Prostaglandins Leukotrienes and Medicine 13: 139-142, 1984

STRUCTURE OF THE ACTIVE SITE OF PROSTAGLANDIN SYNTHASE FROM STUDIES OF DEPSIDES: AN ALTERNATE VIEW.

J.M. Gerrard* and D.A. Peterson** *Manitoba Institute of Cell Biology, 100 Olivia Street Winnipeg, Manitoba R3E OV9 **Clinical Pharmacology, Veterans Administration Hospital, Minneapolis, Minnesota 55417

ABSTRACT

Evaluation of the structure of the lichen depside, 4-O-methylcryptochloro- phaeic acid, the most potent inhibitor of prostaglandin synthesis known, and its potential interaction with heme supports a model of the active site of prostaglandin synthase initially suggested by studies of arachi- donic acid-heme interaction.

INTRODUCTION

Sankawa et al. have studied the depside, 4-0-methylcryptochlorophaeic acid (MCCPA), and a variety of structurally similar compounds in order to evaluate the active site of the prostaglandin synthase enzyme (1). They deduce that the active site is not relatively planar as envisioned by Sherrer (2) or by Gund or Shen (3), but that the arachidonic acid fits into the site in a curving three demensional pattern. Furthermore, Sankawa et al. showed that the carboxyl group of non-steroidal anti- inflammatory inhibitors interacts with an active site on the enzyme, not the site which binds the carboxyl group of arachidonic acid. Similarities between the model proposed by Sankawa et al. and an earlier concept which we put forward (4) led us to evaluate the potential interaction of MCCPA with ferrous heme, a well-recognized and important constituent of the prostaglandin synthase active site (5-8).

139

METHODS

Molecular models of MCCPA and heme were built using Framework Molecular Models (Prentise Hall, West Hyack, NY) and the potential fit of these two compounds was evaluated using these inodels..

RESVLTS AND DISCUSSION _--

Sankawa et al. proposed two active sites where oxygens on the depside MCCPA would interact with the active sites of prostaglandin synthase. Currently the active site of the prostaglandin synthase is presumed to be one of the ligands of the Fe*+ in the Fe*+-heme. The heme could exist either inside the depside (Figure 1) or possibly beside the depside (either above or below the plane of the page of the model as shown in Figure 1). We suspect the model in Figure 1 is the most reasonable interaction of the depside and heme and as can be noted by comparison with the model shown in the paper of Peterson and Gerrard (4) it is very similar to the interaction of arachidonic acid and heme proposed earlier. In this model the Fe of the heme interacts with the COO- of the depside.

The consideration of heme interaction with the sites on the depside proposed by Sankawa et al. allows a look at the proposal that there are two active sites. While the evidence presented by Sankawa et al. provides good support for two sites of interaction of inhibitors with electron donating groups, it does not necessarily follow that one site is the site of addition of the oxygen to form the endoperoxide while the other is the site of addition 'of the oxygen at C-15. Both these oxygen additions likely require the Fe2+-heme, and these two steps almost certainly occur at the same site as there is only one molecule of heme per molecule of prostaglandin synthase. (9,lO). This provides a strong argument against the concept of two active sites as proposed by Sankawa et al.

The two oxygen addition steps could easily both occur at site B of Sankawa et al., as suggested by the model in Figure 1 and the model shown previously (4). Site A of Sankawa et al. could be important for binding of the O-O of the endoperoxide once it has been formed to allow the addition of the oxygen at C-15. Thus, following addition of the oxygen across carbons 9-11 and cyclization, the molecule would have to swing around to position the C-15 in an appropriate place to add the second oxygen. Site A would be in an appropriate position to interact with the O-O of the endoperoxide while the second 02 is added to C-15. Thus, the present model and that proposed earlier (4) are more satisfactory than the model of Sankawa et al., without even considering the several reasons for initially proposing this model. ie. (it can predict four of the six stereospecific steps in the conversion of arachidonic acid to PGG2, it can account for the chain length of fatty acids which can be converted to prostaglandins, it can explain the rapid turnover of the enzyme and it is consistent with other known characteristics of this enzyme) (4,11,12).

140

FIGURE 1: Interaction of the depside MCCPA with the heme of prostaglandin endoperoxide synthase as suggested by the molecular models. The depside is drawn as shown in Sankawa et al. with the carbon (0) and oxygen (0) atoms shown. The heme with its Fe (black circle) is shown positioned in the "mouth" of the depside. For simplicity, the heme is shown in cross section with only one series of carbons visible. For clarity the substi- tuent groups on the heme are omitted from this diagram although they were included in the original model. of the COO- group on the depside.

Immediately above the Fe is an oxygen

1.

2.

3.

ACKNOWLEDGEMENT

We thank Lori Devlin for typing this manuscript.

REFEKENCES

Sankawa U, Shibuya M, Ebizuka Y, Noguchi H, Kinoshita T, Iitaka Y, Endo A, Kitahara N. Depside as potent inhibitor of prostaglandin biosynthesis a new active site model for fatty acid cyclooxygenase. Prostaglandins 24:21, 1982.

Sherrer KA. Introduction to the chemistry of anti-inflammatory and anti-arthritic agents. P29 in Anti-Inflammatory Agents: Chemistry and Pharmacology. KA Sherrer and MW Whitehouse, eds Academic Press, New York, New York, 1974.

Gund P, Shen TY. A model for the prostaglandin synthetase cyclooxygen- ation site and its inhibition by anti-inflammatory arylacetic acid. J. Med. Chem. 20:1146, 1977.

141

4.

5.

6.

7.

8.

9.

10.

11.

12.

Peterson DA, Gerrard JM. A hypothesis for the interaction of heme and arachidonic acid in the synthesis of prostaglandins. Med. Hypotheses 5:683, 1979.

Yoshimoto A, Ito H, Tomita K. Cofactor requirements of the enzyme synthesizing prostaglandins in bovine seminal vesicles. Journal of Biochemistry (Tokyo) 68:487, 1970.

Miyamoto T, Ogino N, Yamamoto S, Hayaishi 0. Purification of prosta- glandin endoperoxide synthetase from sheep vesicular gland microsomes. The Journal of Biological Chemistry 251:2629, 1976.

Van Der Ouderaa FJ, Buytenhek M, Nugteren DH, Van Dorp DA. Purifica- tion and characterization of prostaglandin endoperoxide synthetase from sheep vesicular glands. Biochimica et Biophysics Acta 487:315, 1977.

Ogino N. Ohki S, Yamamoto S, Hayaishi 0. Prostaglandin endoperoxide synthetase from bovine vesicular gland microsomes. The Journal of Biological Chemistry 253:5061, 1978.

Roth GJ, Machuga ET, Strittmayer P. The heme-binding properties of prostaglandin synthetase from sheep vesicular gland. J. Biol. Chem. 256:10018, 1981.

Ueno R, Shimizu T, Kondo K, Hayaishi 0. Activation mechanism of prost- aglandin endoperoxide synthetase by hemoproteins. J. Biol. Chem. 257:5584, 1981.

Peterson DA, Gerrard JM, Rao GHR, White JG. Salicylic acid inhibition of the irreversible effect of acetylsalicylic acid on prostaglandin synthetase may be due to competition for the enzyme cationic binding site. Prostaglandin Med. 6:161, 1981.

Peterson DA, Gerrard JM. Evidence that the peroxidase of the fatty acid cyclooxygenase acts via a Fenton type of reaction. Prostaglandin: Med.: in press.

142