated. Incubations were performed in the dark, at 37’C. with continuous agitation for 60 min using sealed tubes, trapping reagents and termination procedures described under methods. Each experimental value presented in this paper represents the average of triplicate incubation. Catalyst

Substrates

(% decarboxylation)

Fe None

Fe + EDTA

1.2

0.6

H202

HZ02

1.9

1.4

Acid peroxidase

43.1

-

57.1

Peroxidase

39.3

1.3

61.7

+ EDTA

35.1

ence on its activity (Table I). An attempt was made to utilize inhibition by EDTA as a control. In addtion, the earlier observation that hydrogen peroxide eliminated the iron requirement [ 61 was investigated. The effectiveness of hydrogen peroxide as a substrate in the decarboxylation reaction was clearly demonstrated (Table I). The iron assay was inhibited by EDTA as observed previously [4]. However, the peroxidase assay seemed to be only partially inhibited. The data suggested that a non-enzymatic, heme-catalyzed peroxidation of retinoic acid might be involved. If true, some other heme source might catalyze the reaction. Thus, the substitution of hemoglobin for horseradish peroxidase was attempted. The effect of cyanide as an inhibitor was studied simultaneously. TABLE

II

COMPARISON OF HORSERADISH PEROXIDASE TION OF RETINOIC ACID AND ITS INHIBITION PEROXIDE ASSAYS

AND HEMOGLOBIN IN THE DECARBOXYLABY EDTA AND CYANIDE IN THE IRON AND

Bovine hemoglobin was used at a level of 40 pg/assay. Preparation of hemoglobin (0.4 mg/ml) was as described for peroxidase in Table I and it was used in place of peroxidase as indicated. All other assay reagents and conditions were as specified in Table I. Catalyst

Substrates

(% decarboxylation)

Fe

Fe + EDTA

Peroxidase

51.3

Acid peroxidase

51.3 3.3

Hemoglobin

Fe+CN

H202

Hz02

2.2

8.3

66.7

49.3

2.6

-

64.2

45.8

3.3

-

56.2

30.3

+ EDTA

Hz02 40.5

-

+CN

4

Hemoglobin was an effective horseradish peroxidase substitute in the peroxide assay and its response to inhibitors was also similar (Table II). At the concentrations used, cyanide was a less effective inhibitor of the iron assay than was EDTA. Again the inhibition of the peroxide assay by either EDTA or cyanide was only partial. The ineffectiveness of hemoglobin in the iron assay and its essential equivalence to horseradish peroxidase in the peroxide assay suggested that at least one additional reaction step was involved in the iron assay. The inability of cyanide and EDTA to prevent decarboxylation in the peroxide assay supported this conclusion. The characteristics of these two systems are further defined by the data presented in Table III. Substitution of Pipes buffer for phosphate confirmed the phosphate dependence of the iron assay and demonstrated the absence of this requirement in the peroxide assay. Heme equivalent amounts of hemoglobin were ineffective in the iron assay but were the equal of horseradish peroxidase in the peroxide assay. In addition, EDTA was noninhibitory in the Pipesbuffered peroxide assay. Thus, the partial inhibition of the phosphate buffered peroxide assay by EDTA (Tables I and II) appears to result from the elimination of a phosphate dependent contribution. Whatever the nature of this phosphate related phenomenon, it effects only the rate and not the final extent of reaction. In an experiment which utilized Pipes buffer, increased initial hydrogen peroxide concentrations were inhibitory. Decarboxylation in the standard, Pipes buffered, peroxide assay was observed to be 45.5% while a 16-fold increase in peroxide concentration produced only 27.9% decarboxylation. The decline in the extent of decarboxylation was linear with respect to the logarithm of the increasing peroxide concentration. The time courses of the iron and peroxide assays are shown in Fig. 1. The peroxide assay reached completion in 45 min. The iron reaction was considerably slower, approaching completion only after 90 min. It was apparent that neither the initial hydrogen peroxide concentration nor the incubation time was responsible for the inability to attain 100% decarboxylation. However, it was feasible that peroxide might become limiting during TABLE

III

COMPARISON ISH

OF

PEROXIDASE

THE

PHOSPHATE

CATALYZED

REQUIREMENTS

PEROXIDE

ASSAYS

Pipes-buffer

(piperazine-N,N’-bis(-Z-ethanesulfonic

an equimokr in Tables

I and

basis.

Preparation

of

catalysts

and

OF

acid),

PH 6.4,

all other

assay

THE

HEMOGLOBIN

RETINOIC

was substituted reagents

AND

ACID

and

IN

for

HORSERAD-

THE

IRON

phosphate

conditions

were

AND

buffer

on

as specified

II. Buffer

CatalYst

Substrate: Peroxidase Hemoglobin

OF

DECARBOXYLATION

(% decarboxylation)

pi

Pipes

Fe

Fl?

Fe + EDTA

H202

H202

56.6 2.8

5.0 3.3

0.7 -

39.0 37.1

41.7 _

+ EDTA

5

'0

20

40

60 80 MINUTES

100

120

) and hydrogen peroxide (0 -0) Fig. 1. The time course of the standard iron (aprocedures were as described in Materials and Methods and in Table I.

assays. AU

the course of the incubation. It was also possible that the heme catalyst was being exhausted in some manner. Since increases in the initial peroxide concentration were inhibitory, and the peroxide assay was almost linear for the initial 20 min (Fig. l), the concentrations of these reagents were varied by making serial additions of either or both to a set of standard incubations. Each addition was equivalent to the amount(s) of reagent initially present in the standard assay mixture. Additions were made at 20, or 20 and 40 min with all incubations proceeding for 60 min. The 2- and 3-fold increases in the amounts of horseradish peroxidase and/or hydrogen peroxide accomplished in this manner were ineffective. For example, a single (zerotime) addition of all reagents yielded 48.1% decarboxylation while double additions (zerotime and 20 min) of peroxide, peroxidase and both reagents together, yielded 46.2, 47.9 and 47.1%, respectively. Since neither the peroxidase nor the peroxide concentrations were limiting, the consistant inability to increase the extent of the decarboxylation suggested that the primary reaction involved a multisite oxidation of retinoic acid and that decarboxylation of this substrate was merely a side reaction resulting from generation of inherently unstable intermediates. The use of two CO, traps, one for the reaction period and one for the post termination collection period, showed that the bulk of the CO* was released during the reaction period. This at least established that the decarboxylation actually occurred during the reaction and not as a result of subsequent handling procedures. Accepting partial decarboxylation as a possible outcome of the completed reaction, the effect of increasing the retinoic acid content of the hydrogen peroxide assay was investigated. In these experiments, the level of carrier retinoic acid was varied while the [15-14C]retinoic acid was held constant. Since the higher levels of retinoic acid were saturating in the final reaction mixture, it is important to note that the labeled and unlabeled retinoic acid were added first so that the label would be uniformly distributed in the methanolic substrate solution before any precipitation occurred as a result of adding the aqueous assay reagents. Increased substrate concentration did not affect the percent decarboxylation until the quantity added was more than double that used in the standard incubations and then, though percent decarboxylation declined, the actual

6

pg RETINOICACID/ASSAY

HOURS

Fig, 2. The effect of changing retinoic acid concentrations on decarboxylation and product yields in the peroxide assay. The standard assay contained 6 pg retinoic acid. The collective yield of decarboxylation products is stated in terms of its retinoic acid equivalent weight. All other procedures were as described under Materials and Methods and Table I. Fig. 3. The time course of the decarboxylafion of retinoic acid in the hydrogen peroxide reaction at elevated substrate levels. Retinoic acid at 25 pg/assay (o----O and twice the normal ) and 50 Mg/assay ( A---a) peroxidase content (40 /.Wassay) were used. All other procedures were as described in Materials and Methods and in Table I.

yield of decarboxylation products increased (Fig. 2). It is apparent that when decarboxylation reached 50-60%, no substrate capable of releasing labeled CO* remained. Further, all that prevented attainment of this final level of decarboxylation in the more concentrated solutions (Fig. 2) was the peroxidase concentration and/or the incubation time. This is confirmed by the data in Fig. 3. Doubling the horseradish peroxidase level allowed an incubation containing 50 Mugsubstrate (8 times the normal level) to near completion in about 4 h. Simultaneously, 1 h incubations containing 50 Fg retinoic acid and 1, 2, 4, and 8 times the normal peroxidase content were found to yield 16.8, 27.1, 43.0, and 52.6% decarboxylation, respectively. Results similar to those observed with the peroxide assay were obtained using the iron assay (Fig. 4). The intent of these experiments was to increase product yield and this has been accomplished. A minimum 40-fold increase per unit volume is attainable. The relative increases are even more striking since the highly concentrated reaction mixtures require the use of markedly reduced quantities of the other

pg RETINOICACID/ASSAY Fig. 4. The influence of increasing retinoic acid concentrations on decarboxylation and product yields in a modified iron assay. In this experiment the ferrous chloride was doubled and horseradish peroxidase was present at a level of 0.4 mg/assay. The incubation time was increased to 2 h and the head space in the reaction tubes was purged with oxygen immediately prior to sealing the tubes. The collective yield of decarboxylation products is stated in terms of its retinoic acid equivalent weight. All other procedures were as described in Materials and Methods and in Table I.

7

reagents which leads to proportionate reductions in the contaminant load contributed by those reagents. The mechanism(s) of the oxidation and decarboxylation of retinoic acid effected by horseradish peroxidase and hemoglobin have not been conclusively elucidated by these studies. However, preliminary radiochemical analysis of the ether extractable products found in double labeled ([15-“%I- and [10,11-3H]retinoic acid) horseradish peroxidase catalyzed reaction mixtures demonstrated that among the several dozen reaction products, there were many 14C-labeled oxidation (i.e., non-decarboxylated) products. Some of these 14C-labeled compounds were subjected to preliminary mass spectral analysis and were found to have spectral characteristics in common with previously descrioed ketones [7] and epoxides [8,9]. Ultraviolet analysis indicated that some of these reaction products are identical to the previously described compounds while others are either geometric and/or structural isomers of same. These data support the suggestion that the mechanism of these heme catalyzed reactions involves multisite oxidations of retinoic acid, some of which lead to the concurrent and/or subsequent decarboxylation of this substrate. It must be noted that the radiochemical analysis of the reaction products also demonstrated the occurrence of significant exchange of the vinyl tritium label. The 3H/‘4C ratios indicated that the extent of this exchange could range from zero to in excess of 90% in any given product. Thus, with the use of such labile markers, it is feasible that a significant decarboxylation product might easily go unnoticed if radiolabeling were the sole criterion for its detection. Studies are in progress to further characterize the products of these reactions. The possibility that some of the products are identical with biologically derived compounds is not precluded by the apparently nonenzymatic character of the reaction. In addition, these procedures provide a readily available source of standard compounds needed for the development and control of the separations technology and handling procedures necessary for the successful study of compounds of this type. Acknowledgements Journal article 3490 of the Oklahoma Agricultural Experiment Station. This investigation was supported in part by Public Health Service Research Grant AM-09191 from the National Institute of Arthritis, Metabolism and Digestive Diseases. We gratefully acknowledge the technical assistance of W. Dobson and N.L. Rockley. References 1 2 3 4 5 6 7 8

Yagishita, K., Sundaresan, P.R. and Wolf. G. (1964) Nature 203. 410412 Roberts, A.B., and DeLuca. H.F. (1968) J. Lipid Res. 9, 501-508 Lin, R.L. (1969) Ph.D. Thesis, Oklahoma State University, Stillwater, OK Nelson. E.C., Mayberry, M., Reid, R. and John, K.V. (1971) Biochem. J. 121.731-733 McGregor, M.L.. Hopkins. K., Thayer. R.H. and Nelson. E.C. (1974) Fed. Proc. 33,688 Reid, R. (1972) MS. Thesis, Oklahoma State University. Stillwater, OK Rao, M.S.S.. John, J. and Cama. H.R. (1972) Intern. J. Vit. Nutr. Res. 42.368-378 Reid, R., Nelson. E.C., Mitchell, E.D.. McGregor, M.L., WaIIer. G.R. and John, K.V. (1973) 558-565 9 John, K.V., Lakshmanan, M.R. and Cama, H.R. (1967) Biochem. J. 103. 539-543

Lipids 8.