195

Molecular and Cellular Endocrinology, 68 (1990) 195-204 Elsevier Scientific Publishers Ireland, Ltd. MOLCEL

02212

The 56 kDa androgen-binding protein in human genital skin fibroblasts: its relation to the human androgen receptor Fred Pereira ‘, Denise Belsham ‘, Kimberley Duerksen I, Eduardo Rosenmann I, Morris Kaufman ‘, Leonard Pinsky 2 and Klaus Wrogemann ’ ’ Department of Biochemishy & Molecular Biology and Human Genetics, University of Manitoba, Winnipeg R3E 0 W3, Canada, and ’ Lady Davis Institute for Medical Research and Centre for Human Genetics, McGill University, Montreal, Quebec H3A IBI, Canada (Received 14 August 1989; accepted 25 October 1989)

Key words: Androgen receptor; Androgen-binding ing; (Human genital skin fibroblasts)

protein; Methyltrienolone

(R1881);

Mibolerone;

Photochemistry;

Covalent label-

Summary We have recently described in genital skin fibroblasts (GSF) a relatively abundant 56 kDa protein with androgen-binding activity. This protein is missing in GSF of most patients with complete androgen insensitivity syndrome (CAI). The protein has many characteristics compatible with the androgen receptor; it has in fact been tentatively considered as a precursor or degradation form of the prototypic (- 100 kDa) human androgen receptor. We have prepared an antiserum to this protein, which allowed us to detect it as a direct product by in vitro translation of mRNA from GSF. It is thus very unlikely to be a degradation product of a larger precursor. Furthermore, covalent photolytic labeling of this protein with the androgen analogue [3H]mibolerone revealed a much lower affinity for this protein than is known for the androgen receptor. Finally, the GSF of two exceptional patients with complete androgen insensitivity syndrome due to negligible androgen receptor-binding activity express this protein normally, as determined on two-dimensional gels by Western blot analysis with the antiserum and by photolytic covalent labeling with androgen analogues. These data indicate that the protein is not a precursor or a degradation product of the receptor; nor is it androgen-induced. They are more compatible with the idea that the protein is another member of the steroid/thyroid/retinoic acid receptor supergene family, perhaps as an unorthodox product of the human androgen receptor gene.

Introduction We have described in human genital skin fibroblasts (GSF) a relatively abundant 56 kDa

Address for correspondence: Dr. Klaus Wrogemann, Department of Biochemistry & Molecular Biology, Faculty of Medicine, University of Manitoba, 770 Bannatyne Ave., Winnipeg R3E 0W3, Canada. 0303-7207/90/$03.50

protein that shares many characteristics with the androgen receptor (Rosenmann et al., 1982; Thompson et al., 1983; Nickel et al., 1988; Wrogemann et al., 1988). A similar protein (58 kDa) was recently described and thought to be the human androgen receptor (AR) or a degradation product thereof (Kovacs and Turney, 1988). In further studies Kovacs et al. (1989) found that this protein, affinity-labeled with dihydrotestosterone

0 1990 Elsevier Scientific Publishers Ireland, Ltd

196

17P-bromoacetate (DHT-BA) is indistinguishable from non-covalently labeled androgen receptor complexes by a variety of criteria under non-denaturing conditions. Together, we have established that the proteins described by us and by Kovacs et al. are the same (Belsham et al., 1989). The androgen receptor monomer in various species is thought to have a common molecular size of approx. 90-110 kDa (Johnson et al., 1987; Brinkmann et al., 1988; Gyorki et al., 1988; van Loon et al., 1988; Mowszowicz et al., 1989; Mulder et al., 1989). The recent cloning and sequencing of the human androgen receptor cDNA indicates that the intact androgen receptor monomer should be a 94-99 kDa protein (Chang et al., 1988; Lubahn et al., 1988; Trapman et al., 1988; Faber et al., 1989) although in vitro transcription and translation of the cDNA also yields a series of smaller proteins (Chang et al., 1988). Because of its size the 56 kDa protein cannot be considered a precursor of the intact androgen receptor monomer. It could, however, be a part of the human androgen receptor, perhaps a product of its degradation, as suggested previously (Nickel et al., 1988; Kovacs et al., 1989). We present here evidence that the 56/58 kDa protein is not a degradation product of a larger precursor and that its photolytic covalent affinity for androgen binding is several orders of magnitude lower than that of the androgen receptor. Furthermore, on the basis of immunoreactivity, abundance, size, isoelectric points and photoaffinity labeling with two synthetic androgens this protein is indistinguishable from that found in two patients with complete androgen insensitivity syndrome (CAI) whose GSF have negligible androgen receptor-binding activity. Materials

and methods

Cell culturing Human genital skin fibroblasts (GSF) were obtained from the Repository for Mutant Human Cell Strains, Montreal, and from the collection of Dr. L. Pinsky, Montreal. The patients of mutant cell strains NHL and TRL have been described previously (Kaufman et al., 1976, 1979). Cells were grown under standard culture conditions in human McCoy’s 5a medium supplemented with 10% fetal calf serum (FCS) and 100 pg/ml of

penicillin G and streptomycin. All GSF grown until confluency for all experiments.

were

One- and two-dimensional gel electrophoresis One-dimensional sodium dodecyl sulphate (SDS) gel electrophoresis was performed according to Laemmli (1970). For two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) isoelectric focusing (IEF) was performed according to O’Farrell (1975) with the following minor alterations: (i) A mixture of 0.8% pH 5-7 ampholytes (LKB), 0.8% pH 6-8 ampholytes, 0.4% pH 3.5-10 ampholytes and 2% (w/v) CHAPS (3-[(3cholamidopropyl)-dimethylammonio] l-propanesulfonate) (Perdew et al., 1983) was used in the gel for IEF. (ii) IEF gels of 12.5 cm were formed in 17 cm long tubes (2.4 mm i.d.). (iii) About 80 pg of cell proteins in CHAPS-lysis buffer (9.5 M urea, 3% CHAPS, 2% ampholyte mixture, 30 mM dithiothreitol (DTT)) were applied to each gel; 150 pg were applied to preparative gels for protein purification. IEF was carried out for 16 h at 400 V followed by 1 h at 800 V for a total of 7200 volt-hours. IEF gels were equilibrated for 30 min in 10% (w/v) glycerol, 3 mM DTT, 2.3% (w/v) SDS, 62.5 mM Tris HCl pH 6.8. A Bio-Rad Protean unit was used for second dimension SDS-PAGE. The resolving gel was 14 X 13.5 cm and 0.75 mm thick at a constant gel-bis-acrylamide concentration of 10%. Gel electrophoresis was performed at a constant current of 20 mA and at a constant temperature of 17°C. Coomassie blue staining Gels were routinely stained overnight with 0.05% Coomassie brilliant blue R-250 (Bio-Rad) dissolved in 50% ethanol containing 10% acetic acid and destained in a solution of 50% methanol and 10% acetic acid. Silver staining After completion of electrophoresis, gels were stained using a modification of the silver-based colour staining method of Morrissey (1981): Gels were fixed in 7.5% acetic acid for 2 h, soaked in 50% methanol overnight, washed in deionized water 3-4 times, incubated for 30 min in 100 ml of 5 ,ug/ml DTT, followed by incubation in 0.1% AgNO, for exactly 30 min. The gels were then

197

washed 2-3 times in deionized water and developed in 100 ml of 3% Na,CO, containing 0.05% formaldehyde. The reaction was stopped by soaking the gels in 1% acetic acid solution for 30 min. Finally the gels were soaked for 30 min in water before they were dried onto filter paper. Antiserum preparation 56 kDa protein material (right spot, pl 6.7, see Fig. 1A) was cut out from Coomassie blue-stained 2D-PAGE gels and stored frozen at - 80 o C. The protein spots. were washed and homogenized in phosphate-buffered saline (PBS). They were mixed to an emulsion with an equal volume of Freund’s adjuvant to a final volume of 2 ml. This material was injected intradermally in 0.1 ml aliquots into the lower back of the rabbit. The primary inoculation was with 100 gel spots in complete Freund’s adjuvant. Subsequent boosts were given intrape~toneally and consisted of 50 spots in incomplete Freund’s adjuvant in a total volume of 2 ml. Blood was collected by bleeding from the ear vein. The antiserum was tested for reactivity with the 56 kDa protein by immunoblotting to nitrocellulose paper (Schleicher and Schuell). Antigen-bound antibodies were visualized using an enzyme-linked immunosorbent assay with horseradish peroxidase or alkaline phosphatase (Bio-Rad) according to the manufacturer’s recommendations. For Western blot analysis the antiserum was immunoabsorbed with a 56 kDa protein-negative cell strain extract bound to nitrocellulose paper. Immuneprecipitation of in vitro translated proteins Total RNA was isolated from thirty 150 mm dishes of GSF by the method of Chirgwin (1979). The RNA was then passed over an oligo-d(T) cellulose (Pharmacia) column according to Maniatis et al. (1982). The mRNA eluted was precipitated and resuspended into 20 yl of water. In vitro translation was performed using a kit (New England Nuclear) according to the supplier’s manual using 1 ~1 samples of mRNA. The 25 ~1 reaction mixtures were pooled and diluted to 500 ~1 with washing buffer (0.05 M Tris HCl, pH 7.4, 0.15 M NaCl, 5 mM EDTA, 1% SDS and 0.02% sodium azide). This yielded a solution of 1 million cpm/ml of trichloroacetic acid-precipitable counts of [35S]met~o~ne. 10-15 ~1 of specific antiserum to

the pf 6.7 56 kDa protein or preimmune serum were added and the mixture placed on an orbital rotator at 4 o C for 1 h. Subsequently, 50 ~1 of a 5% solution of protein A-Sepharose was added and left overnight on the rotator. The immune complexes bound to the Sepharose were pelleted and washed 9 times in 1 ml of washing buffer and released with 50 ~1 of 1 M acetic acid. The sample was then lyophilized and resuspended into CHAPS-lysis buffer and analyzed by 2D-PAGE. Proteins were visualized by Coomassie blue, and the translation products revealed by fluorography (Bonner and Laskey, 1974). Photoaffinity labeling of intact fibroblast cells with [3H]mibolerone From confluent 100 mm Petri dishes of cultured genital skin fibroblasts the standard medium was removed, and the dishes were rinsed twice with 5 ml of sterile PBS (8 g NaCl, 0.2 g KCl, 1.14 g Na,HPO,, 0.2 g KH,PO,, total volume 1 liter). Dishes were then incubated at 37°C for 2 h with labeled [3H]mibolerone (7a,17a-dimethyl-[l7@methyl-3H]19-nortestosterone Ci/mmol, (85 Amersham)) (Kovacs and Turney, 1988) + a 200fold excess of radioinert mibolerone in 3 ml Eagle’s minimal essential medium (Gibco) with 15 mM Hepes at pH 7.4. For photolysis the incubation medium was removed and the dishes placed in the inverted position on a UV transillu~nator, ’ Chromato-Vue’ Model C-63B (Ultraviolet Products, San Gabriel, CA, U.S.A.) for 10 min for UV exposure. After photolysis, cells were harvested by scraping in a minimal amount of isotonic NH,CO,, pH 7.4. The cells were sonicated in this medium, aliquots removed to measure protein (Bradford, 1976) and then freeze dried. The lyophilized material was resuspended in lysis buffers for one- or tw~dimensional PAGE. Results

The 56 kDa protein in protein maps of total proteins of genital skin fibroblasts The protein map of GSF strain MCH 6 is shown in Fig. 1A. The 56 kDa protein doublet clearly belongs to the class of more abundant protein spots. It must be emphasized that of the 23 normal GSF cell strains of either sex tested to

198

MCHF;

-

Silver

acid

MCH6

stained

alkaline -

+-

acid

56K Ab

alkaline --+

Fig. 1. Protein map of genital skin fibroblasts, strain MCH 6 and Western blot analysis with the antiserum against the 56 kDa protein. A: Silver-stained gel of 12,000X g supematant proteins; B: Western blot reacted with the anti-56 kDa rabbit antiserum. Arrowheads point to the 56 kDa protein doublets of pl 6.5 and pZ 6.7.

date the expression of the protein in strain MCH 6 is the strongest. Yet in all GSF cell strains the 56 kDa protein can be detected by protein staining (Nickel et al., 1988). We estimate on the basis of staining intensities that the 56 kDa protein represents at least 0.1% of total cell protein in strain MCH 6. Antiserum against the 56 kDa protein Cell strain MCH 6 was chosen for purification of the 56 kDa protein by preparative 2-D PAGE. Cut-out right-hand protein spots (Fig. 1A) were used for immunizing a rabbit, because the righthand spot generally is stronger than the left-hand one (Nickel et al., 198S), although not in the preparation shown in Fig. 1. The immune response was tested on Western blots of 2-D gels. An example of such an analysis is shown in Fig. 1B with a 1 : 500 dilution of total rabbit serum. The blot shows that the antiserum is specific for the 56 kDa protein, and detects it reliably at 1 : 500 dilutions. The result also confirms that the two spots are structurally related (Wrogemann et al., 1988), as the antiserum was prepared only against right-hand spots, but the left-hand spot is detected by the antiserum equally well.

In vitro translation of the 56 kDa protein To see whether the 56 kDa protein was synthesized as a larger precursor, its synthesis was tested by in vitro translation. Two-dimensional PAGE maps of total translation products clearly reveal the 56 kDa protein doublet of p1 6.5 and 6.7 (arrowheads, Fig. 2A). This was confirmed by immunoprecipitating the translation product with the anti-56 kDa antiserum and separating the precipitate on 2-D PAGE (Fig. 2B). Thus, the 56 kDa protein is not synthesized as a larger precursor. Covalent labeling of the 56 kDa protein with [3H]mibolerone We had previously shown that this protein could be specifically covalently labeled by photolysis in situ with the androgen analog [3H]methyltrienolone, but the efficiency of labeling was very low and required long exposure times of the fluorograms (Wrogemann et al., 1988). With the new synthetic androgen mibolerone (Bannister et al., 1985) we could achieve specific labeling of the 56 kDa protein with almost 10 times better efficiency than with methyltrienolone. The latter gave an efficiency of only approx. 0.125% (Wrogemann et

Fig. 2. In vitro translation of mRNA isolated from strain MCH 6 genital skin fibroblasts. A: Autoradiogram of total translation products labeled with [35S]methionine and separated on ZD-PAGE. B: Autoradiogram of immuneprecipitate of the translation products separated by 2D-PAGE. Arrowheads point to the 56 kDa protein doublets.

al., 1988). Labeling intact genital skin fibroblasts in situ on Petri dishes, followed by 2D-PAGE and fluorography, the labeled product could be de-

MCH6

-

acid

50 nM [;?-ilMibolerone

alkaline ----)

tected on the total protein maps after as little as 2 days’ exposure (Fig. 3A and B). Labeling of total GSF proteins with mibolerone

MCH6

-

50 nM ?HlMibolerone

acid

+200x

cold

alkaline -----)

3. Photoaffinity labeling with [3H]mibolerone of strain MCH 6 genital skin fibroblasts. A: Cells labeled with 50 nM [3H]mtbolerone m situ, separated by ZD-PAGE and exposed by fluorography. B: As in A, but with the inclusion of 200-fold excess radioinert mibolerone. Exposure: 44 h. Arrowheads point to the positions of the 56 kDa protein doublets.

56K ?HlMibolerone

Labelled

Fig. 4. Effect of mibolerone concentration on covalent labeling of the 56 kDa protein by photolysis. MCH 6 genital skin fibroblasts were labeled by photolysis in situ with increasing concentrations of [‘H]mibolerone i a 200-fold excess of radioinert ligand, followed by SDS-gel electrophoresis and fluorography.

is sufficiently specific and selective for the 56 kDa protein that the analysis can be done on one-dimensional SDS gels (Fig. 4) and this allowed us to study the concentrations of ligand required to saturate this protein with androgen under conditions of photolytic labeling. At 2 nM of mibolerone, which would saturate the prototypic androgen receptor (Gottlieb et al., 1987), labeling of the 56 kDa protein is observed, but competition with a 200-fold excess of radioinert mibolerone is not detected (Fig. 4). This appears to reflect a low affinity of the 56 kDa protein for androgen. Densitometric analysis of the fluorogram of Fig. 4, at different exposure times to avoid film saturation, reveals that half-maximal saturation was only achieved at approx. 4 PM mibolerone (Fig. 5).

-

acid

alkaline +

pM

[3H]-Mibolerone

Fig. 5. Saturation curve for covalent labeling of the 56 kDa protein with mibolerone. Densitometric scans of absorbancy peaks of the fluorogram of Fig. 4 expressed in arbitrary absorption units. Only lanes with 200Xcold mibolerone were scanned for this graph. Half-maximal saturation is reached at approx. 4 FM mibolerone.

Expression of the 56 kDa protein in cell strains from patients with complete androgen insensitivity syndrome To date we have studied 23 independent GSF cell strains which all express the 56 kDa protein. In contrast, 13 out of 15 strains from unrelated patients with complete androgen insensitivity syndrome due to receptor abnormalities do not express this protein (Nickel et al., 1988; Wrogemann et al., 1988; and unpublished data). Importantly, however, the remaining two mutant cell strains do express the protein in spite of no detectable non-

M--

acid

alkaline A

Fig. 6. Western blot analysis with the antiserum against the 56 kDa protein of two genital skin fibroblast cell strains from patients with complete androgen insensitivity syndrome. These cell strains are exceptional in that they have no androgen-binding activity but do express the 56 kDa protein. Arrowheads point to the 56 kDa protein doublets. A: Cell strain NHL. B: Cell strain TRL. The cell strains are described under Materials and Methods.

201

TRL 50 nM [%lMibolerone

TRL 50 nM [%llMibolerone

c-

t-

alkaline --)

acid

7. Photoaffinity

labeling

with [3H]mibolerone

acid

of genital skin fibroblast strain TRL (complete as described in Fig. 3.

covalent androgen receptor-binding activity. If the 56 kDa protein was a degradation product of the prototypic androgen receptor, one might expect the 56 kDa protein in these two mutant strains to be structurally altered. We used the antiserum against this protein to test this possibility. On Western blots of two-dimensional protein maps from these two patients (Fig. 6A and B) we found the protein to react normally in intensity, with no indication of altered size or charge or unusual crossreactive material. The splitting of the lefthand spot in cell strain TRL is due to a gel artifact and was not seen in repeat experiments. Binding of androgen to the 56 kDa protein of ceil strains from patients If the 56 kDa protein resulted from the degradation of the genuine androgen receptor, one would expect the binding of androgen to this protein to be altered in patients with complete androgen insensitivity syndrome and no androgen receptor-binding activity. We have found that the 56 kDa protein in GSF of the two exceptional patients of this category can be covalently labeled with methyltrienolone (Wrogemann et al., 1988; and unpublished data) equally well as the one

+200x cold

alkaline androgen

insensitivity).

Conditions

from GSF of control subjects with normal androgen receptor-binding activity. We show here that this also holds true for photolytic labeling with another synthetic androgen, [3H]mibolerone (Fig.

7). Discussion The 56 kDa protein described here is the same (Belsham et al., 1989) as the one of 58 kDa recently described by Kovacs and Turney (1988) and Kovacs et al., (1989), that was specifically radiolabeled with [3H]dihydrotestosterone 17pbromoacetate (DHT-BA). The protein shares a number of characteristics with the androgen receptor: (i) it is more enriched in GSF than in non-GSF (Kovacs and Turney, 1988; Nickel et al., 1988); (ii) it is missing in most GSF strains from patients with complete X-linked androgen insensitivity (Kovacs and Turney, 1988; Nickel et al., 1988; Wrogemann et al., 1988); and (iii) it binds androgen specifically (Kovacs and Tumey, 1988; Wrogemann et al., 1988). These features allowed Kovacs and Tumey (1988) to conclude that the protein could be the androgen receptor itself. This conclusion was further affirmed by recent studies

202

showing that 58 kDa protein affinity-labeled DHT-BA is indistinguishable under a variety of native conditions from androgen receptor complexed non-covalently with [3H]dihydrotestosterone (Kovacs et al., 1989). However, our previous findings (Nickel et al., 1988; Wrogemann et al., 1988) and especially the data presented here are not compatible with this interpretation because (i) the protein is too abundant, (ii) it is too small, (iii) its affinity for androgen appears low, and (iv) the GSF of two unrelated exceptional patients with complete androgen insensitivity and lacking androgen receptor-binding activity nevertheless express this protein in an apparently unaltered form as judged by all the criteria that define the protein in the GSF of normal individuals. The 56 kDa protein has been detected by staining methods in every normal GSF strain studied (23 strains so far). To detect the classical AR (maximal receptor-binding activity approx. 30 fmol/mg total GSF protein) by this technique in 100 p.g of total cell homogenates is clearly below the detection limit of even the most sensitive staining methods. Based on staining intensities of the 56 kDa protein we have estimated that in the GSF strain MCH 6 this protein is at least two orders of magnitudes more abundant than the androgen receptor. Indeed Kovacs et al. (1989) ,also find under saturating conditions and on the assumption of 100% affinity labeling efficiency, the 56/58 kDa protein binds lo- to 30-fold more androgen than the receptor can bind under noncovalent conditions. These findings are difficult to reconcile with the idea that the 56 kDa represents the androgen receptor, unless, of course, it is a more stable degradative product, which accumulates in the cell to much larger concentrations due to a slower turnover. Covalent labeling of the protein with the synthetic androgen [“Hlmibolerone shows that the 56 kDa protein is specifically labeled, but saturation requires micromolar concentrations of the ligand. This is in accord with our previous observation on the androgen analogue methyltrienolone (Wrogemann et al., 1988), but disagrees with the observation of high-affinity labeling with dihydrotestosterone 17P-bromoacetate (Kovacs and Turney, 1988). Because ligand molecules and receptor proteins may degenerate during the 10 min ex-

posure to UV light, it is possible that the photolytic covalent labeling underestimates the true affinity for the ligand. However, if the affinity of this protein was identical to that of the androgen receptor, binding studies should reveal a larger binding capacity than the 38 fmol/mg protein observed by non-covalent labeling in cell strain MCH 6 (Wrogemann et al., 1988). Similarly, if binding affinity was only somewhat lower, Scatchard analysis should reveal a second component of lower affinity in GSF; this is not observed at concentrations up to 40 nM (Keenan et al., 1984). Hence, most of the excess androgen-binding activity measured in GSF with DHT-BA compared to DHT must represent binding to a very low affinity androgen-binding protein that is not recognized by noncovalent DHT labeling. The androgen receptor cDNA has been cloned and sequenced recently (Chang et al., 1988; Lubahn et al., 1988; Trapman et al., 1988). The data suggest a molecular mass of 94 kDa for the full-length receptor monomer (Chang et al., 1988). Transcription/translation of the complete cDNA produced, apart from the 94 kDa product, peptides of 76, 70, 55, 46, 32 and 30 kDa, all immuneprecipitable with an autoantibody to the AR (Chang et al., 1988). The 94, 76, 70 and 46 kDa are explainable as products of translation from different initiation codons, but it is not yet clear, whether these do occur in vivo. The 55, 32 and 30 kDa forms could have been proteolytic fragments of larger forms (Chang et al., 1988). From other recent studies using purification procedures and photoaffinity labeling techniques it is emerging, however, that the classical species of androgen receptor monomer is clearly larger than the 56 kDa protein, and the values reported between 90 and 110 kDa (Johnson et al., 1987; Brinkmann et al., 1988; Gyorki et al., 1988; van Loon et al., 1988; Mowszowicz et al., 1989; Mulder et al., 1989) correspond fairly well with the full-length coding sequence observed. However, the 56/58 kDa protein described in this paper is unlikely to be a stable degradation product of a larger rare protein, the AR, because its quantity and its size could not be influenced by harvesting cells directly in electrophoresis lysis buffer (Nickel et al., 1988). by the inclusion of antiproteolytic cocktails (Wrogemann et al., 1988), in agreement with

203

Kovacs et al. (1989), or by cell fractionation procedures (data not shown). Using pulse-labeling techniques we were unable to detect any indications of grossly lower turnover rates of this protein compared to other proteins seen on twodimensional GSF protein maps (data not shown). In addition, the clear demonstration of its direct synthesis by in vitro translation makes it highly unlikely that the protein arises proteolytically under all these conditions. As the intact androgen receptor monomer is larger than the 56 kDa protein, one can also not consider the latter a possible precursor for the receptor (Kovacs et al., 1989), given our current understanding of the biosynthesis of protein molecules in general, and of the steroid/thyroid family of receptor proteins specifically. The most compelling evidence in support that the androgen receptor and 56 kDa protein are distinct molecules probably without a precursorproduct relation comes from the study of two mutant CA1 cell strains. While the 56 kDa protein is generally not expressed in CA1 cell strains (Kovacs and Turney, 1988; Nickel et al., 1988; Wrogemann et al., 198X), two exceptional cell strains do express it. In itself this indicates a new level of heterogeneity among genetic lesions of the androgen receptor gene. In this context, however, it is important to see that the mutant cell strains have negligible androgen receptor-binding activities (Nickel et al., 1988), yet the protein appears as determined by its structurally unaltered, immunoreactivity, its size and charge, and most importantly, its ability to be covalently photoaffinity labeled with two synthetic androgens just like the protein in wild-type cell strains. The antiserum used in this study was prepared against denatured protein spots of two-dimensional gels. The antiserum fails to precipitate native receptor preparations (not shown). However, this does not exclude the possibility that our protein shares structural similarities with the AR. In fact, occasional Western blots of 2D-PAGE maps revealed a faint doublet of approx. 110 kDa (not shown), which might be the full-length AR monomer. We have previously reported that the 56 kDa protein is not androgen induced, and it is not the AR itself or a degradation product of the AR as shown here, yet, with specific exceptions, it re-

quires an intact AR gene for its expression. To accommodate these findings best, we have postulated that this protein might be synthesized from the AR gene as well, but as a different protein, involving processes, which may include separate promoters, differential splicing, different polyadenylation sites (Wrogemann et al., 1988) or, as recently found for members of the steroid/thyroid receptor supergene family, transcription from the opposite strand (Lazar et al., 1989). It could thus be another member of the growing supergene family of steroid/thyroid hormone receptors (Evans, 1988). Cloning and sequencing of the cDNA of the 56 kDa is required to answer this question conclusively. With the antiserum described here for the 56 kDa protein this project has now become feasible. Acfcnowledgements This work was supported by the Manitoba Health Research Council and the Children’s Hospital of Winnipeg Research Foundation and the Medical Research Council of Canada Group Grant in Medical Genetics. References Bannister, P., Sheridan, P. and Lososwsky. M.S. (1985) J. Steroid B&hem. 23, 121-123. Belsham, D.D., Rosenmann, E., Pereira, F-A., Williams, S.G., Tumey, M.K., Kovacs, W.J.. Faber. L.E. and Wrogemann, K. (1989) J. Steroid Biochem. 33 (in press). Banner, W.M. and Laskey, R.A. (1974) Eur. J. Biochem. 46, 83-88. Bradford, M. (1976) Anal. B&hem. 72, 248-249. Brinkmann, A.O., Kuiper, G.G., Bolt-de Vries, J. and Mulder, E. (1988) J. Steroid Biochem. 30, 257-261. Chang, C., Kokontis. J. and Liao, S. (1988) Proc. Natl. Acad. Sci. U.S.A. 85, 7211-7215. Chirgwin, J.M., Przybyla. A.E.. MacDonald, R.J. and Rutter, W.J. (1979) Biochemistry 18, 5294-5299. Evans, R.M. (1988) Science 240, X89-895. Faber, P.W., Kuiper, G.G.J.M., van Rooij. H.C.J., van der Korput, J.A.G.M.. Brinkmann, A.O. and Trapman, J. (1989) Mol. Cell. Endocrinol. 61, 257-262. Gottlieb, B., Kaufman, M.. Pinsky, L., Leboeuf, G. and Sotos, J.F. (1987) J. Steroid Biochem. 28, 279-284. Gyorki, S., Wame, G.L. and Funder. J.W. (1988) J. Steroid Biochem. 29, 611-615. Johnson, M.P.. Young, C.Y. Rowley, D.R. and Tindall, D.J. (1987) Biochemistry 26, 3174-3182. Kaufman, M., Straisfield, C. and Pinsky, L. (1976) J. Chn. Invest. 58, 345-350.

204 Kaufman, M., Pinsky, L., Baird, P.A. and McGillivray, B.C. (1979) Am. J. Med. Genet. 4, 401-411. Keenan, B.S., Greger, N.G.. Hedge, A.M. and McNeel, R.L. (1984) Steroids 43, 159-178. Kovacs, W.J. and Turney, M.K. (1988) J. Clin. Invest. 81, 342-348. Kovacs, W.J., Turney, M.K. and Skinner, M.K. (1989) Endocrinology 124, 1270-1277. Laemmli, U.K. (1970) Nature 227, 680-685. Lazar, M.A., Hodin, R.A., Darling, D.S. and Chin, W.W. (1989) Mol. Cell. Biol. 9, 1128-1136. Lubahn, D.B., Joseph, D.R., Sar, M., Tan, J., Higgs, H.N., Larson, R.E., French, F.S. and Wilson, E.M. (1988) Mol. Endocrinol. 2, 1265-1275. Maniatis, T., Fritsch. E.F. and Sambrook, J. (1982) Molecular Cloning. A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Morrissey, J.H. (1981) Anal. Biochem. 117, 307-310. Mowszowicz, I.. Stamatiadis, D., Wright, F., Kuttenn. F. and Mauvais Jarvis, P. (1989) J. Steroid B&hem. 32, 157-162. Mulder, E., van Loon, D..’ de Boer, W., Schuurmans, A.L.. Bolt. J., Voorhorst, M.M., Kuiper, G.G. and Brinkmann, A.O. (1989) J. Steroid B&hem. 32, 151-156.

Nickel, B., Schwartz, A., Rosenmann, E., Kaufman, M., Pinsky, L. and Wrogemann, K. (1988) Clin. Invest. Med. 11, 22-33. O’Farrell. P.H. (1975) J. Biol. Chem. 250, 4007-4021. Perdew, G.H., Schaup. H.W. and Selivonchick. D.P. (1983) Anal. Biochem. 135. 453-455. Rosenmann, E.. Kreis. C.. Thompson, R.G., Dobbs, M., Hamerton, J.L. and Wrogemann. K. (1982) Nature 298. 563-565. Thompson, R.G., Nickel, B., Finlayson. B., Meuser, R., Hamerton, J.L. and Wrogemann. K. (1983) Nature 304. 740-741. Trapman, J.. Klaassen, P., Kuiper. G.G.J.M., van der Korput, J.A.G.M., Faber, P.W., van Rooij. H.C.J.. van Kessel, A.G., Voorhorst, M.M.. Mulder, E. and Brinkman, A.O. (1988) Biochem. Biophys. Res. Commun. 153, 241-248. van Loon, D., Voorhorst, M.M.. Brinkmann, A.O. and Mulder, E. (1988) Biochem. Biophys. Acta 970, 278-286. Wrogemann, K., Pereira, F., Belsham, D., Kaufman, M., Pinsky, L. and Rosenmann, E. (1988) Biochem. Biophys. Res. Commun. 155, 907-913.