[CANCER RESEARCH (SUPPL.
) 42, 3327s-3333s, August1982]
0008-5472/82/0042-0000$02.00
A New Hypothesis Based on Suicide Substrate Inhibitor Studies for the
Mechanism of Action of Aromatase1
Douglas F. Covey2 and William F. Hood
Department of Pharmacology, Washington University School of Medicine, St. Louis, Missouri
Abstract
Recently, it was discovered that 4-hydroxy-4-androstene-
3,17-dione, 4-androstene-3,6,17-trione, and 1,4,6-androsta-
triene-3,17-dione, compounds previously reported to be com
petitive inhibitors of aromatase, cause a time-dependent loss
of aromatase activity in human placental microsomes. We
report here that 1,4-androstadiene-3,17-dione (K¡0.32 fiM,
/W) 0.91 x 10~3/sec) and testolactone (K¡35 ¿UM; kmac,0.36
x 10~3/sec) also cause a similar loss of aromatase activity.
The mechanism which explains the unexpected loss of activity
caused by these five inhibitors is neither established nor ap
parent from current theories of the enzyme mechanism of
action of aromatase. We propose an inactivation mechanism
based on a new hypothesis for estrogen biosynthesis in which
the third enzyme oxidation carried out by aromatase results in
the formation of an enzyme-bound intermediate. This interme
diate is released as an aromatized product via a facile elimi
nation reaction which simultaneously regenerates the unaltered
active enzyme. Various structural modifications made in these
five inhibitors are hypothesized to redirect this elimination
reaction so that the steroid intermediate remains covalently
attached to the enzyme instead of being released as an aro
matized product.
Introduction
Chemically unreactive substrates for an enzyme which are
converted into chemically reactive irreversible inhibitors at the
active site of an enzyme when it carries out its normal catalytic
function are termed "suicide substrates." The rational design
of these inhibitors requires a knowledge of both the substrate
specificity of the target enzyme and its mechanism of action.
Since in the case of aromatase considerable information had
been accumulated in both these areas, we thought that it might
be possible to rationally design a suicide substrate for this
physiologically important enzyme.
Based on the widely accepted conversion of 4-androstene-
3,17-dione to 3,17-dioxoandrost-4-en-19-al via the initial 2
successive hydroxylations of the C-19 methyl group (4), we
rationalized that the unreactive steroid RED3 could be enzy-
matically converted into the reactive conjugated acetylenic
' Presented at the Conference 'Aromatase: New Perspectives for Breast
Cancer," December 6 to 9, 1981. Key Biscayne, Fla. This research was sup
ported in part by Grant CA-23582 awarded by the National Cancer Institute,
Department of Health and Human Services.
1 To whom requests for reprints should be addressed.
3 The abbreviations and trivial names used are: PED, 10-propargylestr-4-ene-
3,17-dione; 4-OH-A, 4-hydroxy-4-androstene-3.17-dione; AT, 4-androstene-3,-
6,17-trione; ATD, 1,4,6-androstatriene-3,17-dione; 1,4-ADD. 1,4-androstadiene-
3,17-dione; 4,6-ADD, 4.6-androstadiene-3,17-dione; testolactone, o-homo-17a-
oxaandrosta-1,4-diene-3,17-dione; DHT, 5o-androstan-17/!-ol-3-one; 5a-AD,
5a-androstane-3,17-dione; testololactone. o-homo-17a-oxaandrost-4-ene-3,17-
dione.
AUGUST 1982
Downloaded from cancerres.aacrjournals.org on July 28, 2020. © 1982 American Association for Cancer Research.
63110
ketone 10-(1-oxo-2-propynyl)-estr-4-ene-3,17-dione (10)
which would then covalently modify and inactivate the enzyme
(Chart 1). We were gratified to find that PED is indeed a potent
suicide substrate of aromatase (9), and we are continuing to
study PED as well as attempting to increase its potency through
further chemical modification of the primary PED structure.
Other investigators also have prepared PED as well as addi
tional C-19-substituted analogs of 4-androstene-3,17-dione for
evaluation as suicide substrates of aromatase (17, 19). For all
these compounds, the potential enzyme-generated reactive
species is postulated to result from enzymatic oxidation of
either the substituted C-19 carbon atom or the substituent
attached to this atom.
It is now clear, however, that the introduction of a nonreactive
C-19 substituent, which is activated by the previously dis
cussed mechanism, is only one way to design suicide sub
strates for aromatase. We have found, for example, that 4-OH-
A, AT, and ATD, compounds which contain an unaltered C-19
methyl group and which were previously shown to be compet
itive inhibitors of human placental aromatase (3, 21), are sui
cide substrates of this enzyme (6, 8). It is obvious that the 4-
hydroxy group of 4-OH-A and the 6-keto group of AT are
necessarily involved in the observed inactivation process since
these are the single structural modifications which distinguish
each of these steroids from 4-androstene-3,17-dione. For ATD,
however, this inhibitor differs from 4-androstene-3,17-dione by
the presence of 2 additional double bonds, and the possibility
existed that only one of these double bonds was causally
related to the ability of the inhibitor to decrease aromatase
activity. We report here results which demonstrate that only the
additional C-1 ,C-2 double bond is required for this process.
Thus, 1,4-ADD but not 4,6-ADD causes a time-dependent loss
of aromatase activity. In addition, testolactone, another steroid
containing a 3-keto-1,4-diene system, also was found to dis
play the kinetic profile expected for a suicide substrate. Finally,
we propose a new mechanism for estrogen biosynthesis which
may explain why these steroids should be aromatase suicide
substrates.
Materials and Methods
Materials. The [l^HH-androstene-S,! 7-dione (46.1 Ci/mmol)
was purchased from New England Nuclear, Boston, Mass. The 4-
androstene-3,17-dione, DHT, 5a-AD, 1,4-ADD, and 4,6-ADD were
purchased from Steraloids, Inc., Wilton, N. H. Testolactone and testo
lolactone were gifts from E. R. Squibb and Sons, Inc., Princeton, N. J.
All steroids used were chromatographically pure. NADPH was pur
chased from Sigma Chemical Co., St. Louis, Mo. Protein determinations
were done with a Bio-Rad Protein Assay Kit 1 purchased from Bio-Rad
Laboratories, Richmond, Calif. Liquid scintillation counting of tritium
was done in ScintiVerse purchased from Fisher Scientific Co., Pitts
burgh, Pa.
Enzyme Methods. Human placental microsomes were prepared by
3327s
 D. F. Covey and W. F. Hood

for protein determination. The remaining resuspended microsomes (1.5

ml) were spun again at 100,000 x g for 60 min at 4°.The supernatants
were discarded while the pellets (P2) were resuspended in buffer (1.5
ml), and portions (0.5 ml each) were taken for aromatase assay and
protein determination.

Results

Initially, 6 steroids known to inhibit the aromatization of 4-

androstene-3,17-dione (21, 22) were tested for their ability to
cause a time-dependent loss of aromatase activity (Table 1).
Chart 1. A proposed mechanism for the inactivation of aromatase by RED. An As expected, each steroid diminished the initially observed
alternate mechanism has been proposed by Metcalf et al. (19).
aromatase activity relative to control (no steroid added) be
cause of competition between inhibitor and radiolabeled sub
our previously described method (9). The specific activities of the 2
preparations used were 190 and 370 pmol/min/mg protein. Time- strate during the assay step. Microsomes appeared to increase
course experiments were done in a shaking water bath at 37° in air in activity when incubated with either DHT or 5a-AD for 30 min.
with initial incubations (3.0 ml) which contained 100 mM KCI, 10 mw This probably was due to the metabolism of these steroids by
potassium phosphate buffer (pH 7.5), 1 mM EDTA, 100 ¡IMNADPH other enzymes in the microsomes to products which had a
(omitted in appropriate controls), propylene glycol (1 drop per 0.5 ml), decreased affinity for aromatase. Three concentrations of 4,6-
and microsomes (0.039 to 0.083 mg protein per 0.5 ml). At various ADD were evaluated and, while minor losses of initial activity
times after the addition of inhibitors dissolved in ethanol (0.01 ml/0.5 occurred, they were not considered to be significant. Since in
ml incubation volume), aliquots (0.5 ml) were removed and added to a competition experiment we found that 4,6-ADD was a tightly
assay tubes containing [1,2-3H]-4-androstene-3,17-dione (0.5 nmol;
bound competitive inhibitor (apparent K, 42 nM) of the aroma
—¿700,000cpm), and remaining enzyme activity was determined by the tization of 4-androstene-3,17-dione (apparent Km 21 nM), it is
method of Thompson and Siiteri (23). After 5 min, the assay reactions
were stopped by the addition of chloroform (5 ml) and vortexing ~40 clear that we tested this steroid at concentrations sufficient to
saturate aromatase and therefore maximize any time-depend
sec. After centrifugation at 1470 x g for 5 min, aliquots (0.10 ml) were
removed from the water phase and added to scintillation mixture (10 ent loss of activity that could be caused by this compound.
ml) for determination of 3H¡>O production. Control experiments dem Testololactone caused a small loss of aromatase activity that
onstrated that: (a) the concentration of [1,2-3H]-4-androstene-3,17- was somewhat greater in the presence of NADPH cofactor,
dione [1 ;IM. which is -22-fold greater than the K, of this substrate and in a competition experiment it was found to be a competi
(14)] added for determination of remaining enzyme activity was suffi tive inhibitor (apparent K¡14 ¿IM)of the aromatization of 4-
cient to protect the enzyme from further significant loss of activity while androstene-3,17-dione. No further experiments were done with
rates of product formation were measured; (t>) complete aromatization testololactone to clarify the reason for the small time-dependent
of the radiolabeled substrate released 60% of the tritium into water;
loss of activity.
and, (c) the assay was linear over the 5-min assay in either the
By contrast, substantial losses of activity were found for 1,4-
presence or absence of inhibitor.
Competition experiments were carried out at 37°in 0.5-ml incubation ADD and testolactone in the presence of NADPH. Failure to
volumes containing [1,2-3H]-4-androstene-3,17-dione (6 different con observe this loss of activity when NADPH cofactor was omitted
centrations from 0.01 to 1 /IM) alone or with inhibitors dissolved in
ethanol (0.01 ml)-100 mM KCI-10 mM potassium phosphate buffer, pH Table 1
7.5-1 m« EDTA-propylene glycol (1 drop) in 10-ml screw cap glass Time-dependent effects of inhibitors on aromatase activity
test tubes. The reaction was started by the addition of microsomes Initial and remaining aromatase activity of microsomes exposed to inhibitors
in the presence (+) or absence (-) of 100 fiM NADPH for 30 min at 37° in air
[0.039 mg protein in 0.10 ml solution containing 0.5 mM NADPH, 100 was measured as described in "Materials and Methods." Activities are expressed
mM KCI, 10 mM potassium phosphate buffer (pH 7.5), and 1 mM EDTA] as relative percentage of the corresponding initial or remaining control (NADPH
to the incubations. Reactions were stopped by the addition of chloro present but no inhibitor) activity after 30 min. Controls routinely lost 16% activity
form (5 ml), and the 3H2O content was determined as described during the 30-min incubations.
previously. Concentrations of 4,6-ADD evaluated were 0.2, 0.5, and 1 Relative % of time-matched
UM, and concentrations of testololactone tested were 50 and 200 ¡JIM. control activity
Kinetic results were analyzed by Lineweaver-Burk plots (16). Regres
sion lines for all kinetic data were drawn according to a least-squares Inhibitor1 Initial75
M)4,6-ADD1
,4-ADD (1 .-,
fit.
6+9+ 1659
Experiments done to investigate the reversibility of the inhibition
caused by 1,4-ADD, testolactone, and 4,6-ADD were done at 37° in
air with initial incubations (3.0 ml) that contained microsomes (0.058 UM2?MTestolactone 64
mg protein per 0.5 ml) with or without added 100 /tM NADPH in the 53+
+ 441573
1968
buffer described earlier. Steroids were added in ethanol (60 ,»l) to yield
final concentrations of 10 /ÕM,400 ¡IM,and 10 MM 'or 1,4-ADD, UM)Testololactone
(400
testolactone, and 4,6-ADD, respectively. Controls received ethanol (60 + 6871 3069
/il) only. Sixty min after the addition of inhibitors, aliquots (0.5 ml) were
/IM)DHTdO/iM)5<>-AD(10
(400
removed and assayed for remaining activity as described earlier. The
68+
+ 588337
remaining incubation volumes (2.5 ml) were spun at 100,000 x g for
60 min at 4°.The supernatants were discarded while the pellets (Pi) 78+
were resuspended in buffer (2.5 ml). Portions (0.5 ml each) were
UM)NADPH 2430min72
removed for the assay of aromatase activity as described earlier and

3328s CANCER RESEARCH VOL. 42

Downloaded from cancerres.aacrjournals.org on July 28, 2020. © 1982 American Association for Cancer Research.
 Suicide Substrates and the Aromatase Mechanism

from the incubations suggests that aromatase enzyme catalysis 70

100
is involved in the inactivation process. Protection experiments 60 90
also support this conclusion. Thus, the inclusion of 0.5 /¿M 4- 80
I -^
50
androstene-3,17-dione in incubations containing NADPH and 70 -_
1 put 1,4-ADD increased the half-life of inactivation from 15.6
E 40 60 %_
to 38.3 min. The half-life of inactivation caused by 400 JUM Õ?,
o 50 2
testolactone (in the presence of NADPH) was increased from 6
33.5 to 315 min when 0.5 JUM4-androstene-3,17-dione was C
30
included in that incubation. The presence of 10 rriM dithiothre- E 40 a
in (V

itol, added as a scavenger of reactive electrophilic compounds, o

had no major effect on the rates of inactivation induced by 20 30 i

either of these inhibitors, indicating that an enzyme-generated
reactive intermediate did not escape from the aromatase active
site and later return to covalently modify and inactivate the 20
12 18 24 30 36 42
enzyme. Incubation Time (minutes)
For 1,4-ADD and testolactone, additional time-course exper Chart 3. Time course for the decrease in aromatase activity by testolactone
iments (Charts 2 and 3) were carried out at several inhibitor in the presence of NADPH. •¿control: testolactone: D. 25 /IM; A, 33.3 JIM;O, SO
»M:•¿.
' 00 ¡M:A, 400 IM. Duplicate experiments were done with 25 \M inhibitor,
but assays for remaining activity were taken at different times in each experiment
O.e., there are no duplicate measurements at any one time point). The data
shown for the remaining concentrations of inhibitor are the average and range of
duplicate experiments assayed at identical times.

concentrations, and the loss in activity was found to follow

pseudo-first-order kinetics. After a correction was made for the
loss of control enzyme activity during the experiments, the
kinetic data of Charts 2 and 3 were analyzed using double
reciprocal plots of the apparent rate constants for inactivation
(kapp = In 2/t,/2) versus the inhibitor concentration (15). For
1,4-ADD, the plot was linear (r = 0.979) and yielded values for
the apparent K, and pseudo-first-order overall rate constant for
decrease in activity of 0.32 ¡IMand 0.91 x 10~3/sec, respec
tively. For testolactone, the linear plot (r = 0.992) gave corre
sponding values of 35 fiM and 0.36 x 10"3/sec.
Finally, Table 2 summarizes the results of centrifugation
experiments done to investigate the irreversibility of the inac
tivation caused by 1,4-ADD and testolactone. The competitive
inhibitor 4,6-ADD was included in these experiments to assess
the difficulty involved in removing a high concentration (~238-
fold its K,) of a reversibly bound inhibitor from the microsomes.
Twice resuspended microsomes initially treated with 1,4-ADD
or testolactone in the presence of NADPH had 12.7 and 33.1 %,
respectively, of the original activity of control microsomes
incubated without added steroid. Based on the half-lives for
loss of aromatase activity found in the previously presented
time-course experiments, the expected remaining activity
would be 4.4 and 28.9%, respectively. Thus, in each case, the
agreement between the calculated and observed remaining
activity after excess reversibly bound inhibitor was removed is
quite good, and this demonstrates that the time-dependent
portion of the loss of activity caused by these inhibitors is not
readily reversible. As expected, twice-resuspended micro
somes treated with these inhibitors in the absence of NADPH,
or with the competitive inhibitor 4,6-ADD in the presence of
NADPH, had activities comparable to those of twice resus
pended control microsomes.
IO I4 I8 22 26 30 34
Incubation Time (minutes) Discussion
Chart 2. Time course for the decrease in aromatase activity by 1,4-ADD in
the presence of NADPH. The data shown for experiments carried out with 0.67 Six steroids were evaluated to obtain new information on the
and 5 /*M are each from single experiments. The data shown for the remaining
concentrations of inhibitor are the average and the range (bars; not shown when molecular structure of aromatase suicide substrates. The cri
this value is smaller than the symbol) of duplicate experiments. teria that we used to establish suicide substrate inhibition

AUGUST 1982 3329s

Downloaded from cancerres.aacrjournals.org on July 28, 2020. © 1982 American Association for Cancer Research.
D. F. Covey and W. F. Hood

Table 2
Evidence for irreversible time-dependent inactivation
Microsomes (0.058 mg protein per 0.5 ml) in the presence (+) or absence (-) of 100 /UMNADPH were incubated with inhibitors for 60 min
at 37°in air. An aliquot was removed for aromatase activity determination, and the remaining microsomes were pelleted at 100,000 X g for
60 min at 4°.The supernatant was discarded and the pellet (Pi) was resuspended for aromatase activity and protein content determination.
A portion of the resuspended microsomes were put through the centrifugation procedure again. The supernatant was discarded, and the pellet
(P2) was resuspended for aromatase activity and protein content determination. Activities are expressed as percentage of initial control
microsome activity.
1,4-ADD
fiM)Control (10 (iM)Testolactone (400
+Initial 4,6-ADD(10/iM) +
microsomes
Resuspended pellet (P,) 95.1 (0.028)6 42.2 (0.037) (0.030) 59.2 (0.033) 31.3(0.030) 73.9 (0.028)
Resuspended pellet (P2)100a 69.6 (0.020)9.4 65.3(0.017)0.83.8 12.7(0.019)14.9 65.9(0.015)16.9 33.1 (0.020)58.8 55.5(0.018)
Specific activity, 65.2 pmol estrogen per 5 min per 0.058 mg protein.
' Numbers in parentheses, actual protein concentration in 0.5-ml assay.

included: (a) a requirement for enzyme catalysis in the initiation

of a time-dependent pseudo-first-order loss of aromatase ac
tivity; (b) substrate protection against this inactivation; and (c)
NADPH,02 | HT rotation
failure to completely reverse the inhibition by twice pelleting
and resuspending the initially treated microsomes in steroid-
free buffer. Accordingly, 1,4-ADD and testolactone, but not
4,6-ADD, testololactone, DHT, or 5a-AD, were considered to hemiketol
formation D NAOPH.O,
be aromatase suicide substrates. The slow inactivation caused
by testolactone may explain how this compound, which has
such a poor affinity for aromatase, successfully lowers circu ENZ-O-Cn
lating estrogen levels in humans (2).
These results and those of our earlier studies (6, 8) demon
strate that 4-androstene-3,17-dione can be converted from a
substrate into a suicide substrate by the introduction of any
one of the following substituents: a 4-hydroxy group, a 6-keto
ENZ-0J
group, or a C-1 ,C-2 double bond. The reason why these
modifications in structure have produced this class of inhibitors
is unknown, and it continues to intrigue those of us interested
in the mechanisms of drug action. Our current experiments
with these inhibitors are guided by an inactivation mechanism
based on a new hypothesis for estrogen biosynthesis which is
outlined in Chart 4.
We have retained the proposal by Osawa (20) that the *X—ENZYME
direction of aromatase monooxygenations is trans to the C-
5.C-10 bond. After the first oxidation, rotation of the C-10.C-
19 bond orients the newly introduced hydroxyl group over Ring
A. This is the conformation of the alcohol observed in the HO
crystalline state by X-ray diffraction analysis (11 ). In the active Chart 4. New hypothesis for the mechanism of action of aromatase.
site of the enzyme, however, we propose that the hydroxyl
group is held over Ring A by virtue of hemiketal formation with moiety should be rapidly lost as formic acid, and the C-3
the C-3 carbonyl group. We envision that the second monoox- carbonyl group would be regenerated. Notice that the formic
ygenation occurs on this hemiketal to give a product which acid still contains the hydrogen present in the original aldehyde
loses water to become the aldehyde. By invoking the hemiketal group and the oxygen atom introduced in the third oxygénation
intermediate, we not only continue to account for the known step as required by the known isotopie labeling data (1). All
ability of the enzyme to distinguish between the C-19 prof? and that remains is for the newly formed covalent link that holds
proS hydrogens (20), but also we readily explain why the first the steroid to the enzyme (thus explaining why no satisfactory
and not the second oxygen atom introduced into the steroid is third intermediate has yet been isolated) to be broken. This
found ultimately in formic acid and not in water (1 ). In the third could be accomplished by an elimination reaction which frees
oxidation, we propose that the addition of oxygen to the C-19 the unaltered enzyme and forms a keto-diene which rapidly
carbon is accompanied by simultaneous bridging of the alde enolizes to the phenol, thus forming the estrone product. We
hyde to the C-3 carbonyl group so that the steroid remains have to further postulate that stereospecific loss of the 1/?-
covalently bound to the enzyme through a peroxy linkage. A hydrogen would occur in the elimination step and that enoli
separate nucleophilic group of the enzyme then attacks C-4 of zation of the keto-diene occurs with stereospecific loss of the
the steroid which results in concurrent cleavage of the peroxy 2/î-hydrogen to accommodate the remaining known isotopie
link to the heme of the cytochrome and transfer of the resulting labeling data (5, 24). We realize that this mechanism is pres
C-19 carbon fragment to C-3. The transferred C-19 formate ently more imaginative than substantive; however, it is not in

3330s CANCER RESEARCH VOL. 42

Downloaded from cancerres.aacrjournals.org on July 28, 2020. © 1982 American Association for Cancer Research.

ENZYME
ENZYME ENZYME
C.
(""BASE
ENZYME
ENZYME
Chart 5. Proposed mechanisms for the ¡nactivation of aromatase by: a, AT;
b, analogs of 4-androstene-3,17-dione containing a C-1.C-2 double bond; c,
4-OH-A.
conflict with known isotopie labeling data, and its heuristic
value will become apparent in the following discussion.
With this new mechanism as a hypothesis, we can now
rationalize why AT is an aromatase suicide substrate. The initial
2 hydroxylations of AT apparently occur without inactivation of
the enzyme as shown by our studies (7) in which the interme
diate 6-keto,19-alcohol and 6-keto,19-aldehyde were unable
to inactivate aromatase without further NADPH-dependent ca
talysis. We postulate that the third oxidation occurs as outlined
in our newly proposed mechanism to liberate formic acid. The
problem arises only when the enzyme has to release the steroid
in the final elimination step. Because of the 6-keto group, the
elimination reaction is redirected so that enolization of the 6-
keto group occurs instead of ejection of the enzyme nucleo-
phile (Chart 5a). Hence, the enzyme remains attached to the
steroid and the enzyme is inactivated.
Our mechanism also allows one to rationalize why several
steroids which contain a C-1 ,C-2 double bond (1,4-ADD, ATD,
and testolactone) inactivate aromatase. Once again, the inac
tivation occurs after all 3 oxidations have taken place when the
enzyme has to release the steroid product. In the presence of
the C-1 ,C-2 double bond, the steroid A-ring will aromatize
AUGUST 1982
Downloaded from cancerres.aacrjournals.org on July 28, 2020. © 1982 American Association for Cancer Research.
Suicide Substrates and the Aromatase Mechanism
without undergoing the final elimination reaction (Chart 5b).
This is disastrous for the enzyme since it needs to perform this
reaction to eject the steroid. Thus, by precluding the elimination
reaction from occurring, the steroid must stay attached, and
the enzyme is inactivated.
The apparent conflict between our 1,4-ADD inactivation
mechanism and the earlier finding that this steroid is enzymat-
ically converted to estrone (13, 18) can be resolved if one
assumes that substantial quantities of the aldehyde intermedi
ate formed after the second hydroxylation step diffuse from the
active site and nonenzymatically break down to estrone. Inter
estingly, the gem-diol equivalent of this aldehyde has been
proposed as a transient intermediate in the nonenzymatic aro-
matization of 2/8-hydroxy-3,17-dioxoandrost-4-en-19-al (12).
By contrast, enzyme inactivation would result from that portion
of the aldehyde which remains at the active site and undergoes
the third oxidation step.
Finally, the inactivation of aromatase by 4-OH-A also can be
envisioned to result from the redirection of this final elimination
reaction. We propose that the 4-hydroxyl group (following rapid
protonation) leaves as water, instead of the normally departing
enzyme nucleophile. Enolization then occurs to aromatize the
steroid A-ring, and once again inactivation results because the
ability of the enzyme to kick off the steroid product has been
precluded (Chart 5c).
It should be stressed that our hypothesis for the mechanism
of aromatase was formulated to explain how several steroids
of diverse structure could all be suicide substrates of this
enzyme. Other investigators interested in the mechanism of
action of these inhibitors will no doubt propose alternative
explanations. Regardless of which mechanism eventually
emerges as the one which best describes this inactivation
phenomenon, any hypothesis which provides the basis for well-
designed experiments will have served its purpose. If, however,
our hypothesis is correct insofar as that all the suicide sub
strates discussed here inactivate after the third aromatase
oxidation step while other suicide substrates like RED inactivate
after the second oxidation step, then the combination of struc
tural features from both classes of inactivators into one mole
cule could lead to the first known examples of "double-jeop
ardy suicide substrates" (i.e., one inactivator would provide
the enzyme with 2 successive opportunities to catalyze its own
demise). We are currently pursuing this possibility.
To our knowledge, of the compounds which we and others
have found to be suicide substrates of aromatase, only testo
lactone has been used to treat breast cancer in women. If
aromatase inactivation is the reason for the observed decrease
in serum estrogen levels caused by testolactone, then it is
possible that these levels could be diminished even further by
the more potent aromatase suicide substrates which are cur
rently available. Whether or not these other inhibitors have
additional undesirable pharmacological properties that pre
clude their use in humans also remains to be established.
References
1. Akhtar, M., Calder, M. R., Corina, D. L., and Wright. J. N. The status of
oxygen atoms in the removal of C-19 in estrogen biosynthesis. J. Churn
Soc. Chem. Commun., 129-130, 1981.
2. Barone, R. M., Shamonki, l. M., Siiteri, P. K., and Jude), H. L. Inhibition of
peripheral aromatization of androstenedione to estrone in postmenopausal
women with breast cancer using A'-testololactone. J. Clin. Endocrino!.
Metab.. 49: 672-676, 1979.
3331s
Discussion
3. Brodie. A. M. H.. Schwarzel, W. C., Shaikh, A. A., and Brodie. H. J. The
effect of an aromatase inhibitor, 4-hydroxy-4-androstene-3,17-dione, on
estrogen-dependent processes in reproduction and breast cancer. Endocri
nology, 700. 1684-1695, 1977.
4. Brodie, A. M. H. Recent advances in studies on estrogen biosynthesis. J.
Endocrinol. Invest., 2. 445-460, 1979.
5. Brodie, H. J., Kripalani, K. J., and Possanza, G. Studies on the mechanism
of estrogen biosynthesis. VI. The stereochemistry of hydrogen elimination at
C-2 during aromatization. J. Am. Chem. Soc., 91: 1241-1242, 1965.
6. Covey, D. F., and Hood. W. F. Enzyme-generated intermediates derived
from 4-androstene-3,6,17-trione and 1,4,6-androstatriene-3,17-dione
cause a time-dependent decrease in human placenta! aromatase activity.
Endocrinology, 108: 1597-1599, 1981.
7. Covey, D. F., and Hood, W. F. Studies on the mechanism of the time-
dependent decrease in human placental aromatase activity caused by 4-
androstene-3,6,17-trione. Fed. Proc., 40: 818, 1981.
8. Covey, D. F., and Hood. W. F. Aromatase enzyme catalysis is involved in
the potent inhibition of estrogen biosynthesis caused by 4-acetoxy- and 4-
hydroxy-4-androstene-3.17-dione. Mol. Pharmacol., 21: 173-180, 1982.
9. Covey, D. F., Hood, W. F., and Parikh, V. D. 10/8-Propynyl-substituted
steroids: mechanism-based enzyme-activated irreversible inhibitors of estro
gen biosynthesis. J. Biol. Chem., 256. 1076-1079, 1981.
10. Covey, D. F.. Parikh, V. D., and Chien, W. W. 19-Acetylenic, 19-hydroxy-
androst-4-en-3,17-diones. Potential suicide substrates of estrogen biosyn
thesis. Tetrahedron Lett., 2105-2108, 1979.
11. Duax, W. L, and Osawa, Y. Steroid structure and function. VI. The molecular
conformation of 19-hydroxy-4-androstene-3,17-dione as an intermediate for
estrogen synthetase. J. Steroid Biochem., 73.383-386. 1980.
12. Fishman, J., and Hosoda, H. Unusually facile aromatization of 2/J-hydroxy-
19-oxo-4-androstene-3,17-dione to estrone. Implications in estrogen bio
synthesis. J. Am. Chem. Soc., 96. 7325-7329, 1974.
13. Guai, C., Morato, T., Hayano, M., Gut, M., and Dorfman, R. I. Biosynthesis
Discussion
Dr. Zumoff: I have a question for Dr. Longcope concerning the
concept he presented of the relative validity of the blood and the urine
methods of measuring conversion. I think that if one is concerned about
how much androstenedione is converted to estrone, which is then
capable of affecting an organ distant from the area in which it is formed,
thereby having to travel through blood, then it is the blood conversion
which is the valid one from the point of view of the distant biological
effect, rather than the overall conversion which is measured in the
urine. As he pointed out from the ratios between the urine and the
blood conversion rate and the overall increase in the urine conversion
rate, I would estimate that very obese women should have virtually no
increase in blood conversion rate. This would account for the fact that
these very obese women who have such a high increase in urinary
conversion rate may have no detectable increase in blood estrone
levels.
Dr. Longcope: If you remember my diagram (Chart 1), Step 1 is the
overall aromatization in the tissues; then Step 2 is the estrogen going
into the pool of free estrogen in the blood. This is a measurement we
make when we determine the conversion in blood. This will also be the
measurement we make for urine if Step 3, the direct conjugation of
estrogen in the tissue, does not occur. From the data of all groups, in
normal-weight individuals the p values measured in blood and urine are
essentially identical, indicating that no conjugation of estrogen occurs
in the tissue without the estrogen first entering the blood pool. This is
different in obese individuals. In their case, the data of McDonald
indicate that it is not the conjugation in the tissue of origin but a very
slow release or entry of estrogen formed in the tissues into the blood
pool. There is no evidence that I am aware of for the necessity for
conjugation first. So if you continue infusion for 48 hr, then blood p
and urine p will be similar. This is an infusion period which I find a little
impractical, and I think most people would. Under those conditions, it
is easier to measure the urinary p, which will be higher, as I indicated,
by a factor in our groups of 4 times. It is not the conjugation in the
tissue of origin that presents a problem here. This is the problem that
occurs when you measure androgen interconversions in the urine, as
opposed to in the blood, where there is conjugation and further metab
olism in the tissue where the conversion occurs. This does not appear
3332s
Downloaded from cancerres.aacrjournals.org on July 28, 2020. © 1982 American Association for Cancer Research.
of estrogens. Endocrinology, 71: 920-925, 1962.
14. Kelly, W. G., Judd, D., and Stolee, A. Aromatization of A"-androstene-3,17-
dione, 19-hydroxy-A4-androstene-3,17-dione, and 1g-oxo-A'-androstene-
3,17-dione at a common catalytic site in human placental microsomes.
Biochemistry, 16: 140-145, 1977.
15. Kitz, R., and Wilson, I. B. Esters of methanesulfonic acid as irreversible
inhibitors of acetylcholinesterase. J. Biol. Chem., 237. 3245-3249, 1962.
16. Lineweaver, H., and Burk, D. The determination of enzyme dissociation
constants. J. Am. Chem. Soc., 56. 658-666, 1934.
17. Marcotte, P. A., and Robinson, C H. Evaluation of a series of 10/3 substituted
estr-4-ene-3,17-dione derivatives/inactivators of human placental estrogen
synthetase (aromatase). Fed. Proc., 40: 1867, 1981.
18. Meigs, R. A., and Ryan, K. J. Enzymatic aromatization of steroids. J. Biol.
Chem., 246. 83-87, 1971.
19. Metcalf, B. W., Wright, C. L., Burkhart, J. P., and Johnston, J. O. Substrate-
induced inactivation of aromatase by allenic and acetylenic steroids. J. Am.
Chem. Soc., 703. 3221-3222, 1981.
20. Osawa, Y. Mechanism of aromatization. In: R. O. Scow (ed.), Endocrinology
Proceedings of the 4th International Congress of Endocrinology, pp. 814-
819. Amsterdam: Excerpta Medica, 1973.
21. Schwarzel, W. C.. Kruggel. W. G.. and Brodie, H. J. Studies on the mecha
nism of estrogen biosynthesis. VIII. The development of inhibitors of the
enzyme system in human placenta. Endocrinology, 92. 866-880. 1973.
22. Siiteri, P. K., and Thompson, E. A. Studies of human placental aromatase.
J. Steroid Biochem., 6. 317-322, 1975.
23. Thompson, E. A., and Siiteri, P. K. Utilization of oxygen and reduced
nicotinamide adenine dinucleotide phosphate by human placental micro-
somes during aromatization by androstenedione. J. Biol. Chem., 249. 5364-
5372, 1974.
24. Townsley. J. D., and Brodie, H. J. Studies on the mechanism of estrogen
biosynthesis. III. The stereochemistry of aromatization of Ci9 and C18 ste
roids. Biochemistry, 7: 33-40, 1968.
to be necessarily the case with the aromatization. Therefore, I think
that increased aromatization does occur in people who are obese
since, as I mentioned, adipose tissue is the site of aromatization. The
more adipose tissue, the more activity.
Dr. Osawa: I have a comment and question to Dr. Covey in relation
to his mechanism that suggests that, for an inhibitor to be suicide
active, it requires first and second hydroxylations before actual etio-
glycosamine inhibition occurs. We have recently found that 19-noreth-
isterone compared to the 17a-ethinyl-19-nortestosterone is an effective
suicide inhibitor the action of which is time and dose dependent and
requires NADPH. It is certainly a better suicide inhibitor than A'-
testololactone. This is the major ingredient of contraceptive pills, so
the pill-taking woman may have in effect an inhibited aromatase system!
Compared to that, 17a-ethinyltestosterone is not an inhibitor at all. This
steroid has an angular methyl group as a requirement in its mechanism
and also a conjugated acetylenic ketone in this case, which is required
in the mechanism. But in this steroid you have to open the D ring up to
make it ketoacetylene. Whether this is true or not has to be tested. Yet
at least there is no hydrogen to be oxidized by the monooxygenase in
the usual sense. Still, they are active suicide inhibitors. How do you
accommodate that in your mechanism?
Dr. Covey: I don't! I'll let Dr. Robinson respond after I do. As you
know, Dr. Osawa, there is some disparity between the way Dr. Metcalf
and his group believe that propargylestradiene works as a suicide
substrate and the way I believe that it works. None of the experiments
which either he or I or my group has done would in fact unambiguously
distinguish one of those mechanisms from the other. So I certainly
could not say for sure that oxygen is not inserted directly into the
acetylene of propargylestrenedione (RED) and that inactivation in some
part might not result from that. In the end, there are 2 ways to resolve
that. You could make more chemical modifications, hoping to subtract
out one of the 2 mechanisms. The drawbacks to that approach are that
you have made a chemical modification and you must accept that, in
attempting to redesign the system to answer one question, you will
necessarily introduce new structural differences which may raise other
questions simultaneously. The second way would be to start to study
the metabolism of the steroid itself by the enzyme. Both of those
CANCER RESEARCH VOL. 42
A New Hypothesis Based on Suicide Substrate Inhibitor Studies
for the Mechanism of Action of Aromatase
Douglas F. Covey and William F. Hood

Cancer Res 1982;42:3327s-3332s.

Updated version Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/42/8_Supplement/3327s

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Subscriptions Department at pubs@aacr.org.

Permissions To request permission to re-use all or part of this article, use this link
http://cancerres.aacrjournals.org/content/42/8_Supplement/3327s.
Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)
Rightslink site.

Downloaded from cancerres.aacrjournals.org on July 28, 2020. © 1982 American Association for Cancer Research.