Regulation and Ligand-binding Specificities of Two Sex-specific Bile Acid-binding Proteins of Rat Liver Cytosol*

Rat liver cytosolic proteins were photoaffinity labeled with the synthetic steroid [3H]methyltrienolone in order to identify and characterize hepatic proteins that may participate in the intracellular binding and transport of steroid hormones and other sterols. A male-specific and a female-specific sterol-binding protein (SBP) that migrated to the 4 S region of a sucrose gradient and had similar molecular weights (male-specific 34-kDa protein (SBP34), female-specific 31-kDa protein (SBP31] were thus identified. Experiments were undertaken to determine the biochemical basis for the sex-specific expression of these two proteins. In vivo hormonal manipulations established that the female-specific expression of SBP31 could, in part, be accounted for by the suppressive effects of androgen on SBP31 levels in male rats. In contrast, androgen stimulated expression of the male-specific SBP34, while estrogen and the estrogen-regulated continuous plasma growth hormone profile that is characteristic of adult female rats were suppressive toward this protein. Unlike several other androgen-dependent hepatic proteins, however, SBP34 did not require an intact pituitary for androgen-stimulated expression, nor was its expression stimulated by the intermittent pulses of plasma growth hormone that are characteristic of adult male rats. SBP34 and SBP31 were not induced but were suppressed to various extents by dexamethasone, phenobarbital, and clofibrate, drugs that are known to induce other hepatic proteins involved in steroid binding and metabolism. Competition experiments revealed that SBP31 has a relatively broad ligand specificity, with significant competition for [3H]methyltrienolone binding exhibited by bile acids (chenodeoxycholic acid and lithocholic acid) and a range of steroid hormones (progesterone, estradiol, testosterone, and 5 alpha-dihydrotestosterone) when present in the low micromolar range. No binding was detected with this protein toward cholesterol, triamcinolone acetonide, 5 alpha-androstan-3 alpha,17 beta-diol, cholic acid, and deoxycholic acid. In contrast, SBP34 exhibited greater binding specificity, with competition for [3H]methyltrienolone binding observed only with primary bile acids (cholic acid and chenodeoxycholic acid) and their metabolites (deoxycholic acid and lithocholic acid). On the basis of these binding specificities and the relatively high concentration of bile acids found in the liver, it is proposed that SBP31 and SBP34 function in the intracellular binding and/or transport of bile acids.

tion of bile acids found in the liver, it is proposed that SBP31 and SBP34 function in the intracellular binding and/or transport of bile acids.
The liver is a major site for both the anabolic and catabolic metabolism of sterols including cholesterol, steroid hormones, and bile acids. For many of these sterols, extra-and intrahepatic transport is mediated by carrier proteins. For example, cholesterol is transported to the liver bound to plasma lipoproteins, its entry into the liver is mediated by specific cell surface receptor proteins, and its synthesis in liver is dependent upon precursor-specific carrier proteins (Scallen et al., 1974;Gill et al., 1985). Proteins have also been identified that function in the plasma transport, hepatic influx, and hepatic efflux of steroid hormones and bile acids (Petra et al., 1988;Ananthanarayanan et al., 1988;Ruetz et al., 1987). In contrast, intrahepatic transport processes for these compounds are poorly understood and the proteins that mediate intrahepatic transport, although inferred experimentally (Myant and Mitropoulos, 1977;Kitani and Kanai, 1981;Stolz et al. 1989;Answer et al., 1976), remain largely uncharacterized.
Although sterol-binding activity has been demonstrated in cytosolic preparations from liver and other tissues, in many cases it has been difficult to ascertain whether the observed binding activity represents cytosolic receptor proteins or proteins that elicit some other sterol-binding function (Roy et al., 1983;Powell-Jones et al., 1980;Kyakumoto et al., 1984;Eagon et al., 1989). Multiple sterol-binding proteins appear to be expressed in hepatic tissue. For example, binding proteins have been identified that migrate to -4 S or to -8 S in sucrose gradients (Powell-Jones et al., 1980;Kyakumoto et al., 1984). Moreover, sterol-binding activity has been reported that is male-specific (Powell-Jones et al., 1980;Dickson et al., 1978;Sarkar et al., 1987), female-specific (Sarkar et al., 1987), or present in both sexes (Dickson and Eisenfeld, 1979). The identify of the proteins responsible for these binding activities and their physiological functions remain largely unknown. Several characteristics can be ascribed to putative intracellular sterol-transport proteins that may facilitate their identification and subsequent characterization. 1) Cellular localization: These proteins would be localized in the cytosol owing to their function in intracellular sterol transport.
2) Sterolbinding capability: These proteins would have the capacity to bind sterols reversibly and with some degree of specificity. 3) Coregulation with sterol-metabolizing enzymes: The liver is sexually dimorphic with respect to many sterol-metabolizing enzymes (Gustafsson et al., 1983). Sterol-bindingproteins that function in concert with these enzymes might also exhibit sexual dimorphism due to hormonal regulatory controls. 4) Responsiveness to drugs: Many hepatic sterol-metabolizing enzymes can be induced or suppressed following in uiuo treat-5654 ment with drugs and other xenobiotics (Waxman, 1986). Induction of these enzymes can facilitate drug activation and inactivation as well as the detoxification and excretion of toxic compounds. Sterol-binding proteins might also participate in the binding of lipophilic drugs and other foreign compounds and therefore may be responsive to drug exposure.
In the present study, we have utilized the synthetic androgen [3H]methyltrienolone in conjunction with photoaffinity labeling techniques to identify hepatic proteins that may function in intracellular sterol transport. A detailed characterization of the localization, regulation by hormones and drugs, and sterol-binding specificities of a male-specific 34-kDa protein and a female-specific 31-kDa protein is presented. Although these proteins were identified by photoaffinity labeling with an androgen, ligand competition experiments revealed that they exhibit greater binding affinity toward bile acids, suggesting that they may serve as intracellular bile acid carrier proteins. (87 Ci/mmol, Du Pont-New England Nuclear) in cytosol buffer containing no glycerol (PL buffer) in a total volume of 57 ~1. Cytosolic methyltrienolone-binding proteins were then photoaffinity labeled using an apparatus similar to that described by Katzenellenbogen et al. (1974). Essentially, a 450-watt mercury vapor lamp was suspended in a glass chamber equipped with a cooling jacket of circulating water (4 "C). This chamber was inserted into a larger glass container filled with a solution of saturated copper sulfate, which served to filter emitted wavelengths <315 nm. The entire apparatus was immersed into the center of a bath of ethylene glycol:water (50:50) that was maintained at 4-6 "C by a jacket of circulating methanol (0 "C). Incubation tubes were suspended into this bath in positions surrounding the lamp and the samples irradiated for 10 min. Preliminary experiments revealed that maximal photolabeling was achieved by 5-7 min. Samples were then incubated for 10 min with an equal volume of dextran-coated charcoal (0.25 g of acid-washed charcoal, 0.025 g of dextran T70 in 50 ml of PL buffer).

Animals
The charcoal was then pelleted in a microcentrifuge and the supernatant containing the photolabeled proteins resolved by SDS-PAGE using 10% acrylamide gels at 25 mA/gel in a Mighty Smalls apparatus (Hoefer Scientific). Proteins were transferred from the gels to nitrocellulose at 120 V for 1 h and the nitrocellulose subsequently dried in an oven for 20 min at 80 "C. The nitrocellulose was sprayed with Enhance@ (Du Pont-New England Nuclear) and the photolabeled proteins visualized after exposure to x-ray film for 3-4 days. Radiolabeled protein band intensity on the x-ray film, which was used as an indicator of the relative amount of the protein present in the sample, was quantitated by laser densitometry . Densitometric quantitations of the radiolabeled protein bands were found to be linear for at least a range of 0.25-4 times the standard load of cytosolic protein from an untreated adult male rat.
Competition binding studies were performed using this same general procedure except that unlabeled competitor (dissolved in ethanol) was added to each tube at the desired concentration and the solvent evaporated under a stream of nitrogen before adding the other constituents.
Density Gradient Centrifugation-Continuous 5-20% (w/v) sucrose gradients were prepared in PL buffer. Photoaffinity labeled cytosol preparations (200-400 ~1 at 10 mg/ml) were layered onto the top of the gradients (5 ml) and the samples centrifuged at 100,000 x g for 2.5 h. Bovine serum albumin (4.6 S) and rabbit IgG (7.1 S) were used as sedimentation markers.
The gradient was then fractionated into 200-~1 aliquots, 150 ~1 of which was used to measure radioactivity by scintillation counting and the remaining sample analyzed by SDS-PAGE and fluorography, as described above, to localize the individual radiolabeled proteins within the gradient.

Identification of SBPs by [3H]Methyltrienolone
Photoaffinity Labeling-Photoaffinity labeling of cytosol from adult male rat liver with [3H]methyltrienolone revealed three major binding proteins having apparent molecular sizes of 27 kDa (SBP27), 34 kDa (SBP34), and 45 kDa (SBP45) (Fig. 1, lane   1). In contrast, in experiments using cytosol from female rats, although the proteins of 27 and 45 kDa were also observed, no 34-kDa protein was detectable. Moreover, a methyltrienolone-binding protein of 31 kDa (SBP31) that was undetectable in the male was observed as a major binding protein in the female cytosol (Fig. 1, lane 8). These same binding patterns were obtained at methyltrienolone concentrations ranging from 10 to 150 nM, (data not shown). In control experiments, photoaffinity labeling of rat serum preparations revealed a single major methyltrienolone-binding protein of M, -70 kDa, indicating that the cytosolic SBPs are not serumderived contaminants (data not shown). The 70-kDa serum protein, which was also detected at a low (albeit variable) level in the cytosol preparations, is presumed to be albumin. No radioactive bands at these molecular weights were detected when microsomal preparations isolated from rat liver were photolabeled with [3H]methyltrienolone, thus demonstrating that the photolabeled SBPs are cytosolic proteins, and not contaminants of the microsomal fraction. Further support for the specific nature of SBP photolabeling was provided by the Rat Hepatic Bile Acid-binding Shown is the response of SBP31 and SBP34 to hormonal alterations. Cytosolic proteins were photoaffinity labeled with [YH]methyltrienolone, subjected to SDS-PAGE, transferred to nitrocellulose, and visualized by fluorography as described under "Materials and Methods." Individual lanes contain liver cytosol prepared from the following rat treatment groups: lane I, unaltered adult male rat (UT); lane 2, castrated male 3 weeks after surgery (CX); lane 3, castrated male that received daily androgen (AND)  Density Gradient Centrifugation-Centrifugation of photolabeled liver cytosolic proteins through sucrose gradients revealed that the majority of binding activity migrated to -4 S in both male and female cytosols ( Fig. 2A, fractions 5-7), with a small shoulder of binding activity also detected at -8 S ( Fig.  2A, fractions 13 and 14). SDS-PAGE and autoradiography of individual fractions of the sucrose gradient revealed that SBP27 and SBP34 from male rat cytosol migrated to -4 S while SBP45 migrated to -8 S (Fig. 2B). In female rat preparations, SBP27 and SBP31 migrated to -4 S and SBP45 migrated to -8 S (data not shown). These results demonstrate that while a single protein is responsible for the majority of ["Hlmethyltrienolone protein binding in the 8 S region of a sucrose gradient, multiple proteins contribute to the observed binding in the 4 S region.
Steroid Hormone Dependence of SBP34 and SBP31-The differential expression of SBP34 and SBP31 in male and female rats prompted an examination of the involvement of gonadal hormones in the expression of these hepatic proteins. Ablation of circulating androgen levels by castration of adult male rats led to a progressive loss of SBP34 while having little consistent effect on SBP27 and SBP45 (Fig. 1, lane 2, Fig.  3A). Treatment of the castrated rats with androgen restored near normal SBP34 levels (Fig. 1, lane 3; Fig. 3A). High level expression of SBP34 was also obtained in ovariectomized female rats treated with androgen (Fig. 30). Conversely, SBP34 levels in unaltered male rats were markedly suppressed following treatment with the estrogen diethylstilbesterol (Fig.  1, lane 7; Fig. 3A). Thus SBP34 expression is positively regulated by androgen and can be suppressed by estrogen.
Expression of the female-specific SBP31 was suppressed by androgen, as was demonstrated by the loss of this protein in ovariectomized female rats administered androgen for 1 week (Fig. 3C) and by its partial expression in castrated males (-20% of normal female level 9 weeks after castration) (Fig.  3B) A, radioactivity associated with individual gradient fractions from both male (0) and female (0) rat liver cytosol. Gradient markers are bovine serum albumin (4.6 S) and rabbit IgG (7.1 S). B, densitometric quantitation of individual photoaffinity labeled proteins from male rat cytosol present in selected gradient fractions. The individual proteins that constituted the binding in the 4 S and 8 S regions of the gradient in A were examined by subjecting individual fractions from the sucrose gradient to SDS-PAGE, transferring the separated proteins to nitrocellulose, subjecting the nitrocellulose to fluorography, and finally measuring band intensities on the resultant x-ray film by laser densitometry. The locations of the individual SBP proteins in the gradient are indicated by the following: SBP27 (O), SBP34 (A), SBP45 (A). In addition, a significant percentage of the radioactivity associated with fractional-9 could be attributed to low molecular weight [3H]methyltrienolone-binding material that migrated with the buffer front during electrophoresis (0). Parallel analysis of female rat liver cytosol revealed that SBP31 migrated in the 4 S region (data not shown).
loss of SBP31 (Fig. 3C), indicating that estrogen does not play an obligatory role in the expression of SBP31 in female rats. These experiments demonstrate that the sex specificities of SBP34 and SBP31 are due, at least in part, to their differential regulation by steroid hormones. In contrast, SBP27 and SBP45 were not significantly affected by these hormones (Fig. 1 and data not shown) and therefore are expressed in both sexes.

Regulation of SBP34 and SBP31 by Pituitary Hormones-
Since many of the effects of androgens and estrogens on hepatic protein expression are mediated by the hypothalamicpituitary axis (Gustafsson et al., 1983), the involvement of pituitary hormones, particularly growth hormone, in the steroid hormone dependence of SBP34 and SBP31 was examined.
SBP34 is dependent upon pituitary factors for full expression as evidenced by its depression in male rats following hypophysectomy (Fig. 1, lane 4; Fig. 4A, lane 3). The magnitude of loss of SBP34 following hypophysectomy was variable among individual rats from several experiments (n = 10) with an average loss of -50%. Growth hormone does not appear to contribute to SBP34 expression since twice daily injections of the hormone, intended to mimic the pulsatile plasma growth hormone pattern characteristic of male rats, had no effect on SBP34 levels (Fig. 4A, lane 4). In contrast, daily androgen administration to hypophysectomized male rats re- sulted in a substantial restoration of SBP34 after 1 week (Fig.   1, he 6; Fig. 4A, lane 6) and fully restored this protein after 2 weeks (data not shown). Similar results were obtained when endogenous androgen production was stimulated in the hypophysectomized male rats by human chorionic gonadotropin (150 IV/kg subcutaneously, daily for 7 days). Thus, in contrast to other androgen-dependent rat hepatic proteins (Gustafsson et al., 1983), SBP34 can be stimulated by androgen by a mechanism that does not require the participation of other pituitary-dependent hormones. Moreover, the loss of SBP34 following hypophysectomy seems to be due to the loss of gonadotropins and ultimately the depletion of circulating androgen.
While androgen, and not intermittent growth hormone, seems to be the pituitary-dependent factor that contributes positively to the expression of SBP34 in male rats, continuous growth hormone exposure can exert a suppressive influence on SBP34 levels. Continuous plasma growth hormone levels, characteristic of adult female rats and achieved in our experiments by use of an osmotic minipump (see "Materials and Methods"), severely suppressed this protein in both hypophysectomized and unaltered male rats (Fig. 1, lane 5; Fig. 4A, lanes 2 and 5).
Female-specific SBP31 was markedly suppressed by hypophysectomy, indicating that it also requires pituitary-dependent factors for full expression (Fig. 1, lane 9; Fig. 4B, lane 8). Continuous infusion of growth hormone to hypophysectomized female rats led to a small increase in the level of this protein, as was also observed in unaltered and hypophysec-tomized male rats treated continuously with growth hormone under the same conditions ( Fig. 1, lanes 5 and 10; Fig. 4B, lanes 2, 5, and 9). This finding suggests that other pituitarydependent hormones may be required to act in concert with growth hormone to achieve full expression of SBP31.
Modulation of SBP34 and SBP31 by Drug Treatment-The expression of many hepatic enzymes involved in sterol metabolism can be modulated following exposure to drugs or other foreign compounds. If SBP34 and SBP31 function in concert with sterol-metabolizing enzymes, then their expression might also be influenced by such drugs. Therefore, the effect on SBP levels of several drugs that are known inducers of hepatic sterol-metabolizing enzymes was evaluated. The male-specific SBP34 was not induced by phenobarbital, dexamethasone, or clofibrate in either sex; rather, this protein was variably suppressed by these compounds, with dexamethasone effecting complete suppression of this protein in males (Fig. 5A). Analysis of serum testosterone levels in the drugtreated rats revealed that dexamethasone also eliminated circulating levels of this hormone (Table I). In view of the androgen dependence of SBP34 (Fig. 3A), it seems likely that the loss of circulating androgen in these animals is responsible for the dexamethasone-induced loss of SBP34. The femalespecific SBP31 was not elevated by any of the drugs tested, although it was partially suppressed in females by phenobarbital (Fig. 5B).
Thus, none of the compounds examined (2%) to drug treatments in male and female rats. Drugs were administered as described under "Materials and Methods" and individual SBPs quantitated using photoaffinity labeling as described in Fig. 1. UT, untreated rats; PB, phenobarbital-treated rats; DEX, dexamethasonetreated rats; CF, clofibrate-treated rats. Brackets, S.E. response from three to four rats per treatment group.  = 3). Values determined by radioimmunoassay as described previously (LeBlanc and .
induced expression of either SBP34 or SBP31 in liver cytosol.
Ligand-binding Specificity of SBPs-The ligand-binding specificities of SBP34 and SBP31 were examined in competition experiments using both anabolic and catabolic sterol metabolites. Included were cholesterol, triamcinolone acetonide (a synthetic glucocorticoid) and progesterone, the gonadal steroids (testosterone, estradiol) and some of their metabolites (5a-dihydrotestosterone, 5a-androstan-3a,l7@diol), and both primary bile acids (cholic acid, chenodeoxycholic acid) and secondary bile acids (deoxycholic acid, lithocholic acid). Female-specific SBP31 exhibited no significant binding affinity toward cholesterol and triamcinolone acetonide as evidenced by the inability of these steroids to compete with [3H]methyltrienolone when included in the photolabeling mixture at 12 PM, corresponding to a IOO-fold molar excess over methyltrienolone (Fig. 6A). In contrast, progesterone was 75% effective at reducing the labeling of SBP31 by [3H] methyltrienolone under comparable incubation conditions (Fig. 6A). The sex steroids (testosterone, 5a-dihydrotestosterone, and estradiol) were moderately effective competitors, with -50% inhibition of [3H]methyltrienolone incorporation at 12 pM, while the secondary testosterone metabolite 5a- Cytosols from adult female rat liver were incubated with, then photoaffinity labeled by 120 nM [3H] methyltrienolone in the presence of 1.2, 6.0, or 12 pM unlabeled steroid compounds. Binding of the steroid compounds to SBP31 was revealed by the reduced labeling of the SBP by [3H]methyltrienolone. The relative level of photoaffinity labeling was then quantitated as described in Fig. 1. Ligands used in each experiment were as follows.  Analyses were carried out as described in Fig. 1 using cytosols from adult male rat liver. The same unlabeled steroid compounds were used as described in Fig. 6. androstan-3a,l7@-dial was ineffective (Fig. 6B). SBP31 did not exhibit significant binding affinity for the primary bile acid cholic acid or its metabolite deoxycholic acid (Fig. 6C). However, chenodeoxycholic acid, and to a somewhat lesser extent its metabolite lithocholic acid, inhibited [3H]methyltrienolone incorporation in the low micromolar range (Fig.  6C). Chenodeoxycholic acid, the more effective competitor, is found in female rat liver cytosol at a concentration of -40 PM, while lithocholic acid is present at a somewhat lower level (Kurtz et al., 1982). This suggests that these bile acids might serve as physiologically significant ligands for SBP31. Male-specific SBP34 exhibited little binding affinity toward cholesterol, triamcinolone acetonide, progesterone, testosterone and its derivatives, and estradiol ( Fig. 7, A and B).
However, this protein did effectively bind the primary bile acids cholic acid and chenodeoxycholic acid, with -80% inhibition of [3H]methyltrienolone binding at a concentration of 6 FM (Fig. 7C). SBP34 bound the secondary bile acid deoxycholic acid with similar affinity as the primary bile acids (Fig. 7C). Lastly, SBP34 also bound lithocholic acid, but less effectively than the other bile acids, particularly at the lower concentrations tested. These results suggest that SBP34 is also a major binder of bile acids. However, SBP34 differs from SBP31 in its much higher affinity for cholic acid and its lower affinity for progesterone and the gonadal steroids.

DISCUSSION
The present study establishes that male and female rats each express a unique hepatic protein with the properties of a cytosolic bile acid-binding protein. These sex-dependent proteins, which are of similar size and have similar sedimen-tation characteristics, are differentially regulated by steroid and peptide hormones. It is this differential hormone regulation that confers the sex-specific expression that is characteristic of these hepatic proteins.
The expression of the male-specific SBP34 appears somewhat analogous to that of a male-specific liver microsomal steroid 16a-hydroxylase cytochrome P-450 enzyme, designated P-4502, (P-450 gene product IICll).
Like SBP34, P-4502, is dependent upon androgen for full expression (Waxman et al., 1985), and is also suppressed by several drugs that are known to induce other hepatic proteins (Dannan et al., 1983;Yeowell et al., 1987). However, the regulatory mechanisms that dictate the male-specific expression of these two hepatic proteins apparently differ. P-450*, expression is largely dependent upon pulsatile growth hormone secretions, and the role of androgen in the expression of this protein is believed to reflect the stimulation by androgen of a pulsatile growth hormone secretory profile (Morgan et al., 1985;Kato et al., 1986;Jansson and Frohman, 1987). SBP34 expression, however, is unresponsive to pulsatile growth hormone and, in addition, its responsiveness to androgen does not require an intact pituitary gland. Both proteins, however, are suppressed by estrogen and by continuous plasma growth hormone levels (this study and Morgan et al., 1985).
Like SBP34, SBP31 is also under multiple hormonal regulatory controls. The expression of this female-specific protein is suppressed by androgen and appears to be positively influenced by growth hormone and perhaps by estrogen as well. The pattern of regulation of this protein by hormones is similar to that of the female-predominant microsomal enzyme steroid 5a-reductase, whose expression is partly derepressed by castration of adult male rats, and can be stimulated (albeit to a fuller extent that SBP31) by continuous growth hormone treatment of unaltered males or hypophysectomized male and female rats (Mode et al., 1981;Waxman et al., 1989a).
Other investigators have identified sterol-binding activity associated with proteins that have sucrose gradient sedimentation characteristics similar to the sterol-binding proteins identified in the present study (Roy et al., 1983;Powell-Jones et al., 1980). Roy et al. (1983) characterized an androgen-and estrogen-binding activity from rat liver cytosol that sediments at -3.5 S and was proposed to be a receptor for 5a-dihydrotestosterone. The protein(s) responsible for this binding activity exhibited high affinity toward testosterone, 5a-dihydrotestosterone, and 17&estradiol and appeared to be expressed only in male rats.
Similarly, Powell-Jones et al. (1980) reported a male-specific protein with high binding affinity for both androgens and estrogens. This binding activity was greatly reduced in cytosol from hypophysectomized male rats, but was elevated in cytosol from hypophysectomized female rats. Unfortunately, since these characterizations only reflect changes in total ligand-binding activity in a crude cell fraction, it is impossible to ascertain whether a single or multiple proteins contribute to the observed ligand-binding profiles. Nonetheless, competition binding experiments performed in the present study with SBP34 indicate that testosterone, 5a-dihydrotestosterone and estradiol are relatively poor ligands for this protein, suggesting that SBP34 is probably not responsible for the male-specific binding activity characterized in the earlier studies.
SBP34 expression in male rat liver was suppressed by the estrogen diethylstilbesterol, while the antiandrogen hydroxyflutamide (SCH 16423) had little effect on the level of SBP34 expression.* These observations are similar to those made with a male estrogen-binding protein designated MEB (Eagon et al., 1989). Other similarities between SBP34 and MEB are their male specificity and their decreases following castration and hypophysectomy.
However, these proteins are apparently distinct insofar as MEB does not bind methyltrienolone (Eagon et al., 1989) and has a high binding affinity toward estradiol (Eagon et al., 1989) that is apparently not shared by SBP34.
Recently, Demyan et al. (1989) purified a 31-kDa cytosolic androgen-binding protein, designated CAB, that is expressed in male but not female rat liver. Earlier studies by the same investigators revealed the existence of a female-specific 29-kDa cytosolic androgen-binding protein in rat liver (Sarkar et al., 1987). Since both proteins were visualized using the [3H]methyltrienolone photoaffinity labeling techniques employed in the present study, it is concluded that they are, in fact, the same as the male-specific SBP34 and the femalespecific SBP31 described in the present study and that the apparent differences in size in the two studies (CAB at 31 kDa uersus SBP34 at 34 kDa; female-specific 29 kDa uersus SBP31 at 31 kDa) reflect interlaboratory differences in the calibration of the SDS-PAGE systems rather than true differences in molecular mass. The physiological significance of these proteins has not been established. However, the malespecific CAB protein (i.e. SBP34) appears to be coregulated with a2u-globulin (Sarkar et al., 1987), a major urinary protein synthesized in male but not female rat liver (Roy et al., 1983).
The CAB protein has been proposed to be a mediator of the androgen-dependent expression of &u-globulin (Sarkar et al., 1987) and was suggested to be related to the androgen-binding domain of the androgen receptor (Demyan et al., 1989). The present demonstration of a significantly higher binding affinity of this male-specific protein toward bile acids as compared to androgens suggests, however, that it is unlikely to carry out androgen-dependent functions in the hepatocyte. Competition binding experiments revealed that SBP34 and SBP31 both bind bile acids with comparatively high affinity. This observation, in the context of the high levels of bile acids found in the liver, strongly suggests that bile aids are the principal endogenous ligands for these proteins. The two sexspecific SBPs do differ, however, in their relative affinities for individual bile acids. The male-specific SBP34 was capable of binding both primary bile acids and their metabolites with high affinity, but exhibited no appreciable binding toward the various steroid hormones tested. In contrast, the femalespecific SBP31 exhibited high binding affinity toward the primary bile acid chenodeoxycholic and its metabolite lithocholic acid, but did not bind cholic acid or its metabolite deoxycholic acid to any significant extent. SBP31 also exhibited significant binding affinity toward progesterone and to a lesser degree, other steroid hormones. Considering the higher circulating levels of progesterone found in mature female rats as compared to males, SBP31 might also function in some progesterone binding capacity in the female rat. It is likely that the competition experiments carried out in this study underestimate the true binding affinities of the SBPs for the bile acids, insofar as the bile acids are bound reversibly, while the photolabeling with [3H]methyltrienolone is covalent and effectively irreversible. Future studies with purified SBPs may provide an estimate of ligand binding affinities which will facilitate a more definitive identification of the physiologically relevant ligands of these proteins.
The physiological significance of sex-specific hepatic bile acid-binding proteins is currently unknown. However, male and female rat liver cytosols are known to have significantly different levels of individual primary and secondary bile acids (Kurtz et al., 1982). Furthermore, sex differences have been observed for a number of enzymes that carry out bile acid metabolism, including those active in sulfate conjugation (Barnes et al., 1979), hydroxylation (Yousef et al., 1973) and oxidoreduction (Bjorkhem et al., 1973). Thus, sex specificity with regard to ligand carrier proteins is not unexpected. Further investigation is required to address critical questions regarding the functions of these proteins. This includes a more thorough evaluation of the ligand specificities of these proteins, the extent to which they might confer protection from bile acid toxicity, and their potential roles in the metabolism of bile acids or in mediating their biological actions, such as the feedback inhibition of cholesterol 7a-hydroxylase, a key enzyme of the bile acid biosynthetic pathway (Myant and Mitropoulos, 1977). Further characterization of these proteins may ultimately lead to a fuller understanding of their activities as well as the factors that contribute to pathological conditions related to bile acid transport and metabolism (Borgstrom et al., 1985;Bjorkhem, 1985).