J. Steroid Biochem. Molec. BioL Vol. 45, No. 6, pp. 467--476, 1993 0960-0760/93 $6.00 + 0.00 Printed in Great Britain Pergamon Press Ltd

K I N E T I C E V I D E N C E F O R A U N I Q U E

T E S T O S T E R O N E - R E C E P T O R C O M P L E X I N

5 ~ - R E D U C T A S E S U F F I C I E N T G E N I T A L S K I N

F I B R O B L A S T S A N D T H E E F F E C T S O F 5 ~ - R E D U C T A S E

D E F I C I E N C Y O N I T S F O R M A T I O N

MORRIS KAUFMAN, 1'2. LEONARD PINSKY, 1'2A.5'6 MARK TRIFIRO, 3A ROSE LUMBROSO, 1 NELLY SABBAGHIAN 2"5 and BRUCE GOTTLIEB I

~Cell Genetics Laboratory, Lady Davis Institute for Medical Research, 2Centre for Human Genetics, 3Division of Endocrinology, 4Department of Medicine, SDepartment of Biology and 6Department of

Pediatrics, McGill University, Montreal, Canada

(Received 2 October 1992; accepted 2 February 1993)

Summary--When 5~-reductase-sutficient genital skin fibroblast (GSF) monolayers are incu- bated with testosterone (T), they first form androgen (A)-receptor (R) complexes that dissociate at a fast rate [k(37°C = 0.024 min- ~]. As T is converted to 5~-dihydrotestosterone (DHT), this population of T-R complexes is eventually replaced by one that dissociates much more slowly [k(37°C)= 0.006 min-~], at a rate typical of DHT-R complexes. During the course of T to DHT conversion, one may observe a population of A-R complexes that has a linear (monophasic) intermediate dissociation rate constant [k(37°C)= 0.012 min-q; this population cannot simply reflect a mixture of T - and DHT-R complexes. The rate at which the complexes are processed from one dissociative form to the next varies with the incubation temperature and the presence or absence of serum in the medium; it also varies within and among GSF strains under apparently constant conditions. To explain these facts, we propose a model that enables the 5ct-reductase enzyme to influence the processive dissociative behaviour of T-R complexes by engaging in some sort of coupling with the AR. The proposal is strengthened by a set of observations in cells with constitutive, mendelian or inhibitor- induced 5~-reductase deficiency that preclude a simple quantitative relation between A-R complex processing and the extent of T to DHT conversion.

INTRODUCTION

A major unresolved aspect of androgen-depen- dent mammalian development is the division of labour between testosterone (T) and 5~-dihy- drotestosterone (5~-DHT). This partnership has come to light by ontogenetic studies relating target tissue 5~-reductase activity to internal and external male genital morphogenesis[l] , and by the phenotypic analysis of genetic [2] or teratogenetic [3, 4] 5~-reductase deficiency in male humans or rats, respectively. For instance, Wolffian duct differentiation is T-dependent, while penoscrotal morphogenesis is DHT- dependent.

*To whom correspondence should be addressed at: Sir Mortimer B. Davis--Jewish General Hospital, 3755 Chemin Cote Ste-Catherine, Montreal, Quebec, Canada H3T IE2.

Androgens form complexes with a cognate intracellular receptor protein (AR) that can find and bind to specific regulatory sequences of D N A in order to promote or repress target gene transcription selectively. It is not yet known whether T-receptor (R) complexes are intrinsically less potent than D H T - R complexes in DHT-dependent targets, or whether such targets have transcriptional cofactors that favour D H T - R complexes or disfavour T - R complexes; the same questions apply to T - R complexes in T-dependent targets that lack 5at-reductase.

We have previously studied interaction of different androgens with the AR by measuring temperature-dependent changes in the rate con- stants of dissociation of mibolerone (MB)-, D H T - , methyltrienolone ( M T ) - and T - R com- plexes. Arrhenius plots of these data have shown that the AR exists in a hierarchy of

467

468 MORRIS KAUFMAN et al.

energy states when complexed to these different androgens [5]. In order of increasing thermo- dynamic stability, the relative energies in this hierarchy were: MB-R (state I ) > D H T - R (state II) > MT-R (state III) > T-R (state IV). We have also provided evidence [6] indicating that A-R complexes "mature" by occupying sequential states in this hierarchy, and that while choice of mature state is androgen- restricted, the process must also involve a fac- tor(s) that interacts non-covalently with the AR [7]. Here we extend these approaches to examine time-dependent changes in the dis- sociation rate constant of A-R complexes that appear during the intracellular conversion of T to DHT.

On the basis of these data we propose a model which envisions that: (1) T-R complexes are normally processed to an "intermediate MT-like (k = 0.12min-I) '' state while part of an "enzyme-receptor complex", (2) natural or inhibitor-induced 50t-reductase deficiency impairs formation or function of the "com- plex", (3) T can be converted to DHT not only when T is free but also when it is bound to the postulated "enzyme-receptor complex"; and (4) deficient 5~-reductase or AR activity may sometimes result from primary defects in their respective coupling partners that impair "complex" formation, but may remain otherwise covert.

EXPERIMENTAL

Cell culture

The genital skin fibroblast strains (GSF) were developed in our laboratory from small pieces of labium majus, scrotal or preputial skin obtained from healthy volunteers (con- trois), with informed consent according to the protocols approved by the Hospital Ethics com- mittee. GSF were also developed from patients referred for diagnosis of presumed androgen resistance. The strains from the patients with 5~-reductase deficiency were donated by J. Imperato-McGinley[8] and M. Hodgins[9]. The tumour cell line, DDT, was obtained from J. E. Norris [10]. Monolayers were grown to confluence in 5cm 2 dishes or 750 cm 2 roller bottles with Eagle's minimal essential medium supplemented with 10% foetal calf serum, 1 mM pyruvate, 10 mg/l garamycin and 60 mg/1 each of penicillin G and streptomycin sulphate. 24h before binding assays monolayers were

preincubated in a serum-free (s-f) version of the above medium made with Hank's salts and additionally buffered to pH7.4 with 15mM HEPES. In some experiments preincubation was in the same medium containing 10% serum.

Binding of [3HIT to ARs within intact GSF

Replicate monolayers in 5 cm 2 dishes were iocubated in a humidified 37°C incubator sup- plied with 5% CO2:95% air, with s-f medium containing 3nM [1,2,6,7-3H]T (105Ci/mmol) alone (to measure total binding) or together with a 200-fold excess of radioinert T (to measure non-specific binding). After incubation for various times the dishes were placed on a bed of ice, washed twice with 5 ml of Tris-HCl (20 mM, pH 7.4) containing 0.15 M NaC1 and 0.2% bovine serum albumin (BSA) and twice with the same buffer lacking BSA, and the monolayers were treated for 5 min at room temperature with 0.1% trypsin. The loosely adherent cells were scraped with a rubber policeman and centrifuged (4°C) at 200g for 5 min. The cells were resuspended in Tris buffer lacking BSA and recentrifuged. The final cell pellets were solubilized with 0.5N NaOH (1.5 ml) and sampled for protein [11] and radio- activity, the latter in 10 ml of a Beta-Max (ICN Biochem) scintillation fluid and counted in a Canberra Packard scintillation counter at an efficiency of 60%. 17fl-N,N-diethylcarbamoyl- 4-methyl-4-aza-5~-androstan-3-one (DMA, 10 -7 M; [12]) was added to the assay media to inhibit 5ct-reductase in control GSF (Merck Sharp & Dohme, Rahway, N J).

Dissociation of A -R complexes within intact cells

To measure the dissociation of particular A-R complexes within cells, the assay medium in some dishes was replaced by a "chase" medium containing 0.6pM of the respective radioinert A. After various periods of time at 37°C, the activity remaining was plotted semi- logarithmically. Rate constants were calculated from slopes of regression lines fitted to these data. In all assays, the concentration of specific A-R complexes formed or remaining at each time was computed by subtracting the non- specific binding measured in duplicate from total binding in triplicate.

Dissociation of A - R complexes in extracts of labelled cells

After incubation for 2 h with [3HIT (3 nM), the monolayers in 5 750 cm 2 roller bottles were

F u n c t i o n a l c o u p l i n g be tween t e s t o s t e r o n e - r e c e p t o r complexes a n d the e n z y m e 5ct- reductase

scraped into 10 ml Tris-HCl buffer and cen- trifuged (200g, 10 min, 4°C). The pellets were resuspended in I ml of 5 mM phosphate buffer Strain (pH 7.4) containing 0.4 M KC1, 1.5 mM mer- MCH-51 1 captoethanol and 50raM dithiothreitol and 307404 1

11863 I homogenized by 50 strokes of a Dounce appar- 2 atus (4°C). The homogenate was centrifuged 1927 i

2 (100,000g, 1 h, 4°C) in a Beckman Ti-50 rotor. 3

1107 1 The supernatant was applied at 4°C to a 2 Sephadex G-75 (0.9 x 30cm) column packed 3 with phosphate buffer lacking KC1; the void 61578 1

2 volume was collected and sampled for pro- 96095

2 tein. An aliquot was resuspended in 1 ml of 15678

phosphate buffer containing 0 . 5 0 charcoal and 2 14388 0.05% dextran T-70. The mixture was vor-

2 texed, centrifuged (10rain, 2000g, 4°C), and 2150

2 the supernatant was counted by radioscintil- 44365

lation as above. The remainder was placed in 2 a circulating 37°C water bath, and sampled 1315

1208 periodically as above. Non-specific binding, 855192 determined on material extracted from cells 1379

2140 labelled with [3H]androgen in the presence of 0 . 6pM radioinert androgen, was found 2019

78517 to be <10% of [3H]androgen bound in its absence. 2200

Identification of androgens in A - R complexes or in assay media

Briefly, a high-speed supernatant containing the [3H]A-R complexes, prepared as above, was loaded onto a column (0.9 x 10 cm) packed with matrix gel "green A" (Amicon Corp., Lexington, MA) in a 10mM TES buffer,

100 J

80

= 60

~ 40

E 8 ~ 20

I I I

0 50 1 O0 150

minutes

Fig. 1. D i s soc i a t i on o f A - R complexes wi th in in t ac t cells f r o m pa t i en t s wi th 5~t-reductase deficiency. M o n o l a y e r s were label led wi th 3 n M [3HIT ( O ) , [3HIMT ( l l ) a n d [ 3 H ] D H T ( A ) fo r 2 h a t 37°C a n d the [ 3 H ] A - R complexes c h a s e d a t 37°C wi th a 200-fo ld excess o f the respect ive

r ad io ine r t a n d r o g e n .

469

Table 1. Processive dissociation rate constants of A - R complexes

Rate constant (rain i)

Experiment 2 h 4 h 24 h

0.023 0.023 0.023 0.023 0.023 0.023 0.023 0.023 0.023 0.023 0.023 0.023 0.023 0.023 0.023 0.020 0.020 0.008 0.014 0.014 0.014 0.021 0.021 0.021 0.021 0.012 0.007 0.020 - - 0.008 0.020 0.020 0.008 0.008 0.008 - - 0.023 0.012 0.007 0.020 0.014 0.014 0.023 0.023 0.023 0.015 0.014 0.012 0.020 - - 0.007

- - 0.013 0.008 0.023 0.008 - - 0.023 0.008 0.008 0.014 0.014 - - 0.007 0.007 0.007 0.011 0.011 0.005 0.011 0.007 0.007 0.015 0.015 0.006 0.014 - - 0.014 0.014 0.007 0.007 0.013 0.013 - - 0.014 0.014 0.014 0.008 0.008 - - 0.008 - - - - 0.008 0.008 0.008 0.006 0.006 - -

To determine dissociation rate constants GSF were incubated at 37°C for 2, 4, and 24 h with [~H]T and their A - R complexes were allowed to dissociate (37°C) in the presence of a 200-fold excess of radioinert T.

pH 8.5, containing 0.5 mM 2-mercaptoethanol and 5 mM EDTA. The bound radioactivity was recovered from the gel at approx. 1.2 M KC1 with a linear salt gradient (0 .4-2M KCI).

[ 1 4 C ] T and -DHT (10,000 dpm of each) as well as 10 #g each of radioinert T and androstene- dione (ENE) were added to samples of A - R complexes or assay media before extracting them twice with diethyl ether. T was separ- ated from D H T on TLC plates (20 cm 2, alumi- num backed, silica gel F-250, Brinkmann Instruments) developed 2 to 3 times in 99.5:0.5 chloroform-methanol. The chromatographic procedure was monitored by the separation of T and ENE under shortwave u.v. light. After overnight autoradiography the [ I a C ] T

and D H T spots were cut off the TLC plate and placed into scintillation vials containing 10ml Beta-Max (ICN Biochem) for double- label counting in a Canberra Packard scintillation counter. Recovery of both ~4C- metabolites > 90%.

470 M O R R I S K A U F M A N et al.

RESULTS

Time-dependence of dissociation profiles within intact GSF and DDT cells

In cells from patients with 5¢-reductase deficiency exposed to androgens for 2 h, the dissociation rates of T - R complexes > M T - R complexes > D H T - R complexes ([5], Fig. 1) and did not differ in replicate cultures incu- bated for 24 h (not shown). In cells from con- trols and patients with androgen resistance not associated with 5~-reductase deficiency, on the other hand, the dissociation rates of the complexes were time-dependent when formed with T, but not with MT, or DHT. The rate at which the linear profiles changed from "T-like" [k(37°C) =0.024min-~], to "MT- like" [k(37°C) = 0.012 min-~], to "DHT-like" [k(37°C) = 0.006 min - 1] was extremely variable both within or among different cell strains (Table 1). In 17 assays the profiles were "T- like" after a 2 h incubation: 6 remained "T- like"; one was "MT-like" at 24 h; the remainder became "DHT-like" at either 4 or 24 h. In 10 assays the dissociation profiles were "MT- like" at 2 h; 4 of these became "DHT-like" at either 4 or 24h; the remainder showed no change. In 6 trials, rate constants were already "DHT-like" at 2 h.

The dissociative behaviour of A-R com- plexes in GSF incubated with T was unex- pected in two respects. The first was the remarkably great variation in the apparent rate of procession from one profile to the next. Prototypic processivity (Fig. 2) was exem- plified by strains 1107 (experiment 2) and 96095

1001

8O

~ 60

o~ ~ 40

~, 20 <

I ! !

0 50 1 O0 150

minutes

Fig. 2. Dissociation of A-R complexes within 5~-reductase- sufficient cells. Monolayers were labelled with 3 nM [3H]T at 37°C for 2h (O), 4 h (m) and 24h (A) and the [3H]A-R complexes chased at 37°C with a 200-fold excess of radioin-

ert T.

(experiment 1) in Table 1. After 2 h at 37°C A -R complexes dissociated linearly, at a rapid rate [k(37°C)=0.024min- t ] , that is typical of T - R complexes in 5~-reductase-deficient cells. After 4 h, the A -R complexes also dis- sociated linearly but their rate constants [k(37°C) = 0.012 min- 1] were typical of auth- entic M T - R complexes. After 24h incu- bation, A-R complexes dissociated at a rate [k(37°C) = 0.007 min- 1] identical to those formed in monolayers labelled directly with DHT. The frequency and extent of deviation from the prototypic behaviour was strongly reminiscent of the exceedingly variable 5~- reductase activity, largely still unexplained, that we [13] and others [14] have observed in GSF.

A B

-~:

rr 20 20

I I

0 50 100 150 0 50 100 150

minutes

Fig. 3. Theoretical dissociative behaviour of mixtures (20:80 [O]; 50:50 [1~]; 80:20 [A]) of DHT- and T-R complexes. These profiles were calculated arithmetically from the intracellular (A) and extracellular

(B) behaviour of T ( 0 ) - and DHT ( l l ) - R complexes.

Functional coupling between testosterone-receptor complexes and the enzyme 5,,-reductase 471

The second surprising observation was the linearity and remarkably slight variation of the dissociation rate constants within the class of "MT-like" complexes (t½ = 55 __+ 5.4 min; SD). In this regard, we calculated [Fig. 3(A)] that to simulate the dissociative behaviour of a pure population of M T - R complexes, a mixed population would have to be com- posed of T - and D H T - R complexes in equal parts.

The dissociation profiles of T - R complexes in the DDT tumour line were "T-like" at 2 h and "MT-like" after 24 h (n = 6). Less than 10% of the media [3H]T was metabolized to D H T at the latter time. In 2 other experiments the profiles were "T-like" at 24 h.

Composition and extracellular behaviour of A-R complexes

To gain further insight into these surpris- ing results, we extracted A - R complexes from GSF (n = 6) labelled with [3H]T, measured their T : D H T content and determined their extra- cellular dissociative behaviour. Three be- haviours were identified (Fig. 4). The one in Fig. 4(C) was linear and identical to A - R complexes extracted from cells labelled with [3H]DHT [7]. As expected > 90% of the hormone extracted from these complexes was DHT. The non- linear dissociation profile in Fig. 4(A) was iden- tical to that of T - R complexes extracted from 5~-reductase-deficient cells [15]. Predictably, >90% of the androgen bound to these com- plexes was [3H]T. The profile of A - R com- plexes in Fig. 4(B) was similar to extracted

M T - R complexes [5]. The narrow range of these dissociation profiles (shaded area; n = 4) con- trasted sharply with wide variation in the T: D H T (80-20% T--*20-80% DHT) that was isolated from these different preparations. Furthermore, they are incompatible with dissociation profiles calculated for different mixtures of T - and D H T - R complexes [Fig. 3(B)] based on extracellular dissociative behaviour of each component.

Temperature-dependence of dissociation profiles

Dissociation rate constants of A - R com- plexes formed in 5a-reductase-sufficient GSF labelled with [all]T, but not -DHT or -MT, were dependent on the incubation temperature. In 14 experiments GSF that were labelled (2h) and dissociated at 37°C yielded A -R complexes that dissociated linearly [Fig. 5(A)]: their rate constant was identical to D H T - R complexes [k(37°C) = 0.007 rain- i + 0.006 SD]. On the other hand, A - R complexes formed in replicate cultures incubated at 24°C, but dissociated at 37°C, did so more rapidly: their rate was akin to M T - R complexes [k(37°C)= 0.013 min-~ _+ 0.006 SD]. Another frequent pair of profiles is shown in Fig. 5(B). In 11 experiments GSF that were labelled and dissociated at 37°C had A -R complexes that dissociated more rapidly: their rates were identical to M T - R complexes. Pre- dictably, A -R complexes in their replicates labelled at 24°C dissociated even faster: their rates were identical to bone fide T - R complexes (k = 0.023 rain- z + 0.003 SD).

100

80

60

40

20

!

50 100

A

e- r -

E

x ¢D Q .

E 0 o

n"

i

0 150 0 50 I I I

100 150 0 50 100 150

minutes

Fig. 4. Extracellular dissociation of A-R complexes. Monolayers from different 5a-reductasv-sufficient cells were labelled for 2 h at 37°C with 3 nM [3H]T, and A-R complexes were extracted with a phos- phate buffer containing 0.4 M KCI. The supcrnatants from a high-speed ultracentrifugation were desalted and the [3H]A-R complexes dissociated at 37°C. The dissociation profiles were: (i) "T-like" in A, (ii) "MT-like" in B: range of results in 4 independent assays are depicted by the shaded area,

and (iii) "DHT-Iike" in C.

E

E 8 ¢, < :

Mo~ds KAUFMAN et al.

O ) ¢- r.-

E P u~

_.¢ Q . E O o

n "

100 ~

80

60

40

20

100 I

60

40

20

472

! I I I I I

0 50 1 O0 150 0 50 1 O0 150

minutes

Fig. 5. Effect of incubation temperature on the dissociation of A-R complexes. Different strains in A and B were labelled for 2 h at 37°C (ll) and 24°C (0) with 3 nM [3H]T and the [3H]A-R complexes thus

formed chased at 37°C with a 200-fold excess of radioinert T.

Effect of serum on dissociation profiles

Dissociation profiles in 5~-reductase-suff/- cient GSF labelled with [3H]T, were affected by the presence o f foetal calf serum in their incubation media. In its absence, A - R com- plexes formed and dissociated at 37°C [Fig. 6(A)] had rates identical to D H T - R complexes (k = 0.006 m i n - ~ + 0.001 SD; n = 7): replicates assayed in its presence dissociated faster, at rates equivalent to M T - R complexes (k = 0 . 0 1 3 m i n - ~ + 0 . 0 0 1 8 SD). In other exper- iments (n = 6) an effect o f serum was observed when GSF were labelled at 24°C but dissoci- ated at 37°C: the profiles were either (i) " M T - like" in the absence o f serum and "T-like" in its presence [n = 3; Fig. 6(B)]; or (ii) " D H T - like" in its absence and "MT-l ike" in its presence (not shown; n = 3).

Effect of steroidal inhibitors on dissociation profiles

Steroidal inhibitors o f 5~-reductase activity retarded the processivity o f A - R complex dis- sociation without altering androgen-binding activity. Profiles o f A - R complexes in 5~- reductase-sufficient G S F coincubated with 10 -7 and 1 0 - S M D M A were "T-like" and " M T - like", respectively; in its absence they were "DHT- l ike" [n = 6 ; Fig 7(A)]. In 8 other assays the profiles were "T-like" at both inhibitor concentrat ions while, in the absence o f D M A , they were "MT-l ike" and " D H T - like" on 3 and 5 occasions, respectively. D M A was also able to block t ime-dependent changes in the dissociation rate constants when it was added to the medium after incubation with [3H]T had begun [n = 3 ; Fig. 7(B)]. When

100

80

60

40

20

I I I

50 1 O0 150

1001 80l 60

40

20

I I

50 100

I

150

minutes

Fig. 6. Effect of foetal calf serum in preincubation and assay media on the dissociation of A-R complexes. 5g-reductase-sufficient cells were labelled for 2 h at 37°C (A) or 24°C (B) with 3 nM [3H]T in media which contained (0) or lacked ( I ) 10% serum and the [3h']A-R complexes thus formed chased at 37°C with

a 200-fold excess of radioinert T.

Functional coupling between testosterone--receptor complexes and the enzyme 5~,-reductase 473

100a 100 B

i

10 - 10 0 50 100 150 0 50 100 150

minutes

Fig. 7. Effect of steroidal inhibitors of 5~-reductase on the dissociation of A-R complexes. Monolayers in A were labelled for 2h at 37°C with 3 nM [3H]T in media without ( n ) or with 10 -7 ( 0 ) or 10 -8 (~.) M DMA. [3H]A-R complexes were chased at 37°C with a 200-fold excess of radioinert T. Monolayers in B were labelled in media without ( n ) or with I0-7M DMA for the last 90 min of the 2 h assay (&)

or for the entire incubation period (0) .

1 0 - 7 M w a s added to GSF for the last 90min of a 2 h incubation with [3HIT, the dissocia- tive behaviour of their A-R complexes was "MT-like"; in contrast, it was "DHT-like" in the absence of DMA and "T-like" in the pres- ence of 10-TM DMA during 2h labelling. Importantly, 10-7M DMA failed to block the formation of intermediate "MT-like" com- plexes in the DDT tumour cell line (n = 4; data now shown); nor did it alter the dissociation profiles of A-R complexes in GSF labelled with [3H]MT or -DHT (n = 2; data not shown).

DISCUSSION

It is not understood how DHT becomes bound to the AR within intact cells that have formed it from T by the action of 5~-reductase. It is usually assumed that upon entering cells, T binds preferentially to the AR because of its much lower affinity for 5~-reductase. T-R com- plexes dissociate ,~ 4 × faster than those formed directly with DHT[16, 17]. Therefore, as the intracellular concentration of DHT accumulates and that of T decreases, T-R complex dis- sociation will permit the "once-liganded" AR to become occupied with DHT. According to this model, the dissociative behaviour of an A-R complex depends solely on the identity of its androgen; and the dissociative behaviour of a mixed population of T- and DHT-R complexes should have a vectorial behaviour that is pre- dictable from the separate behaviours of its components (Fig. 3).

We have discovered that populations of A-R complexes formed within cells labelled with

[3H]T under various conditions (of time, tem- perature, presence or absence of serum, and presence or absence of 5~-reductase inhibitors) can very often be shown to adopt an "MT-like, intermediate" state on their way to becoming typical DHT-R complexes. According to the above model these data can only be explained by assuming that the profiles with rates of 0.012 min- ~ are composed of 50: 50 mixtures of T- and DHT-R complexes. Our failure to observe intermediate profiles with rates other than ~0.012min -] in more than 50 exper- iments, even though these GSF are known to have extremely variable rates of T---,DHT con- version [13, 14], suggests that this model cannot explain the observed dissociation profiles. Fur- thermore, the extracellular dissociation of A-R complexes formed by labelling 5~-reductase- sufficient GSF with [3HIT was linear (37°C) and "MT-like" [5], irrespective of the proportion of T and DHT bound to the receptor,

We have previously[18] proposed that the AR is a conformationally mobile protein and that different androgens (T, MT, DHT and MB) enable it to adopt successive confor- mational states (IV, III, II and I) in a sequence that is intrinsic to the receptor. According to the scheme in Fig. 8, we have reasoned that when the AR is labelled directly with DHT, DHT-R complexes start in state IV (or higher) and reach state II by undergoing a pair of time- and temperature-dependent transitions through state III. Likewise, MB-R complexes would reach state I by passing, respectively, through states IV, III and II. In this hierarchical view the dissociative behaviour of a particular A-R

SBMB 45/6~B

474 MORRIS KAUFMAN et al.

MB-R

STATE 4 3 2 1

Reaction Coordinate

Fig. 8. A model of the state transitions of the AR. The ordinate is in units of relative energy. The abscissa is the reaction coordinate and is meant to depict the energy changes of a receptor complex as the binding reaction

proceeds.

complex would depend more on the confor- mational state of its receptor than on the iden- tity of its androgen. Indeed, we have reported [5] different patients with receptor-defective andro- gen resistance whose mutant MB-R complexes dissociated, respectively, at rates identical to normal D H T - M T - and T - R complexes. Ac- cordingly, the dissociation of an "intermediate" state III A- R complex would be the same whether it contained T or DHT bound to it. The present data thus support a hierarchical view of A-R complex transitions by demonstrating that T - R complexes within 5ct-reductase-suf f ic ient

GSF undergo a time- and temperature-depen- dent transition from state IV to state III.

We have previously discovered that within the cells of some patients with qualitative AR defects, various A - R complexes dissociated lin- early, but more rapidly than normal, appar- ently because they had not undergone the usual number of transitions [19, 20]. When these complexes were extracted with high salt, on the

other hand, they were able to undergo time- and temperature-dependent transitions in confor- mity with the hierarchy[21]. Likewise T - R complexes extracted from cells of patients with 5~t-reductase deficiency dissociated non-lin- early, even though they did so linearly within cells. Furthermore, the rate constant of the "slow" component of this non-linear profile was that of an "MT-like" complex. These results suggested [5] that the conformation state attained by a particular intracellular A-R com- plex was regulated not only by the androgen, but also by an extra-receptor factor(s) that interacted non-covalently with the receptor.

In this paper we present data that implicate the 5ct-reductase molecule in the role of the extra-receptor factor. Figure 9 is a schematic representation of this idea. We visualize the AR (R) to contain two binding domains, one for androgen (here occupied by T), the other for 5~t-reductase (E). An initial state IV " T - R - E complex" changes conformationally to the in- termediate " T - R - E complex". Enzymatic con- version of T-- ,DHT could occur at this point, generating mixtures of "intermediate" com- plexes with either hormone. A subsequent con- formational change of "intermediate" state III D H T - R complexes would yield mature state II D H T - R complexes (with or without bound enzyme). Steroidal inhibitors of the enzyme activity probably act by binding to the cata- lytic site (possibly covalently as a Schiff base) thereby affecting, in some unknown fashion, the mobility of the state IV " T - R - E complex".

The mechanism by which "serum factors" may affect the behaviour of " T - R complexes" is not known. Greco and Gorski[23] postu- lated that a serum component is responsible for the decreased accumulation of progesterone

J R

State I V t I I I J I I

Fig. 9. Cooperative interaction of the AR and 5ct-reductase. This model depicts the androgen receptor (R) as having a binding site for 5¢t-reductase (E) as well as for testosterone (T). Metabolism of T to DHT

is postulated to occur via the state III "T-R-E complex".

Functional coupling between testosterone-receptor

receptors in estrogen-treated pr imary rat uter- ine, and M C F - 7 cells. By contrast , Ort iz-Caro et

al. [24] found that serum not only affected levels of T3-b inding in cul tured glial C6 cells, but also reduced the affinity of T3 for its receptor.

In G S F from some normal individuals and in the D D T tumour cell line, state IV T - R com- plexes did undergo a conformat iona l change to the " in termedia te" form despite unmeasure- able 5~t-reductase enzyme activity. Remarkably , 10 -7 M D M A blocked the format ion of " inter- mediate" complexes in these part icular GSF, but not in the D D T cells. This implies that D D T cells lack 50t-reductase enzyme molecules, while these G S F have enzymatically inactive molecules.

We observed interexperimental var ia t ion in the processive dissociative behaviour of some G S F strains under nomina l ly identical con- ditions. G S F strains have been shown to conta in cells (clones) with "h igh" or " low" 5~-reductase activities[14]. Al though the physicochemical basis for the dis t inct ion between "high" and " low" activity has no t been identified, it is reasonable to propose that the var iat ion we observed reflects concomi tan t f luctuat ion of the ratio of "high-" and " low-" activity cells in a s train at the time of different experiments.

Final ly, G S F from some patients with pre- sumed defects of the A R had unmeasurab le 5~t-reductase activity and their T - R complexes could not a t ta in the " in termedia te" state. This observat ion prompts us to postulate that a pr imary defect of the 5c~-reductase enzyme may rarely dis turb its coupling with the A R so as to result in clinically significant A R dysfunc- tion. Contrar i ly, a pr imary defect of the A R may rarely dis turb its coupling with the 5ct- reductase enzyme so as to result in clinically significant 5ct-reductase dysfunction. This may explain why 5~-reductase deficiency secondary to A R defects [25] is much less frequent than expected from the fact that the A R is necess- ary for androgenic induct ion of 5~t-reductase activity in several parts of the body [26-30].

Acknowledgements--This research was supported by the Medical Research Council of Canada, the Fonds de la Recherche en Sant6 du Qurbec, and the Fonds pour la Formation de Chercheurs et l'Aide ~ la Recherche du Qurbec. We thank Carlos Alvarado and Rhona Rosenzweig for faithful assistance.

REFERENCES

1. Sliteri P. K. and Wilson J. D.: Testosterone formation and metabolism during male sexual differentiation in

complexes and the enzyme 5ct-reductase 475

the human embryo. J. Clin. Endocr. Metab. 119 (1974) !13-125.

2. Impcrato-McGinley J., Guerrcro L., Gautier T. and Peterson R. E.: Steroid 5~t-reductase in man: an inher- ited form of male pseudohermaphroditism. Science 186 (1974) 1213-1215.

3. Imperato-McGinley J., Binienda Z., Arthur A., Minin- merg D. T., Vaughan E. D. Jr and Quimby F. W.: The development of a male pseudohermaphroditic rat using an inhibitor of the enzyme 5ct-reductase. Endocrinology 116 (1985) 807-812.

4. Blohm T. R., Laughlin M. E., Benson H. D., Johnston J. D., Wright C. L., Shatzman G. L. and Weintraub P. M.: Pharmacological induction of 5~t-reductase deficiency in the rat: separation of testosterone- mediated and 5ct-dihydrotestosterone-mediated effects. Endocrinology 119 0986) 959-966.

5. Kaufman M., Pinsky L., Gottlieb B. and Trifiro M.: The state transitions of normal and mutant androgen- receptor complexes in human genital skin fibroblasts. J. Steroid Biochem. 36 (1990) 651-657.

6. Kanfman M. and Pinsky L.: A single-site allosteric model of intracellular androgen-receptor interaction. J. Steroid Biochem. 32 (1989) ll3-119.

7. Kaufman M., Pinsky L., Simard L. and Wong S. C.: Defective activation of androgen-receptor complexes: a marker of androgen insensitivity. Molec. Cell Endocr. 25 151-162.

8. Imperato-McGinley J., Gautier T., Pichardo M. and Shackelton C.: The diagnosis of 5ct-reductase deficiency in infancy. J. Clin. Endocr. Metab. 63 (1986) 1313-1318.

9. Hodgins M. B.: Binding of androgens in 5~t-reductase- deficient human genital skin fibroblasts and its inhi- bition by progesterone. J. Endocr. 94 (1982) 415-427.

10. Syms A. J., Nag A., Norris J. S. and Smith R. G.: Glucocorticoid effects on growth and androgen receptor concentrations in DDT 1-MF-2 cell lines. J. Steroid Biochem. 28 (1987) 109-116.

11. Lowry O. H., Rosebrough N. J., Farr A. L. and Randall R. J.: Protein determination with the folin phenol reagent. J. Biol. Chem. 193 (1951) 265-275.

12. Liang T. and Heiss C. E.: Inhibition of 5~-reductase, receptor binding, and nuclear uptake of androgens in the prostate by a 4-methyl-4-aza-steroid. J. Biol. Chem. 256 (1981) 7998-8005.

13. Pinsky L., Kaufman M., Straisfeld C., Zilahi B. and Hall C. S.: 5 alpha-reductase activity of genital and nongenital skin fibroblasts from patients with 5 alpha-reductase deficiency, androgen insensitivity, or unknown forms of male pseudohermaphroditism. Am. J. Med. Genet. 1 (1978) 407-416.

14. Grittin J. E., Allman D. R., Durrant J. L. and Wilson J. D.: Variation in steroid 5 alpha-reductase activity in cloned human skin fibroblasts. Shift in phenotypic expression from high to low activity upon subeloning. J. Biol. Chem. 256 (1981) 3662-3666.

15. Kaufman M., Pinsky L., Schwartz A. and Long- Simpson L.: Possible implications of the comparative dissociative behavior of testosterone- and 5alpha-dihy- drotestosterone-receptor complexes formed by human skin fibroblasts. J. Steroid Biochem. 19 (1983) 561-565.

16. Kaufman M. and Pinsky L.: The dissociation of testos- terone- and 5 alpha-dihydrotestosterone-receptor com- plexes formed within cultured human genital skin fibroblasts. J. Steroid Biochem. 18 (1983) 121-125.

17. Maes M., Sultan C., Zerhouni N., Rothwell S. and Migeon C. J.: Role of testosterone binding to the androgen receptor in male sexual differentiation of patients with 5~t-reductase deficiency. J. Steroid Biochem. 11 (1979) 1385-1390.

18. Kaufman M., Pinsky L., Gottlieb B., Schweitzer M., Brezezinski A., von Westarp C. and Ginsburg J.: Androgen receptor defects in patients with minimal

476 Moams KAUFMAN et aL

and partial androgen resistance classified according to a model of androgen-receptor complex states. Horm. Res. 33 (1990) 87-94.

19. Kaufman M., Pinsky L., Bowin A. and Au M.: Famil- ial external genital ambiguity due to a transformation defect of androgen-receptor complexes that is expressed with 5 alpha-dihydrotestosterone and the synthetic androgen methyltrienolone. Am. J. Med. Genet. 18 (1984) 493-507.

20. Pinsky L., Kaufman M., Killinger D. W., Burko B., Shatz D. and Volpe R.: Human minimal androgen insensitivity with normal dihydrotestosterone-binding capacity in cultured genital skin fibroblasts: evidence for an androgen-selective qualitative abnormality of the receptor. Am. J. Hum. Genet. 36 (1984) 965-978.

21. Gottlieb B., Kaufman M., Pinsky L., Leboeuf G. and Sotos J. F.: Extra-cellular correction of the androgen- receptor transformation defect in two families with complete androgen resistance. J. Steroid Biochem. 211 (1987) 279-284.

22. LeGoff J. M., Martin P. M., Ojasoo T. and Raynaud J. P.: Non-Michaelian behaviour of 5a-reductase in human prostate. J. Steroid Biochem. 33 (1989) 155-163.

23. Greco T. L. and Gorski J.: Monohydroxytamoxifen inhibits progesterone receptor concentrations in pri- mary uterine cells in serum-free medium. Endocrinology 124 (1989) 2837-2843.

24. Ortiz-Caro J., Montiel F., Yusta B., Pascual A. and Aranda A.: Down-regulation of thyroid hormone nuclear receptor levels by L-triiodothyronine in cultured C6 cells. Molec. Cell Endocr. 49 (1987) 255-263.

25. Jukier L., Kaufman M., Pinsky L. and Peterson R. E.: Partial androgen resistance associated with secondary 5 alpha-reductase deficiency: identification of a novel qualitative androgen receptor defect and clinical implication. J. Clin. Endocr. Metab. 59 (1983) 679~88.

26. Kuttenn F. and Mauvais-Jarvis P.: Testosterone 5a reduction in the skin of normal subjects and patients with abnormal sex development. Acta Endocr. 79 0975) 164-176.

27. Mauvais-Jarvis P.: Androgens and Antiandrogens. Raven Press, New York (1977).

28. Kuttenn F., Mowszowicz I., Schaison G. and Mauvais- Jarvis P.: Androgen production and skin metabolism in hirsutism. J. Endocr. 75 (1977) 83-91.

29. Mowszowicz I., Melanitou E., Doukani A., Wright F., Kuttenn F. and Mauvais-Jarvis P.: Androgen binding capacity and 5a-reductase activity in pubic skin fibro- blasts from hirsute patients. J. Clin. Endocr. Metab. 56 (1983) 1209-1213.

30. Mowszowicz I., Melanitou E., Kirchoffer M.-O. and Mauvais-Jarvis P.: Dihydrotestosterone stimulates 5at- reductase activity in pubic skin fibroblasts. J. Clin. Endocr. Metab. 56 (1983) 320--325.