An NADH-induced conformational change that mediates the sequential 3 beta-hydroxysteroid dehydrogenase/isomerase activities is supported by affinity labeling and the time-dependent activation of isomerase.

3β-Hydroxysteroid dehydrogenase (3β-HSD) and steroid Δ-isomerase were copurified as a single protein from human placental microsomes. Because NADH is an essential activator of isomerase (Kact = 2.4 μM, Vmax = 0.6 μmol/min/mg), the affinity alkylating nucleotide, 8-[(4-bromo-2,3-dioxobutyl)thio]adenosine 5′-diphosphate (8-BDB-TADP), was synthesized. 8-BDB-TADP activates isomerase (Kact = 338 μM, Vmax = 2.1 μmol/min/mg) prior to inactivating the enzyme. The inactivation kinetics for isomerase fit the Kitz and Wilson model for time-dependent, irreversible inhibition by 8-BDB-TADP (KI = 314 μM, first order maximal rate constant kobs = 7.8 × 10−3 s−1). NADH (50 μM) significantly protects isomerase from inactivation by 8-BDB-TADP (100 μM). The isomerase activity is inactivated more rapidly by 8-BDB-TADP as the concentration of the affinity alkylator increases from 67 μM (t1/2 = 8.4 min) to 500 μM (t1/2 = 2.4 min). In sharp contrast, the 3β-HSD activity is inactivated more slowly as the concentration of 8-BDB-TADP increases from 67 μM (t1/2 = 4.8 min) to 500 μM (t1/2 = 60.0 min). We hypothesized that the paradoxical kinetics of 3β-HSD inactivation is a consequence of the activation of isomerase by 8-BDB-TADP via a nucleotide-induced shift in enzyme conformation. Biophysical support for an NADH-induced conformational change was obtained using stopped-flow fluorescence spectroscopy. The binding of NADH (10 μM) quenches the intrinsic fluorescence of the enzyme protein in a time-dependent manner (rate constant kapp = 8.1 × 10−3 s−1, t1/2 = 85 s). A time lag is also observed for the activation of isomerase by NADH. This combination of affinity labeling and biophysical data using nucleotide derivatives supports our model for the sequential reaction mechanism; the cofactor product of the 3β-HSD reaction, NADH, activates isomerase by inducing a conformational change in the single, bifunctional enzyme protein.

Our studies of purified human placental 3␤-HSD/isomerase have suggested that NADH mediates a shift in enzyme activity from dehydrogenase to isomerase by inducing a conformational change in the enzyme protein. NADH is the coenzyme product of the 3␤-HSD reaction and is a potent, essential activator of isomerase activity (7). NADH, but not NAD ϩ , completely protects both the 3␤-HSD and isomerase activities from inactivation by the substrate-site affinity alkylators, 2␣-bromoacetoxyprogesterone (5) and 5,10-secoestr-4-yne-3,10,17-trione (8). Finally, the isomerase-site-directed secosteroid, 5,10-secoestr-4-yne-3,10,17-trione, inactivated the isomerase activity with the expected first-order kinetics but inactivated the 3␤-HSD activity in an unexpected manner. As the concentration of the alkylating secosteroid increased, the rate of 3␤-HSD inactivation paradoxically decreased (8) instead of increasing in accordance with the Kitz and Wilson model of irreversible enzyme inhibition (9).
In this report, the interaction of 3␤-HSD/isomerase with coenzyme is studied using the cofactor-site affinity alkylator, 8-[(4-bromo-2,3-dioxobutyl)thio]adenosine 5Ј-diphosphate (8-BDB-TADP). In addition, the time-dependent effects of NADH binding on the intrinsic enzyme fluorescence and on isomerase activation are measured. 8-BDB-TADP, as well as the 6-and 2-(4-bromo-2,3-dioxobutyl)thio-derivatives of ADP, have been employed previously to study nucleotide binding by pyruvate kinase (10) and isocitrate dehydrogenase (11,12). Demonstration that the binding of substrate triggered a time-dependent quenching of protein fluorescence (13) or stimulation of enzyme activity (14) provided evidence for a ligand-induced conformational change in Escherichia coli dihydrofolate reductase. The unprecendented use of both kinetic data from affinity labeling and spectroscopic measurements of coenzyme binding furnishes a definitive test of our hypothesis that NADH induces a conformational change that is critical to the sequential reaction mechanism of 3␤-HSD/isomerase.
Enzyme Purification-3␤-HSD/isomerase was purified from human placental microsomes by our previously described method (1). The purified enzyme, which expresses both 3␤-HSD and isomerase activities, is a homogeneous protein according to SDS-polyacrylamide gel electrophoresis, the NH 2 -terminal sequence of amino acids, and fractionation of each activity during gel filtration chromatography (1,15). The "UVinvisible" nonionic detergent, Genapol C-100 (0.2% w/v), was substituted for UV-absorbing nonionic detergent, Emulgen 913 (0.2% w/v), as the eluting reagent during the DEAE chromatography to provide purified enzyme suitable for fluorescence spectroscopy. The enzyme purified using the Genapol C-100 was identical to enzyme purified using Emulgen 913 according to the 3␤-HSD and isomerase specific activities as well as the mobility of the single protein band during SDS-polyacrylamide gel electrophoresis. Protein concentrations were determined by the method of Bradford (16) using bovine serum albumin as the standard.
Synthesis and Stability of 8-BDB-TADP-The affinity alkylating nucleotide was synthesized as previously described by DeCamp et al. (10). The identity of each product in the three-step synthesis was verified by thin layer chromatography on cellulose sheets developed with isobutyric acid:concentrated NH 4 OH:H 2 O (66:1:33) and by ultraviolet spectra. The identity and purity of the 8-BDB-TADP were verified by the thin layer chromatography system (single spot, R f ϭ 0.63), ultraviolet spectrophotometry ( max ϭ 278 nm), and determination (11) of bromide content (1.1:1.0 molar ratio of hydrolyzable bromide to 8-BDB-TADP). These values agree with those previously reported for 8-BDB-TADP (10).
The decomposition rate of 8-BDB-TADP in 0.03 M MES buffer, pH 7.0, was determined by the loss of bromide at 22°C as previously described (11). The half-time for bromide loss was 56 min, in agreement with the reported half-life of 50 min under similar conditions (10).
Inactivation and Assay of the Enzyme-Inactivation of the pure enzyme (1.0 M) was carried out in experimental incubations that con- In protection studies, the control and experimental mixtures contained the same concentration of the potentially protecting steroid or cofactor with no increase in final solvent content compared to incubations without protector. The concentrations of these ligands were at least three times the K m or K I measured for 3␤-HSD or isomerase activity to facilitate competition with a subsaturating concentration of 8-BDB-TADP (100.0 M).
Nonspecific alkylation by 8-BDB-TADP was evaluated by preincubating the enzyme with ethyl bromoacetate (100.0 M) for 30 min followed by the addition of 8-BDB-TADP (100.0 M) using the conditions described above for enzyme inactivation. The rates of isomerase inactivation by 8-BDB-TADP were compared with or without preincu-bation with ethyl bromoacetate.
Assays that monitored the loss of 3␤-HSD or isomerase activity during enzyme inactivation were performed on aliquots taken from the incubation mixtures at appropriate time intervals in duplicate according to our published conditions (5). The slope of the initial linear increase in absorbance at 340 nm (due to NADH production during the oxidation of pregnenolone) per unit time was used to determine 3␤-HSD activity. Isomerase activity was measured by the initial absorbance increase at 241 nm (due to progesterone formation from 5-pregnene-3,20-dione) as a function of time. Changes in absorbance were measured with a Varian (Palo Alto, CA) Cary 219 recording spectrophotometer. Nonspecific and spontaneous enzyme activity were determined using blanks that contained either no steroid substrate or no enzyme, respectively. The incubation conditions cited in this report minimized spontaneous activity in the enzyme assays, and all measurements of enzyme activity were corrected for any observable nonspecific conversion of substrate.
Time-dependent Activation of Isomerase by NADH-Purified 3␤-HSD/isomerase (5.0 g) was dissolved in 0.02 M potassium phosphate buffer, pH 7.4, under three different preincubation conditions. 1) Isomerase substrate (5-pregnene-3,20-dione, at K m ϭ 10.0 M) was preincubated with the enzyme for 2.0 min, and NADH (at K act ϭ 2.4 M) was added to start the reaction at zero time.
2) The enzyme was preincubated in buffer alone, and the reaction was started by adding a mixture of the isomerase substrate and NADH. 3) The enzyme was preincubated in buffer plus NADH, and the isomerase substrate was added to start the reaction. In the spectrophotometric assay (241 nm) for isomerase activity (5), the substrate or NADH was quickly mixed (Ͻ5 s) in the cuvette at zero time (after the 2.0-min preincubation) so that the nmol progesterone formed could be measured from the Varian Cary 219 recorder tracing.
Fluorescence Spectroscopy-The time-dependent change in enzyme fluorescence due to the binding of NADH was measured using an Applied Photophysics (Surrey, UK) stopped-flow spectrophotometer. To measure the intrinsic fluorescence of the protein, the excitation wavelength was 290 nm, and fluorescence was observed using a 305-nm cutoff filter. The fluorescence data were collected as 1000 data points over 200 s and fitted to an exponential curve using software provided with the Applied Photophysics instrument.
The stoichiometry and dissociation constant for the binding of NADH to the enzyme were determined using a Proton Technology International (South Brunswick, NJ) Alpha Scan 4000 spectrofluorometer. The decrease in intrinsic enzyme fluorescence produced by titration with NADH yielded a value for maximal fluorescence change at site saturation (⌬F max ). After correction with a blank tryptophan titration, each NADH concentration yielded a ⌬F value used to obtain ϭ ⌬F/⌬F max . The data was plotted as 1/(1-) versus [NADH]/ according to Stinson and Holbrook (17) to yield a stoichiometry and dissociation constant for NADH binding. For the stoichiometry study, the actual concentration of enzyme protein was determined by ultraviolet spectral measurements in 6.0 N guanidine (34 Tyr plus 16 Trp residues/enzyme dimer yielded a molar extinction coefficient at 280 nm of 1.346 ϫ 10 5 /dimer).

Evidence That 8-BDB-TADP Is a Site-directed Alkylator of
Isomerase-Like NADH, 8-BDB-TADP activated isomerase according to Michaelis-Menton kinetics ( Table I). The K act and V max values reported for 8-BDB-TADP were calculated from initial velocities (Ͻ1.0 min) because isomerase was measurably inactivated within 1-3 min by the higher concentrations of the affinity alkylator. Although 8-BDB-TADP exhibited much lower affinity (higher K act ) for the enzyme compared to NADH, 8-BDB-TADP activated isomerase to a 3.5-fold higher maximal velocity than did NADH. In mixed activator analysis, isomerase activity was measured with NADH alone, 8-BDB-TADP alone, or a mixture of the two nucleotides (each at 0.5 ϫ K act ). The mixture of activators yielded an isomerase activity of 789 nmol/min/mg. Because this value was substantially less than the sum of the isomerase activities (910 nmol/min/mg) produced by NADH alone (144 nmol/min/mg) and 8-BDB-TADP alone (766 nmol/min/mg), competition between the NADH and 8-BDB-TADP for the same activator site of isomerase is strongly supported.

Inactivation of 3␤-HSD and Isomerase by 8-BDB-TADP-
The affinity alkylating ADP derivative inactivated both the 3␤-HSD and isomerase activities in an irreversible manner over time. The rate of 3␤-HSD inactivation paradoxically decreased as the 8-BDB-TADP concentration increased from 67.0 to 500.0 M (Fig. 1A). Conversely, isomerase was inactivated, as expected, at progressively faster rates when 8-BDB-TADP was increased over the same range of concentrations (Fig. 1B).
The Kitz and Wilson analysis (9) determined an inhibition constant (K I ϭ 314 M) and a first-order maximal rate constant (k obs ϭ 7.8 ϫ 10 Ϫ3 s Ϫ1 ) for the inactivation of isomerase by 8-BDB-TADP. In contrast to the first-order kinetics of isomerase inactivation, the inactivation of 3␤-HSD by 8-BDB-TADP exhibited an intriguing reversal of the expected kinetics that could not be analyzed by the Kitz and Wilson model.
The presence of 8-BDB-TADP in the experimental incubation mixtures stimulated the 3␤-HSD activity at zero time relative to the 3␤-HSD activity in control mixtures that contained ADP in place of 8-BDB-TADP (Fig. 1A). This indirect effect followed a concentration dependence that was consistent with the K act (338 M) measured for isomerase activation by 8-BDB-TADP. Because the inactivation data are plotted as percent of zero time activity, the controls appear to decrease as the ADP concentration increases. In fact, the control 3␤-HSD activity was quite similar at all of the ADP concentrations (mean Ϯ S.D. ϭ 67.1 Ϯ 4.7 nmol/min/mg), and the 3␤-HSD activity in the experimental mixtures increased as the 8-BDB-TADP concentration increased (values given in Fig. 1 legend). In contrast, the isomerase controls are plotted at 100% activity in Fig. 1B because aliquots of the control and experimental mixtures are diluted 20-fold into the isomerase assay mixture where saturating levels of isomerase substrate steroid and activating nucleotide (NAD ϩ ) induce the isomerase conformation. These assay conditions reverse the stimulation of isomerase by 8-BDB-TADP in the enzyme inactivation mixture. Stimulation by 8-BDB-TADP is evident in Fig. 1A because the 3␤-HSD assay conditions do not promote the isomerase conformation to obscure the effect of 8-BDB-TADP on the enzyme protein.
Time-dependent Activation of Isomerase-When the enzyme was preincubated with isomerase substrate steroid, the addition of NADH activated isomerase gradually over 1 min (t 1/2 ϭ 20 s). A time-dependent increase in isomerase activity was also measured when NADH and isomerase substrate were added simultaneously to enzyme preincubated in buffer alone (t 1/2 ϭ 18 s). In contrast, isomerase immediately exhibited maximal activity without the time-dependent lag when the substrate steroid was added to enzyme that had been preincubated with NADH (Fig. 3).
The spectrophotometric measurement of the activation of isomerase by 8-BDB-TADP did not reveal a time lag, suggesting that 8-BDB-TADP induces the conformational change in the enzyme more rapidly than NADH (data not shown).
Fluorescence Spectroscopy-The change in the intrinsic fluorescence of the enzyme protein induced by the binding of NADH was measured by stopped-flow fluorescence spectroscopy (Fig. 4). The initial binding event was visible as the very rapid voltage change during the first second (Fig. 4, inset). The subsequent gradual change in voltage over 200 s characterized the NADH-induced conformational change in the enzyme protein (k app ϭ 8.1 ϫ 10 Ϫ3 s Ϫ1 , t 1/2 ϭ 85 s) and is represented by the best-fitting curve in Fig. 4.
The stoichiometry and dissociation constant (K d ) of the bind-

DISCUSSION
As our studies of purified human placental 3␤-HSD/isomerase with affinity alkylators have progressed (5, 7, 8, 18 -20), it has become increasingly clear that the two-step enzyme mechanism is more complex than a dehydrogenase reaction followed by an isomerase reaction along a single protein with separate substrate and coenzyme sites for each activity. In addition to the NADH/NAD ϩ protection (5,8) and the secosteroid inactivation (8) studies discussed above, NADH (20) and pregnenolone (19) protected the same tryptic peptides (Arg-250, Lys-175) in the enzyme from affinity radioalkylation by 2␣bromo[2Ј- 14 (19,20). This shift in affinity radiolabeling that was produced by NADH protection, but not by pregnenolone pro- In an identical assay, the enzyme was preincubated in buffer, and the reaction was started by adding a mixture of the substrate steroid and NADH (f). In another identical assay, the enzyme was preincubated with NADH, and the isomerase substrate was added to start the reaction (E). Additional experimental conditions are described in the text. The isomerase activity (nmol of progesterone formed) was measured at 15 equal intervals during the first minute and at 4 equal intervals during the second minute to obtain the enzyme velocity versus time plots. tection, provided further support for our hypothesis: NADH formed by the 3␤-HSD reaction induces a conformational change in the enzyme protein that activates isomerase. The hypothesis has now been definitively tested by measuring the inactivation of 3␤-HSD and isomerase with the NADH sitedirected, affinity alkylating nucleotide, 8-BDB-TADP, as well as by obtaining direct spectroscopic measurements of the timedependent activation of isomerase by NADH.
The mixed activator analysis and protection studies are consistent with 8-BDB-TADP binding at the NADH site on the enzyme. NADH completely protected against the inactivation of both 3␤-HSD and isomerase by all other affinity alkylators we have studied (2␣-BAP (5), FSBA (7), and secosteroid (8)). Complete protection against inactivation is an unusual observation and suggests that NADH protected in these cases by inducing a conformational change in the enzyme rather than by simple competition. The significant, but less than complete, protection seen with 8-BDB-TADP plus the mixed activator results suggest that NADH directly competes with this affinity alkylator.
NAD ϩ failed to protect either activity from 8-BDB-TADP. There has been no significant protection by NAD ϩ against any alkylator we have studied thus far, with the notable exception of the inactivation of 3␤-HSD by FSBA (7). NAD ϩ slowed the FSBA inactivation of 3␤-HSD by 3-fold but did not significantly protect isomerase from inactivation by FSBA. Although both FSBA and 8-BDB-TADP are alkylating analogs of adenosine, the respective alkylating groups are located at "opposite" ends of the adenosine molecule (8-position of adenine in 8-BDB-TADP versus the 5Ј-position of ribose in FSBA). FSBA is the weakest activator of isomerase in the group of nucleotide-analogs studied (Table I). Finally, FSBA inactivates 3␤-HSD by the expected first-order kinetics (not "reverse" kinetics) because FSBA does not significantly activate isomerase. The current study with 8-BDB-TADP indicates that FSBA binds at the NAD ϩ site when the enzyme is in the 3␤-HSD conformation, whereas 8-BDB-TADP binds at the NADH site after the 8-substituted nucleotide induces the enzyme to assume the isomerase conformation.
Because 8-BDB-TADP is a highly efficacious activator of isomerase, a definitive model has been developed to explain the "reverse" kinetics of 3␤-HSD inactivation by the affinity alkylating nucleotide. According to this model, the concentrationdependent activation of isomerase by increasing levels of 8-BDB-TADP (I) converts progressively more molecules of enzyme from the 3␤-HSD conformation (E) into the isomerase form (EЈ⅐I). Enzyme alkylated in the isomerase form (EЈ Ϫ I) retains significant dehydrogenase activity when an aliquot from the alkylator/enzyme mixture is diluted 10-fold in the 3␤-HSD assay cuvette, where the dehydrogenase conformation is favored at pH 9.7 (3␤-HSD optimum) with pregnenolone as substrate. However, enzyme alkylated in the isomerase conformation has no activity during the isomerase assay because the conformation is not shifted to the dehydrogenase form under these incubation conditions (at the isomerase optimal pH 7.5 with 5-pregnene-3,20-dione as substrate). This model is illustrated by the following reaction scheme: 3␤-HSD assay E ϩ I … EЈ ⅐ I 3 EЈ Ϫ I 3 EЈ Ϫ I ͑active for 3␤-HSD͒ P Isomerase assay P 3 EЈ Ϫ I ͑inactive for isomerase͒

REACTION II
As the concentration of 8-BDB-TADP is increased, more alkylated enzyme exists in the active EЈ Ϫ I form during the assay used to measure 3␤-HSD inactivation. Thus, the induction of the isomerase conformation (EЈ⅐I) by the reversible binding of 8-BDB-TADP to the enzyme is directly responsible for the decrease in the rate of 3␤-HSD inactivation as the concentration of 8-BDB-TADP increases. At each of the 8-BDB-TADP concentrations used (67-500 M), a portion of the enzyme molecules remains in the dehydrogenase form (E) during the inactivation. Because formation of the reversible enzyme-alkylator complex induces the isomerase conformation (EЈ⅐I), the enzyme in the dehydrogenase form is inactivated by 8-BDB-TADP via a bimolecular mechanism (E ϩ I 3 E Ϫ I). Based on the measurements of 3␤-HSD inactivation, enzyme alkylated by 8-BDB-TADP while in the dehydrogenase conformation (E Ϫ I) has no activity in the 3␤-HSD assay and partial activity in the isomerase assay.
The inactivation of isomerase by 8-BDB-TADP fits the equation for biphasic enzyme inactivation (10), where kfast and kslow represent the values of k obs ϭ 0.693/t 1/2 that were measured for the inactivation of isomerase and 3␤-HSD, respectively, at each concentration of 8-BDB-TADP (determined from Fig. 1, A and B). The variable F represents the fractional residual activity of isomerase. The calculated curves fit the data points measured for the inactivation of isomerase (Fig. 6B) until less than 20% of the initial activity remains. At this point, the observed data points are higher than the predicted curves because all available enzyme in the isomerase  (t) , where kfast and kslow represent the values of k obs ϭ 0.693/t 1/2 that were measured for the inactivation of isomerase and 3␤-HSD, respectively, at each concentration of 8-BDB-TADP (determined from Fig. 1, A and B). The remaining parameters and the significance of the fits are discussed in the text. Fractional enzyme activity (E/E 0 , E 0 ϭ 1.0) is plotted on a logarithmic scale along each ordinate, and time is represented by the linear scale on each abscissa. form (EЈ⅐I) has been inactivated as EЈ Ϫ I, leaving a mixture of E and E Ϫ I. Because the biphasic inactivation equation assumes that the alkylated enzyme (EЈ Ϫ I or E Ϫ I) has no activity, the observed data points exceed the values of the calculated points due to the partial isomerase activity of enzyme alkylated in the 3␤-HSD form (E Ϫ I).
The equation for biphasic enzyme inactivation overestimates the observed rate of 3␤-HSD inactivation. However, the data fits the equation for single-phase enzyme inactivation (Fig. 6A), where kobs represents the value of k obs ϭ 0.693/t 1/2 that was measured for 3␤-HSD inactivation at each 8-BDB-TADP concentration (determined from Fig. 1A). The variable P represents the fractional residual 3␤-HSD activity when an inactivation plateau is reached. Because 8-BDB-TADP decomposes relatively slowly (t 1/2 ϭ 56 min) compared to the rates of inactivation measured for 3␤-HSD, reagent decomposition causes the observed data points to exceed the predicted values only at the lower 8-BDB-TADP concentrations (67 and 100 M) after 20 min of inactivation. The need to switch from a two-phase to single-phase equation to fit the data obtained for the inactivation of 3␤-HSD supports the concept that enzyme alkylated in the isomerase conformation has full activity in the 3␤-HSD assay. Because a greater proportion of enzyme molecules are in the isomerase conformation as the 8-BDB-TADP concentration increases (Table I), higher concentrations of 8-BDB-TADP inactivate 3␤-HSD more slowly than lower concentrations to yield the "reverse" kinetics of 3␤-HSD inactivation.
Our hypothesis that NADH activates isomerase by inducing a conformational change in the enzyme protein is indirectly supported by this affinity labeling study: 8-BDB-TADP binds at the NADH site, activates isomerase, and produces the reverse kinetics of 3␤-HSD inactivation. Moreover, direct evidence for the NADH-induced conformational change has been obtained by observing the time dependence of both the activation of isomerase (Fig. 3) and the quenching of intrinsic protein fluorescence by NADH (Fig. 4). The fact that the time frame for the fluorescence change (t 1/2 ϭ 85 s) is longer than for the activation of isomerase by NADH (t 1/2 ϭ 20 s) suggests that a point is reached during the conformational change where the isomerase substrate is brought into proper juxtaposition with the amino acid residues that catalyze the reaction. Once that threshold is reached, isomerization proceeds at the maximal rate.
The stoichiometry of 1 mol of NADH bound/mol of enzyme dimer can be interpreted in two ways: 1) NADH induces the isomerase conformation in just one of the two subunits or 2) both subunits form a single NADH site when the enzyme is in the isomerase form. Whether one or both subunits participate in the isomerase activity will require studies of tertiary and quaternary structure by nuclear magnetic resonance spectroscopy or x-ray crystallography.
The inactivation data obtained with the NADH site-directed affinity alkylator, 8-BDB-TADP, complement the direct measurements of the NADH-induced activation of isomerase to validate our proposed mechanism for the sequential 3␤-HSD/ isomerase activity. As the 3␤-HSD activity oxidizes the 3␤-hydroxy-5-ene steroid (pregnenolone or dehydroepiandrosterone) to the 3-oxo-5-ene intermediate, NAD ϩ is reduced to form NADH. This NADH induces a conformational change in the enzyme protein that activates isomerase to produce the 3-oxo-4-ene steroid (progesterone or androstenedione). After the product steroid and NADH dissociate, the enzyme converts back to the dehydrogenase form and can again catalyze the reaction sequence. Understanding how the enzyme shifts from the first to the second reaction in the sequence will help us evaluate the relationship between the individual reaction mechanisms for 3␤-HSD and isomerase, which are currently being studied in our laboratory.