THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1987 by The American Society of Biological Chemists, Ine.

Vol. 262, No. 18, Issue of June 25, pp. 6840-8844. 1987 Printed in U.S.A.

The Active Site of Aromatase Cytochrome P-450 DIFFERENTIAL EFFECTS OF CYANIDE PROVIDE EVIDENCE FOR PROXIMITY OF HEME-IRON AND CARBON-19 IN THE ENZYME-SUBSTRATE COMPLEX*

(Received for publication, December 12, 1986)

James T. Kellis, Jr.S and Larry E. Vickeryt From the Department of Physiology and Biophysics, University of California, Imine, California 92717

19-Norandrostenedione and androstenedione are shown to be metabolized by purified, reconstituted hu- man placental aromatase cytochrome P-450. Kinetic evidence indicates that both steroids share a common catalytic site: 19-norandrostenedione is a competitive inhibitor of androstenedione aromatization, and the Ki value for its inhibition (120 nM) is similar to the K,,, value for its metabolism (132 nM). The two substrates differ, however, in their sensitivity to inhibition by the heme-iron ligand cyanide; 19-norandrostenedione is approximately %fold more sensitive to cyanide inhi- bition. Spectroscopic studies show that this differential inhibition by cyanide occurs because androstenedione competes with cyanide, whereas 19-norandrostene- dione promotes cyanide binding to the heme-iron. It is proposed that these opposite effects on cyanide-iron coordination are due to the proximity of the heme-iron and (2-19 of androstenedione in the enzyme-substrate complex, which results in steric exclusion of cyanide from the active site by the C-19 methyl group of an- drostenedione. Dioxygen is not excluded from binding to the heme-iron during catalysis, presumably because it bonds at an angle, in contrast to the linear bond of iron-cyanide complexes. A model for the active site of aromatase cytochrome P-450 is presented.

Aromatase cytochrome P-450 (P-450arom)’ catalyzes the aromatization of androgens to form estrogens. This enzyme plays a key role in endocrine physiology and several estrogen- dependent diseases, and it has been most extensively studied in human placental microsomes. The aromatization reaction is believed to involve three sequential dioxygen- and NADPH- dependent oxygenations, as depicted in Scheme 1 (1-3). Cur- rent evidence suggests that a single enzyme species carries out all three steps of the reaction and that androgens differing in D-ring substitution (androstenedione, testosterone, and 1601-hydroxytestosterone) can each serve as substrate (4). 19- Norandrostenedione (lg-norAD), which lacks the C-19 an-

* This work was supported in part by American Cancer Society Grant BC444 and National Institutes of Health Grant DK37551. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Recipient of a Monsanto Graduate Fellowship in Gene Research and Biotechnology.

Recipient of National Institutes of Health Research Career De- velopment Award AM1005. To whom correspondence should be ad- dressed.

’ The abbreviations used are: P-450arom, aromatase cytochrome P-450; AD, 4-androstene-3,17-dione; 19-norAD, 4-estrene-3,17-dione; P-450cam, camphor hydroxylase from Pseudomonas putida.

Androstenedione

NADPH L O 2

Estron SCHEME 1

gular methyl group of androstenedione (AD), also undergoes metabolism in human placental microsomes leading to aro- matization (5,6). There has been some question as to whether the same enzyme is responsible for the metabolism of both AD and 19-norAD, because the substrates differ from each other in their sensitivity to inhibition by the heme-iron li- gands, cyanide (7), and carbon monoxide (8). 19-NorAD aro- matization is more sensitive to inhibition in both cases.

In the present study, we have used purified human placental P-450arom to investigate the metabolism of 19-norAD. The results indicate that 19-norAD is metabolized by the purified enzyme and that 19-norAD binds to the same substrate site as AD. The compounds nevertheless exhibit differential sen- sitivity to cyanide inhibition, as observed in microsomes. Spectroscopic studies suggest that the differential inhibition arises because AD competes with cyanide for binding to P- 450arom, whereas 19-norAD promotes cyanide binding. We interpret these findings in terms of the structure of the enzyme-substrate complex and the proximity of C-19 to the heme-iron.

EXPERIMENTAL PROCEDURES

Steroids-AD and 19-norAD were purchased from Steraloids, Inc. (Wilton, NH). [1,2-3H]AD (80% l&?8-3H) and [6,7-3H]estrone were purchased from New England Nuclear. [3H]AD was purified by high pressure liquid chromatography prior to use (9).

Enzyme Preparations-Microsomes from human term placentas were isolated as previously described (4), except that microsomes were isolated in the absence of AD for the experiment shown in Fig. 5. P-450aron was solubilized from placental microsomes and purified to 12 nmol/mg protein (4). Cytochrome P-450 reductase was purified from livers of phenobarbital-induced rabbits (10) and had a specific activity of 20.6 rmol of cytochrome c reduced/min/mg protein.

Assay Procedures-Aromatase activity was determined by measur- ing tritium release from [lp,2g-3H]AD (11) or by radioimmunoassay for estrone (9). These two assay methods have previously been shown to agree quantitatively (4,9), and linear reaction rates were observed in all experiments reported. The reaction volume was 0.5 ml and contained AD or 19-norAD, purified P-450arom, 100 nM purified

8840

The Active Site of P-450arom 884 1

rabbit liver cytochrome P-450 reductase, 0.003% Nonidet P-40, 0.3 mM dithiothreitol, 25 mM glucose 6-phosphate, and 0.25 units glucose- 6-phosphate dehydrogenase. Reactions were initiated with 20 pM NADPH. When 19-norAD was the substrate, no attempt was made to differentiate between its aromatization and lp-hydroxylation, both of which occur in human placental microsomes (6). This assay was performed by terminating the enzyme reaction with 5 p l 6 N NaOH, which causes A-ring dehydration of l&hydroxy-l9-norAD, thereby yielding estrone (6). When AD aromatization was measured by ra- dioimmunoassay for estrone, reactions were terminated in the same manner.

Spectroscopic Studies-Human placental microsomes or purified P-450arom were diluted into 200 mM sodium phosphate buffer, pH 7.2. Spectra were recorded using 1-cm path length cuvettes at ambient temperature. Steroids were added to the sample cuvettes from stock solutions in ethanol. The ethanol concentration did not exceed 2%; this amount caused no significant spectral perturbations. The spec- trophotometer (Cary model 17D) was interfaced to a Zenith 2-100 computer adapted for data aquisition by On-Line Instruments (Jef- ferson, GA).

PDP 11/23 computer using the programs XTAL and MOLGRP, Molecular Graphics-Molecular structures were calculated with a

which were written in C. The structures were displayed on an Ad- vanced Electronic Design 767 terminal and photographed directly. The crystal coordinates for AD and iron protoporphyrin IX were obtained from Busetta et al. (12) and Timkovich and Dickerson (13), respectively. Bond lengths and angles for Fe-0-0 and Fe-C-N com- plexes were obtained from Phillips (14) and Scheidt et al. (15), respectively.

RESULTS

We determined that 19-norAD is a substrate of purified, reconstituted P-450arom in an assay designed to measure both aromatization and la-hydroxylation of 19-norAD. 19- NorAD undergoes metabolism at a linear rate under the conditions of the experiment shown in Fig. 1. In addition, the rate of estrogen formation is linear with respect to the con- centration of P-450arom, indicating that the cytochrome is rate-determining. A Lineweaver-Burk plot for 19-norAD me- tabolism by purified P-450arom shows that the substrate obeys Michaelis-Menten kinetics (Fig. 2). 19-NorAD exhibits a K,,, value of 132 nM and a maximal turnover number of 0.65 min". Whereas the K , value for 19-norAD is comparable to that for AD, the Vmax value is about one-tenth of that observed for AD, which is 6 min" (cf. Fig. 3 and Ref. 4). Consequently V,,,,JK,, the specificity constant for 19-norAD, is an order of magnitude lower than that for AD.

The possibility existed that the metabolism of 19-norAD and AD might result from discrete enzymes in the purified P- 450arom preparation; kinetic experiments were therefore car- ried out to address this question. A Lineweaver-Burk plot for

Time, min

FIG. 1. Time course of 19-norAD metabolism by purified, reconstituted P-4SOrrrom. The concentration of 19-norm was 1 p~ and the concentrations of P-450orom are shown on the right- hand side of the graph. The points plotted are the mean of duplicate determinations, and no individual values fell outside the symbols.

0 I 2 3 4 I

'19-norAD ( uM)-l FIG. 2. Kinetics of 19-norAD metabolism by purified, re-

constituted P-460arom. The concentration of P-450arom was 2 nM, and the incubation time was 8 min. The points plotted are the mean of duplicate determinations, and error bars indicate individual values where they fell outside the symbol. The line drawn is a least- squares fit to the data, which yielded a correlation coefficient (r) of 0.994.

L

K, 110 nM 0 2 4 6 .!/AD (NM)"

FIG. 3. Kinetic analysis of the inhibition of AD aromatiza- tion by 19-norAD. Aromatase activity was measured by tritium release from labeled AD. The concentration of purified P-450arom was 2 nM, and the incubation time was 8 min. The points plotted in double reciprocal form are the mean of two separate experiments, each of which employed duplicate determinations. The combined data were normalized to the V,. value of each experiment (6.3 and 5.6 min") and error bars indicate standard deviations which fell outside the symbols. 0, no inhibitor; A, 200 nM 19-norAD; El, 400 nM 19- norAD. The inset shows a replot of the slopes of the double reciprocal plot uersus 19-norAD concentration.

the inhibition of the aromatization of AD by 19-norAD reveals that 19-norAD binding is competitive with respect to AD aromatization, suggesting that 19-norAD binds to the same site as AD (Fig. 3). A replot of the slopes of the Lineweaver- Burk plot as a function of 19-norAD concentration yields a Ki value of 120 nM, which is close to the K , value (132 nM) determined in Fig. 2. Because the binding constant for 19- norAD inhibition of AD metabolism is similar to the binding constant for 19-norAD metabolism, the same binding site is likely to be responsible for the metabolism of both steroids. In experiments similar to those shown in Figs. 2 and 3, we observed that the K , and K; values for 19-norAD were also comparable to one another in human placental microsomes (data not shown).

Whereas it appears that both substrates share a common binding site, AD and 19-norAD differed in their sensitivity to inhibition by cyanide. Fig. 4 shows a dose-response experi-

8842 The Active Site of P-450arom

I + 2 0 m M KCN

KCN, mM FIG. 4. Dose-response data for the inhibition of P-46Oarom

metabolism of AD and 19-norAD by cyanide. The concentration of purified P-450arorn was 2 nM, and the incubation time was 8 min. Estrone formation from both substrates was measured by radioim- munoassay. The points plotted are the mean of duplicate determi- nations, and error bars indicate individual values where they fell outside the symbol. The control specific activities were 0.35 min" for 19-norAD metabolism and 2.4 min" for AD metabolism.

D . I

I "" + 19-nwAD ,"---- .

360 380 400 420 440 460 480 WAVELENGTH (nml

FIG. 5. Difference spectra produced in human placental mi- crosomes by AD and 19-norAD in the absence (upper panel) and presence (lower panel) of cyanide. The microsomal protein concentration wai 0.8 mg/ml, and the concentration of steroid added to the sample cuvette in each case was 10 p ~ . A fresh sample was used for each spectrum.

ment in which the concentration dependence of cyanide in- hibition was assayed using AD or 19-norAD as substrate for P-450arom. In this experiment, the substrate concentrations were twice their K,,, values, to give equal saturation of the enzyme with substrate in both cases. 19-NorAD is approxi- mately 3-fold more sensitive to cyanide inhibition than is AD under these conditions, with Is,, values for cyanide of 2 and 5.5 mM, respectively.*

Previous observations using C-19-substituted AD deriva- tives as probes of the active site of P-450arom indicated that the C-19 methyl group of androgen substrates might be posi- tioned in close proximity to the heme-iron when bound to P- ~

* For reversible, noncooperative enzyme inhibition, K, is related to Im by the equation Ki = Im/(l + S/K,,,), where S is the substrate concentration. This expression yields Ki values for cyanide inhibition of 19-norAD and AD metabolism of 0.67 and 1.8 mM, respectively. If the Ki value for cyanide inhibition was actually the same for the steroid substrates, it can be derived that the estimates of K,,, value for both substrates would have to be incorrect by greater than 2.3- fold to account for the disparity in Im values.

I T

0.07 I t I9-norADI

0.06 -/ W

I I I I I I I I I I I I I I I

350 370 390 410 430 450 470

WAVELENGTH (nm)

FIG. 6. Effects of AD and 19-norAD on the absorption spec- trum of purified P-46Oarom in the presence of cyanide. The concentration of P-450arom was 0.4 p ~ . The spectra were recorded in the sequence in which they are numbered. Spectrum 1, P-450arom in the presence of 0.8 p~ AD and 20 mM cyanide. Spectra 2-5, sequential addition of 0.1, 0.2, 0.6, and 2 p~ 19-norAD; difference spectra were generated by digital subtraction of spectrum 1 from the spectra induced by the 19-norAD additions. Spectrum 6, reversal of 19-norAD effect by addition of 30 p~ AD.

450arom (16). This suggested the possibility that the C-19 methyl group of AD might sterically interfere with cyanide binding to the heme-iron. Because 19-norAD lacks this methyl group, it would not interfere with cyanide binding, thus giving rise to the observed differential inhibition. We tested this hypothesis directly by measuring the effect of each steroid on cyanide binding to P-450arom as monitored by absorption spectroscopy. Difference spectroscopy was used initially to examine P-450arom in situ in human placental microsomes. AD and 19-norAD each yield "type 1" difference spectra (Fig. 5, upper panel); at saturating concentrations of each steroid, spectra of the same magnitude are produced, similar to pre- viously reported results (17). This is consistent with both steroids binding to the same cytochrome and causing a shift from the hexacoordinate, low spin form of P-450arom ( Amax = 417 nm) to the pentacoordinate, high spin form (Amax = 393 nm).

Using new microsome samples, we partially saturated the enzyme with 5 mM cyanide in both sample and reference cuvettes in order to determine the effects of AD and 19- norAD on cyanide binding (Fig. 5, lower panel). The addition of AD produces a difference spectrum with a peak at 390 nm and a broad trough which extends from 420 to 440 nm. The peak at 390 nm and the portion of the trough at 420 nm indicate an increase in the concentration of substrate-bound P-450arom and a decrease in the concentration of the sub- strate-free form of the enzyme. The cyanide complex of cy- tochrome P-450cam has a Soret absorption maximum at 439 nm (18); hence the portion of the trough at 440 nm probably results from a decrease in the concentration of the cyanide- bound form of P-450arom. Therefore AD binding to P- 450arom displaces cyanide. The addition of 19-norAD, on the other hand, induces a difference spectrum of the opposite sign (Fig. 5, lower spectrum). 19-NorAD produces a trough at 420 nm, indicative of a decrease in the concentration of substrate- free P-450arom and a broad peak near 450 nm, indicative of an increase in the concentration of the cyanide complex of the enzyme. Therefore 19-norAD promotes cyanide binding to P-450arom.

The differential effects of AD and 19-norAD are more

FIG. 7. Model for the structure of heme-ligand-substrate complexes of P-450arom. Upper panels show sche- matic representations and lower panels show computer-generated models with standard van der Waals radii. Left, heme-cyanide-19-norm complex. Right, heme-dioxygen-AD complex. The fol- lowing bond distances and angles were used in constructing the Jigand com- plexes (14, 15): Fe-C, 1.91 A, 90 degrees from the mean plane of the heme; C-N, 1.15 A, collinear with Fe-C; Fe-0, 1.83 A, 90" from- the mean plane of the heme; 0-0, 1.22 A, 115 degrees from Fe-0. In the models shown, the dihedral angle between the mean plane of the heme and the mean plane of the steroid A-ring is appoximately 45'. In both models, the heTe-iron is approximately equidistant (7 A) from C-1, C-2, and C-10 of the steroid. In the heme-dioxygen complex, the iron-bound oxygen atom is approxi- mately 5 A from C-1 and C-2 and 4 A from C-19 of the steroid.

The Active Site of P-450arom

Aromatase Active Site Model 8843

P-450- CN- 19-norAD P-450 - 0 2 AD

clearly seen using purified P-450arom and examining absolute spectra of the various complexes (Fig. 6). Spectrum 1 shows P-450arom in the presence of 0.8 p M AD and 20 mM cyanide. This spectrum is indistinguishable from the AD complex in the absence of cyanide, presumably because AD prevents cyanide from binding to P-450arom. Difference spectra 2-5 were produced by successive additions of 19-norAD (0.1-2 p ~ ) , and the spectra exhibit peaks at 440 nm and troughs near 390 nm. Spectrum 5, also shown in absolute form, is very similar to that of the cyanide complex of P-450carn, which has Amax = 439 nm and e = 78.5 nM". cm" (18), indicating that the presence of 19-norAD allows cyanide to coordinate to the heme-iron of P-450arom. Spectrum 6 shows the reversal of these spectral changes by the addition of a high concentra- tion of AD (30 FM). Thus, in the presence of cyanide, 19- norAD produces a spectrum characteristic of cyanide coordi: nation to the heme-iron of P-450arom, whereas in the absence of cyanide, 19-norm produces a high spin complex with a spectrum indistinguishable from that produced by AD (cf. Fig. 5A).

DISCUSSION

The findings presented establish that 19-norAD is metab- olized by purified P-450arom. The affinity of the purified enzyme for 19-norm (K, = 132 nM) is comparable to that for AD (K , = 110 nM); thus the C-19 methyl group of AD does not contribute significantly to its binding affinity. It appears that the two substrates share a common binding site, because the binding constant for 19-norAD metabolism is similar to its binding constant for competitive inhibition of AD aromatization (120 nM). However, the metabolism of 19-

norAD is approximately 3-fold more sensitive to cyanide inhibition than is AD metabolism. Whereas this could result from differences in rate-determining steps of the reaction cycles for the two steroids, the spectroscopic findings suggest that AD competes with cyanide binding to P-450arom, whereas 19-norAD promotes cyanide binding to the enzyme.

We propose that the differential effects of the two steroids are a consequence of the close proximity of the heme-iron and C-19 in the enzyme-substrate complex. AD, possessing a methyl group bonded to C-10, sterically competes with cya- nide binding. 19-NorAD, on the other hand, lacks this methyl group, and its binding opens a pocket immediately adjacent to the heme-iron, allowing cyanide to coordinate to the iron in the presence of the steroid. Positioning of C-19 close to the heme-iron is not unlikely, because C-19 is hydroxylated in the initial step of the aromatization reaction (Scheme 1). Moreover, the heteroatoms of 10P-substituted AD derivatives can apparently coordinate to the heme-iron of P-450arom while bound to the substrate site (16). In addition, the stere- ospecificity of isomeric forms of these 108-substituted AD derivatives indicates that the heme is likely to be positioned above C-1 and C-2 of the A-ring of the steroid in the enzyme- substrate complex (16).

A model for the possible position of androgen substrates in the active site of P-450arom is depicted in Fig. 7. In the left panels, 19-norAD is shown immediately below the heme moiety, and cyanide is shown bonded to the iron in a linear, end-on fashion (15). Because the C-19 methyl group is absent, there is sufficient space to accommodate the cyanide ligand. In the case of AD, the C-19 methyl group would sterically interfere with cyanide binding. During enzymatic turnover,

8844 The Active Site of P-450arom

dioxygen must bind to the heme-iron following reduction; however, unlike cyanide, dioxygen forms a bond with heme- iron at an angle of 115-130 degrees (14, 19). This configura- tion allows dioxygen to occupy a more restricted space. The right panel of Fig. 7 shows a heme-dioxygen complex with AD positioned in a manner identical to 19-norAD in the leftpanel; the C-19 methyl group does not interfere with oxygen binding because of the angle of the iron-dioxygen bond.

The close proximity of C-19 to the heme-iron in androgen- P-450arom complexes may explain the relative insensitivity of AD aromatization to carbon monoxide inhibition (8). Car- bon monoxide bonds to ferrous heme-iron in a linear, end-on fashion (20) and would thus be subject to the same steric factors which affect cyanide binding in the ferric state. Earlier spectroscopic studies with human placental microsomes showed that, like cyanide, carbon monoxide binding to cyto- chrome P-450 is antagonized by AD but augmented by 19- norAD (21). Thus, the proximity between C-19 and the heme- iron of P-450arom appears to exist in the ferrous state as well.

Evidence for close proximity of heme-iron and site of sub- strate hydroxylation has also been observed with other cyto- chromes P-450. With cholesterol side chain cleavage cyto- chrome P-450, active site-directed inhibitors (22-25), stopped-flow kinetic studies (26), flash photolysis experi- ments (27), and equilibrium binding studies (28) indicate that C-22 of the cholesterol side chain is located close to the heme- iron in the enzyme-substrate complex. The crystal structure of the camph9r complex of P-450cam has been solved (29), and the 1.6 A structure reveals that the target atom for hydroxylation ((2-5) is held in the active site at a distance of 4.2 A from the heme-ir~n.~ It thus appears that for the aforementioned cytochromes P-450, the apoprotein serves as a scaffold to position the substrate close to the iron and in the correct orientation for direct regio- and stereospecific attack by an iron-bound oxidant. This aspect of steroid hy- droxylase cytochromes P-450 has been important in the de- sign of inhibitors which possess high affinity and specificity by being capable of simultaneously occupying the substrate binding site and coordinating to the heme-iron (16,22-25,30, 31).

Acknowledgments-We thank Nghi Ta for preparing human pla- cental microsomes, Dr. John Dawson for supplying us with unpub- lished absorption spectra of P-450cam, Dr. Jay Edelman for his role in writing the molecular graphics computer program, and Steven Sands for purifying rabbit liver cytochrome P-450 reductase.

REFERENCES 1. Meyer, A. S. (1955) Biochim. Biophys. Acta 17,441-442 2. Akhtar, M., and Skinner, S. J. M. (1968) Biochem. J. 109, 318-

321

T. L. Poulos, personal communication submitted for publication.

3. Thompson, E. A., Jr., and Siiteri, P. K. (1974) J. Biol. Chem.

4. Kellis, J. T., Jr., and Vickery, L. E. (1987) J. Biol. Chem. 262 ,

5. Ryan, K. J. (1959) J. Biol. Chem. 2 3 4 , 268-272 6. Townsley, J. D., and Brodie, H. J. (1966) Biochem. J. 101, 25c-

7. Ganguly, M., Cheo, K. L., and Brodie, H. J. (1976) Biochim.

8. Meigs, R. A., and Ryan, K. J. (1971) J. Biol. Chem. 246,83-87 9. Kellis, J. T., Jr.; and Vickery, L. E. (1984) Endocrinology 114,

10. Johnson, E. F., Schwab, G. E., and Muller-Eberhard, U. (1979)

11. Rabe, T., Rabe, D., and Runnebaum, B. (1982) J. Steroid

12. Busetta, B., Comberton, G., Courseille, C., and Hospital, M.

13. Timkovich, R., and Dickerson, R. E. (1976) J. Biol. Chem. 2 5 1 ,

14. Phillips, S. E. V. (1980) J. Mol. Biol. 142, 531-554 15. Scheidt, W. R., Lee, Y. J., Luangdilok, W., Haller, K. J., Anzai,

K., and Hatano, K. (1983) Znorg. Chem. 22,1516-1522 16. Kellis, J. T., Jr., Childers, W. E., Robinson, C. H., and Vickery,

L. E. (1987) J. Biol. Chem. 262,4421-4426 17. Thompson, E. A., Jr., and Siiteri, P. K. (1974) J. Bwl. Chem.

18. Sono, M., and Dawson, J. H. (1982) J . Biol. Chem. 257, 5496- 5502

19. Jameson, G. B., Molinaro, F. S., Ibers, J. A., Collman, J. P., Brauman, J. I., Rose, E., and Suslick, K. S. (1980) J. Am. Chem.

20. Yu, N-T., Benko, B., Kerr, E. A., and Gersonde, K. (1984) Proc.

21. Juchau, M. R., and Zachariah, P. K. (1975) Biochem. Biophys.

22. Sheets, J. J., and Vickery, L. E. (1982) Proc. Natl. Acad. Sci. U.

23. Sheets, J. J., andvickery, L. E. (1983) J. Biol. Chem. 258,1720-

24. Sheets, J. J., and Vickery, L. E. (1983) J. Biol. Chem. 2 5 8 ,

25. Nagahisa, A., Foo, T., Gut, M., and Orme-Johnson, W. (1985) J.

26. Tuckey, R. C., and Kamin, H. (1983) J. Biol. Chem. 2 5 8 , 4232-

27. Mitani, F., Iizuka, T., Shimada, H., Ueno, R., and Ishimura, Y.

28. Heyl, B. L., Tyrrell, D. J., and Lambeth, J. D. (1986) J. Biol. Chem. 261,2743-2749

29. Poulos, T. L., Finzel, B. C., Gunsalus, I. C., Wagner, G. C., and Kraut, J. (1985) J. Biol. Chem. 260, 16122-16130

30. Sheets, J. J., Zuber, M. X., McCarthy, J. L., Vickery, L. E., and Waterman, M. R. (1985) Arch. Biochem. Biophys. 242 , 297- 305

31. Johnson, E. F., Schwab, G. E., Singh, J., and Vickery, L. E. (1986) J. Bwl. Chem. 261,10204-10209

249,5364-5372

4413-4420

27c

Biophy~. Acta 43 1,326-334

2128-2137

Mol. Pharmacol. 15, 708-718

Biockm. 17, 305-309

(1972) Cryst. Struct. Commun. 1, 128-133

4033-4046

249,5373-5378

SOC. 102,3224-3237

Natl. Acad. Sci. U. S. A. 81,5106-5110

Res. Commun. 65,1026-1032

S. A. 79,5773-5777

1725

11446-11452

Bwl. Chem. 260,846-851

4237

(1985) J. Biol. Chem. 2 6 0 , 12042-12048