Biochemical Homology between Rat Dorsal Prostate and Coagulating Gland PURIFICATION OF A MAJOR ANDROGEN-INDUCED PROTEIN*

The anatomically distinct organs, rat dorsal prostate and coagulating gland, were found to display remark- able homology in protein composition, including two major androgen-dependent secretory proteins, referred to as dorsal proteins I and II. Dorsal protein I has been purified and found to be a dimer composed of two identical subunits with sedimentation coefficient 4.6 S, Stokes radius 32 A, and M, = 71,000 (62,000 by sodium dodecyl sulfate (SDS)-polyacrylamide gel electropho- resis). The dimer (6.8 S, 46 A, M, = 150,000) dissociates at high ionic strength and can be partially reconstituted by removal of salt. Dorsal protein I is a basic protein (PI 9) with high lysine content and binds to phospho- cellulose but not to DEAE-Sepharose. Schiff's staining shows that it contains carbohydrate. Quantitative rocket imm~oelectrophoresis using a rabbit antiserum indicates dorsal protein I is produced only in dor- sal prostate and coagulating gland. The protein consti-tutes approximately 26% of total cytosol protein in both organs, yet makes up only 6% of coagulating gland luminal fluid and ejaculated seminal fluid. It was not detected in the rat dorsal prostate tumor (Dunning R3327H). Dorsal protein I1 is a larger protein (Mr = 80,000 by SDS gel electrophoresis) with higher carbo- hydrate content. Under nondenaturing conditions, it has a Stokes radius of >200 A, corresponding to M, of >300,000. Dorsal protein I1 represents a smaller pro- portion of total cytosol protein than does dorsal


Rat Dorsal Prostate and Coagulating Gland Protein
Homology 10947 in 40% methanol and 10% acetic acid overnight and dried under vacuum on low heat.
Sucrose Gradient Centrifugation-Linear 5.4-ml gradients of 5 to 20% (w/v) sucrose in 1 mM EDTA and 50 mM Tris, pH 7.5, contained no KCI, 0.15, or 0.5 M KC1 as indicated. Gradients were centrifuged in a Beckman SW 50.1 rotor at 44,000 rpm for 18 h a t 2°C. Purified dorsal protein samples of 200 pg in 0.2 ml were applied. Sedimentation coefficients were determined by analyzing gradients containing 1 mg of each marker protein: myoglobin (2.0 S), ovalbumin (3.6 S), rat serum albumin (4.4 S), bovine y-globulin (7 S), and catalase (11.3 S). Gradients were fractionated by suction from the bottom of the tube. Fractions containing 12 drops (0.18 ml) were analyzed for protein content by the method of Lowry (1951).
Preparation of Antisera-A White New Zealand adult female rabbit was immunized according to the procedure of Vaitukaitis et al. (1971). Purified dorsal protein (300 pg in 45 pl of H20) was combined with 20 mg of M. butyricum, 2 ml of Freund's complete adjuvant, and 2 ml of sterile saline. The mixture was blended on ice in a Sorvall Omni-Mixer a t increasing speeds for 4 min.
The white colloidal suspension was transferred to a syringe and the contents injected into multiple intradermal sites in the form of rosettes in the shoulder and lower back of the shaved rabbit. B. pertussis antigen (1 ml) was injected separately a t three sites, intradermally. Blood was collected immediately following immunization from an ear vein and used as control serum. A booster injection, administered 12 weeks after the fvst immunization, consisted of 130 pg of dorsal protein I in 2 ml of Freunds complete adjuvant emulsified in the Omni-Mixer. The suspension was administered a t four subcutaneous sites on the back. Two weeks later, the rabbit was bled from the ear with a syringe under suction. The blood clotted at room temperature for 1 to 3 h followed by overnight a t 4°C. Serum was obtained by centrifugation and was stored at -20°C.
Rocket Immunoelectrophoresis-Antiserum was used to quantitate dorsal prostate protein I by rocket immunoelectrophoresis, as previously described (Axelsen et al., 1973). A 1% agarose solution (300 mg in 30 ml) in the electrophoresis buffer, 2 mM diethylbarbituric acid (Na barbital), 0.4 mM Ca lactate, and 73 mM Tris, pH 8.6, is prepared by diluting 5-fold the following stock buffer, 44.& g of diethylbarbital, 88.6 g of Tris, and 1.08 g of Ca lactate in 2 liters, pH 8.6. The agarose is dissolved by heating with stirring and then equilibrated at 50-55°C for 30 min. Antiserum (100 pl) is added and the gel is formed between a coated (with 0.5% agarose in H20) and uncoated plate (11 x 20.5 cm). After cooling (30 min) a t room temperature and removal of the uncoated plate, sample wells are prepared using a 2.5-mm, 5-pl template. Samples are appropriately diluted with electrophoresis buffer and applied with a Hamilton syringe under 50 volts to minimize diffusion. Electrophoresis continues overnight a t 100 volts with cooling. Plates are soaked in saline (0.9% NaCI) for 30 min. The gel is dried and stained for 15 min in 0.3% Coomassie blue in 45% methanol A and 10% acetic acid and destained in 45% methanol and 10% acetic acid. The relationship between rocket length and protein concentration was linear for rockets less than 25 mm in length. Analytical Electrofocusing-Thin layer polyacrylamide gel electrofocusing was carried out on Ampholine (2.4% w/v) polyacrylamide gel plates containing 5% polyacrylamide which have a pH range from 3.5 to 9.5. Samples (1 to 7 mg/ml) were saturated into application filters (approximately 15 pl) and applied to a cooled gel plate. The anode wick contained 1 M HaP04 and the cathode, 1 N NaOH. All samples were electrofocused simultaneously from anode and cathode. The voltage was increased at IO-min intervals from 200 to 950 volts. while the current decreased from 25 to 2 mA over a period of 2'12 h. Hemoglobin (PI = 6.8) and cytochrome c (PI = 10.65) were standards.
Protein concentration-The method of Lowry (1951) was used to estimate protein concentration with bovine serum albumin as standard.

Anatomy of the Rat Prostate
The orientation of rat prostate lobes is illustrated in Fig. 1. Surrounding the neck of the bladder on the urethra are four large ventral lobes (two are visible in Fig. l ) , two smaller lateral and two dorsal (or dorsocaudal) prostate lobes. Removed from the bladder region and lying contiguous on the concave side of the seminal vesicles are the coagulating glands, also known as anterior, cranial, or dorsocranial prostate. Although not illustrated, we have observed that the right coagulating gland often lies outside and free of the connective tissue sheath of the seminal vesicle, while the left coagulating gland is characteristically connected to the seminal vesicle. No structural association could be found between the coagulating gland and dorsal prostate. The lower end of each coagulating gland forms a narrow single duct draining into the urethra. Dorsal and lateral prostates share a common boundary that can be separated along a connective tissue plane of cleavage (Fig. 1B). Unlike other lobes of the prostate, but like the seminal vesicle, each coagulating gland (shown separated from seminal vesicle in Fig. 1B) has a lumen containing fluid of high protein concentration (~3 0 0 mg/ml).

Proteins of Rat Prostate
SDS-polyacrylamide gel electrophoresis of prostate cytosol proteins reveals a strong similarity between rat dorsal prostate

Rat Dorsal Prostate and
Coagulating Gland Protein Homology (Fig. 2, Gel 2) and coagulating gland (Gel 3). In addition, their protein patterns clearly differ from ventral (Gel 9) or lateral (Gel 10) prostate or the Dunning tumor (Gel 5). Two proteins predominate in cytosols from dorsal prostate and coagulating gland, as well as in coagulating gland fluid (Fig. 2, Gels 4 and 7). These bands are also observed in a saline extract of ejaculated rat semen (Fig. 2, Gel 8). Purification and characterization of the M , = 62,000 protein, referred to here as dorsal protein I, are described in this report. Preliminary studies on the larger protein of about M , = 80,000, referred to as dorsal protein 11, suggest that this protein is high in carbohydrate content, as revealed by dark staining with periodic acid-Schiff reagent. Chromatography of coagulating gland fluid on Sepharose 4B in a 1 mM EDTA and 50 mM Tris, pH 7.4, buffer separates the M , = 62,000 dorsal protein I from protein 11.
The high Stokes radius (>200 A) suggests that protein I1 exists under nondenaturing conditions as a large complex in excess of M, = 300,000. This protein has not been further characterized, but is clearly a predominant protein in coagulating gland fluid and seminal fluid (Fig. 2, Gels 7 and 8).
Purification of Dorsal Protein I Dorsal prostate was removed from either Sprague-Dawley or Copenhagen Fischer, Dunning tumor-bearing rats that were either intact or castrated 18 h prior to decapitation. The portion of the urethra adhering to the dorsal prostate was cut away, and the tissue was rinsed in saline, blotted, frozen in liquid N2, and stored at -6OOC. All procedures were carried out at 4°C unless specified otherwise.
Step 1: Preparation of Cytosol-Frozen dorsal prostate (10 g) was crushed under liquid N2 and homogenized in 3 volumes (grams/ml) of 10% glycerol, 1 mM EDTA, and 50 mM Tris, pH 7.5, using an Ultraturrax for approximately 5 min with intermittent cooling. The homogenate was centrifuged for 120 min at 33,000 rpm in a Beckman 35 rotor. The supernatant (cytosol) was removed below the fat layer, aliquots were taken for protein determination, and the sample either used immediately or frozen on dry ice and stored at -60°C.
Step 3: Sephadex G-200-The lyophilized DEAE-Sepharose flow through fractions were resuspended in 1 mM EDTA and 50 m~ Tris, pH 7.5, and further purified by chromatography on Sephadex G-200 (65 x 2.6 cm) equilibrated in the same buffer. A typical protein elution pattern is shown in Fig.  3. Peak fractions with Stokes radii of 46 and 32 8, both contained one predominant band of M , = 62,000 by SDSpolyacrylamide gel electrophoresis.
Step 4: Phosphocellulose Chromatography-The lyophilized fraction of the 32 A geak from Sephadex G-200 chromatography was resuspended in distilled H20, dialyzed for at least 2 h against distilled H20, and applied to a phosphocellulose column (0.9 x 7 cm). Phosphocellulose was previously washed with 1 M KC1, followed by H 2 0 until a chloride electrode indicated no residual C1-. The column was then equilibrated with 50 mM Tris, pH 7.5, and 1 mM EDTA until the pH of the eluant was 7. After application of the sample, a salt gradient of 200  described above was obtained by rocket i~unoelectrophoresis, as summarized in Table I. The apparent low recovery following DEAE-Sepharose may have been due to nonspecific adsorption of protein since little dorsal protein I was detectable in the 1 M KC1 column wash. Sephadex G-200 (Fig. 3) effectively separated what appeared to be monomer from dimer (see below). Chromatography of the monomer on phosphocelluiose separated away the seven remaining contaminating protein bands, as visualized on SDS-polyacryl~de gels.
Approximately 5 mg of purified dorsal protein were recovered from 10 g of dorsal prostate taken from 35 rats.
Although not described here, dorsal protein I can also be readily purified from coagulating gland fluid by chromatography on Sepharose 4B, followed by the phosphocellulose step. A monomer were pooled, dialyzed against dHtO overnight, and lyophilized. The protein (6 mg) was resuspended in 0.3 ml of dHpO, dialyzed for 2 h against distilled H20, and applied to a phosphocellulose column (4 x 0.9 cm) prepared as described (see "Results," purification Step 4). A linear gradient from 0 to 0.3 M KC1 in 1 KIM EDTA and 50 m~ Tris, pH 7.5, was applied while collecting 30-drop (1.1 ml) fractions. C1' ion concentration was estimated with a C1electrode (Orion) as indicated (X).

Molecular Weight and Subunit Structure
Sedimentation coefficients of the 32 and 46 A fractions were 4.6 and 6.8 S, respectively, when determined by linear sucrose gradient centrifugation. As indicated, both fractions yielded identical bands of M, = 62,000 by SDS-polyacrylamide gel electrophoresis. Their h y~o d~a m i c molecular weights were 71,000 and 150,000, as summarized in Table 11. Frictional ratios of 1.16 and 1.3 were estimated for the 32 A, 4.6 S and 46 A, 6.8 S forms, respectively.
Interconversion of the 32 A, 4.6 S and 46 A, 6.8 S Forms Addition of 0.5 M KC1 to the sucrose gradient (Fig. 5A) or Sephadex G-200 (not shown) buffers caused a shift from 6.8 S, 46 A to 4.6 S, 32 A. Subsequent removal of salt by dialysis resulted in partial reco~titution of the 6.8 S form (Fig. 5B). Attempts to c o n f i i the presence of identical subunits by partial NHz-terminal sequence analysis were unsuccessful due to what appeared to be a blocked terminal amino acid.
The ratio of dimer to monomer in cytosol (in 50 mM Tris, pH 7.5, 1 mM EDTA, and 10% glycerol) is usually about 1 to 1 as judged on Sephadex G-200, although Fig. 3 shows significantly more dimer. The pooled dimer peak from Sephadex G-200 could be repeatedly rechromatographed in 50 mM Tris, pH 7.5, and 1 mM EDTA with no apparent dissociation. It is not clear whether both forms exist in uiuo. Our inability to completely reconstitute dimer from monomer may be due to some alteration in the protein during dissociation and reassociation.
Other Properties The purified dorsal protein I stains lightly with periodic acid-Schiff reagent, suggesting that it is a glycoprotein. It was not retained, however, on concanavalin A-Sepharose. Treatment with n e~a m~d a s e had no effect on monomer mobility in SDS-polyacryla~de gels.
The purified dorsal protein is basic in charge since it is not retained by DEAE-Sepharose and adsorbs to phosphocellulose at neutral pH. Analytical electrofocusing in 5% polyacrylamide gels indicates an isoelectric point of approximately pH 9. In agreement with its basic properties, amino acid analysis revealed an unusually high lysine content (Table 111).
The dorsal protein precipitates between 40 and 60% satu-and Coagulating Gland Protein Homology rated (NH*)~SO~ as determined by SDS-polyacryl~de gel electrophoresis or by rocket immunoelectrophoresis.
Antiserum to Dorsal Protein I Antiserum to purXed dorsal protein I formed a precipitin line of cathodal migration by rocket immunoelectrophoresis (Fig. 6). Crossed ~u n o e l e c~o p h o r e s i s of dorsal prostate cytosol further indicated the specificity of the antiserum (Fig.  7). Typical rocket immunoelectrophoresis patterns are shown for an ejaculated seminal fluid-saline extract, coagulating Determined on linear sucrose gradients as described under "Methods," using the standard proteins ovalbumin (3.6 S), bovine yglobulin (7 S), and catalase ( is the elution volume, and V, is the total column volume. A semilogarithmic plot of Ka,1'3 versus Stokes radius was used to estimate the Stokes radii of monomer and dimer. Determined as illustrated in Fig. 2. Estimated as previousiy described (Siege1 and Monty, 1966) using the equation M, = 6 r q NaSl(1-Up), where 7 is viscosity of medium, 0.01 g/s/cm; N is Avogadro's number, 6.02 X lOZ3/g-m0l; a i s Stokes radius, lo-' em; S is Svedberg unit in s; 6 is partial specific volume, 0.743 cm3/g; and p is density, 1.03 g/cm3. The partial specific volume was calculated from the amino acid composition, 5 =

(xc S 7 & / (~s w,)
where U is the partial specific volume of each residue (Cohn and Edsalt, 1941) and w is the weight fraction.
Estimated using the equation f/fo = a/(3~7M,/4?rN)"~ with abbreviations as indicated above. gland fluid, and cytosols from -dorsal prostate, coagulating gland, and ventral prostate (Fig. 8). The dark staining protein of coagulating gland fluid with anodal migration likely consists primarily of the high molecular weight glycoprotein complex (dorsal protein 11); it dispiays no cross-reactivity with the antiserum.

Tissue Distribution
Rocket immunoelectrophoresis and SDS gel electrophoresis were used to determine the d~~b u t i o n of the dorsal protein in various organs of the rat. In cytosols from dorsal prostate and coagulating gland, the concentrations of dorsal protein If a M, = 71,000 is assumed, then the per cent concentration is multiplied by 5.1 to determine moles of amino acid/mol of protein. were similar and ranged from 22 to 30% of total protein. The concentration of dorsal protein in coagulating gland fluid was surprisingly less than in cytosol and similar to that in saline extracts of seminal fluid (Table IV). Fluid could not be collected from dorsal prostate due to the absence of a luminal compartment. Separation of sperm plug from the primary ejaculate revealed negligible dorsal protein I in the plug and high amounts in the fluid portion containing the sperm. The low amount of dorsal protein I detected occasionally in lateral prostate (Table IV) is likely due to incomplete separation from dorsal prostate. The dorsal protein was undetectable in ventral prostate and in the Dunning dorsal prostate tumor (R3327H). It was also not detected in seminal vesicle cytosol or fluid, rat serum, or cytosols of rat testis, epididymis, liver, or kidney. No cross-reaction was noted with human seminal fluid.

Androgen Control
Rocket immunoelectrophoresis was used to quantitate the puritied protein in dorsal prostate and coagulating gland cy-I FIG. 6. Rocket immunoelectrophoresis using antiserum to purified dorsal protein I. Electrophoresis was performed as described under "Methods." The purified 32 A monomer was applied in 5 pI containing (left to right) 1.6,0.8, and 0.4 pg of protein, as measured by the Lowry assay. I 2 3 4 5 6 7 8 9 1 0 FIG. 7. Crossed immunoelectrophoresis of dorsal prostate cytosol. Cytosol was diluted 20-fold with electrophoresis buffer (see "Methods"). Two aliquots of 5 pl each were electrophoresed for 1.3 h a t 200 volts in a 1% agarose slab gel (10 x 10 cm). Bromphenol blue (0.05%) and pyronin Y (0.1%) (Eastman) were placed in separate sample wells containing carrier cytosol protein in order to monitor anodal and cathodal migration. Electrophoresed strips (1.8 x 9 cm) were cut out and placed on another glass plate so that antibodycontaining agarose (100 pl of antisera/30 m l ) could be poured on both sides of the strips. Rocket immunoelectrophoresis was performed as described under "Methods." Shown is the sample electrophoresed toward the anode. The sample electrophoresed toward the cathode was identical except that a single visible rocket was displaced an equal distance in the opposite direction from the starting well.
FIG. 8. Rocket immunoelectrophoresis of prostate cytosols and fluids. Aliquots (5 pl) of each of the following were analyzed by rocket immunoelectrophoresis as described under "Methods." 1, Ejaculated seminal fluid, 1.3 mg/ml (collected after decapitation of rats and homogenized in 10% glycerol, 1 mM EDTA, and 50 mM Tris, pH 7.5, and centrifuged a t 100, OOO X g); 2 to 3, coagulating gland fluid (approximately 300 mg/ml) diluted 1/200 and l/8Q 4 to 6, dorsal cytosol (22.6 mg/ml) diluted 1/50, 1/25, and l / l Q 7 to 9, coagulating gland cytosol (11.8 mg/ml) diluted 1/50, 1/25, and l / l Q and 10, ventral prostate cytosol (12.4 mg/ml) undiluted.  Effects of age and castration on dorsal protein I. Rocket immunoelectrophoresis was used as described under "Methods" to quantitate dorsal protein I in cytosols from rats of differing ages (A) and in adult rats at 2 or 4 weeks after castration ( B ) . Cytosols (100,ooO X g supernatants) were prepared from dorsal prostate (0) or coagulating gland (0). A rocket height of 58 mm corresponded to approximately 1 pg of pure dorsal protein. Total protein was measured by the procedure of Lowry (1951).
tosols from rats of differing ages (Fig. SA) and in adult rats castrated for 2 or 4 weeks (Fig. 9B). The concentration of protein I increased linearly beginning near 20 days of age. This is near the time of onset of puberty when a surge in gonadotropin release stimulates androgen production. Following castration of adult rats, the concentration of dorsal protein I decreased to a low level by 4 weeks. The concentration of dorsal protein I in both glands is thus similarly dependent on androgen stimulation.

DISCUSSION
A major protein of rat dorsal prostate and coagulating gland cytosol, referred to as dorsal protein I, has been purrled, characterized, and a specific antiserum prepared. The protein is a dimer composed of two identical subunits, each with 32 A Stokes radius, 4.6 S sedimentation coefficient, and M, = 71,000 (62,000 by SDS-polyacryl~ide gel electrophoresis). The dimer has a Stokes radius of 46 A, s e d~e n~t i o n coefficient of 6.8 S, and an estimated M, = 150,000. Dorsal protein I appears unique to the dorsal prostate and coagulating gland since it is antigenically undetectable in ventral or lateral prostate, seminal vesicle cytosol or luminal fluid, the Dunning dorsal prostate tumor, or other organs of the rat.
A striking similarity in cytosol proteins of dorsal prostate and coagulating gland suggests that these anatomically distinct organs have similar patterns of gene expression. Two proteins predominate, dorsal protein I, which we have purified, and dorsal protein 11, a larger protein complex with high carbohydrate content. Other prostate lobes display diverse protein profiles. The coagulating gland differs structurally from dorsal prostate in that it is a tubular gland with a luminal c o m p~m e n t containing fluid of high protein content. Dorsal prostate, on the other hand, is a tubuloalveolar gland, like ventral and lateral prostate, draining directly into the urethra via minute ducts. It was, therefore, possible to collect and analyze protein of the secreted fluid from coagulating gland lumen, but not from dorsal prostate. The cytosol-fluid distribution of dorsal proteins I and I1 differs markedly. Dorsal protein I composes approximately 25% of cytosol, yet o d y 5% of fluid total protein, suggesting that it may be compartmentalized within the cell. Dorsal protein 11, on the other hand, represents a smaller proportion of cytosol protein, yet is clearly the predominant protein of coagulating gland fluid.
This relationship is consistent with the observation that coagulating gland secretion is strongly periodic acid-Schiff positive, while the cytoplasm reacts weakly (LeBlond, 1950).
The rat coagulating gland, first described by Walker (1910), is known for its secretion of the heat-labile enzyme, vesiculase (Gotterer and Williams-Ashman, 1957). This enzyme catalyzes coagulation of a protein secreted by the seminal vesicle through a transamidation reaction (Notides and Williams-Ashman, 1967;Williams-Ashman et al., 1972;Beil and Hart, 1973). Both dorsal proteins reported here salt out near 50% saturated (NH4)2S04, the same concentration used to recover vesiculase activity (Gotterer and Williams-Ashman, 1957). Although the enzymatic properties of dorsal proteins I and I1 have not been established, their high concentration in cytosol and fluid argues against an enzymatic function. Other proteins of the coagufating gland are involved in the f o~a t i o n of an intrauterine gel. Joshi et al. ( 1972) observed that, post coitus, the uterine contents of the rat form a viscous gel resulting from the interaction of two major glycoprotein components of coagulating gland fluid. The bulk of the gel consisted of a protein with high carbohydrate content and sedimentation coefficient of 17 S (similar to dorsal protein 11). Gelation of this glycoprotein was dependent on the presence of a 4.7 S protein which contained less carbohydrate (similar to dorsal protein I).
The nearly identical protein profiies observed in dorsal prostate and coagulating gland suggest that these organs may perform the same function in the rat, even though they are anatomically distinct and have a different overall gross structure. In agreement with this idea, earlier reports demonstrated similarities in their epithelial cells histochemically (Korenchevsky and Dennison, 1935), histologically (Price and Williams-Ashman, 1961;Brandes and Groth, 1961), and functionally (Humphrey and Mann, 1949;Mann, 1964). Some differences have been described, however, in the location of nuclei and organization of cisternal spaces (Korenchevsky and Dennison, 1935;Brandes and Groth, 1961). Dorsal prostate and coagulating gland appear to represent organs of unique embryological origin which have differentiated in nearly an identical manner. Each of the different lobes of rat prostate, including coagulating gland, is thought to arise out of the urogenital sinus from a separate epithelial bud (Price, 1963).
These organs are of particular interest because they ais0 relate emb~ologically to parts of the human prostate. Rat dorsal prostate is thought to be homologous to the outer region of human prostate (Price, 1963), the common primary site of prostatic carcinoma, while rat coagulating gland is homologous to the inner region of human prostate, the site of benign prostatic hypertrophy.
Dorsal protein I displays an androgen dependency characteristic of a secretory protein of the prostate or other accessory sex organs. Castration generally affects secretory processes initially, followed by histochemical and weight changes in the organ (Price and Williams-Ashman, 1961). The rather slow decline to near antigenically undetectable levels of dorsal protein I at 4 weeks postcastration may reflect, in part, the presence of stored pools of the dorsal protein within the cell. However, postcastrate regression of the coagulating gland, cytologically noticeable in 10 days, is slower than ventral prostate, observed by 2 days (Price and Williams-Ashman, 1961). All prostatic lobes are androgen dependent, although they atrophy at different rates in response to castration (Price and Williams-Ashman, 1961) or administered antiandrogens (Dahl and Kjaerheim, 1974). All can be subsequently maintained by administration of testosterone.
A goal of this work was to establish a marker that could be used in studies on androgen control of gene expression in the dorsal prostate and Dunning tumor. The tumor (Dunning R3327H) is becoming a promising model for the study of prostate cancer. It is a well differentiated, androgen-dependent adenocarcinoma that is believed to have arisen spontaneously from rat dorsal prostate (Dunning, 1963). The tumor has been propagated by subcutaneous transplan~tion since 1961 at the Papanicolaou Cancer Research Institute in Miami. Like the normal dorsal prostate, it contains both androgen and estrogen receptors ( M a r~a n d et al., 1978;M~~a n d and Lee, 1979;Heston et al., 1979 Wilson andFrench, 1979). That the tumor is truly of dorsal prostate origin has been substantiated in part by an analysis of the activity and distribution of numerous enzymes (Smolev et at., 1977;Muntzing et ai., 1978). The presence of the secretory form of acid phosphatase, an enzyme unique to the prostate, clearly supports at least its prostatic origin (Isaacs et al., 1979). SDS-polyacrylamide gel electrophoresis of tumor cytosol reveals, however, a pattern of proteins markedly different from normal dorsal prostate, coagulating gland, or other prostate lobes. Neither of the bulk proteins, dorsal proteins I or 11, is a major protein in the tumor. Moreover, no antigenic cross-reaction of tumor cytosol or secreted fluid was detectable with antiserum to dorsal protein I by rocket immunoelectrophoresis or radioimmunoassay. Thus, the tumor displays a marked alteration of androgen-controlled gene expression associated with transformed epithelial cells that have remained well differentiated and androgen responsive.