Expression of Recombinant Androgen Receptor in Cultured Mammalian Cells
Valerie E. Quarmby, Jon A. Kemppainen, Madhabananda Sar, Dennis B. Lubahn, Frank S. French, and Elizabeth M. Wilson Laboratories for Reproductive Biology Department of Pediatrics, Cell Biology, and Anatomy (M.S.), Pathology (D.B.L.), and Biochemistry (E.M.W.) University of North Carolina Chapel Hill, North Carolina 27599
Full-length rat and human androgen receptor (AR) cDNA clones were expressed in COS-7 and CV1 monkey kidney cells to analyze the AR protein using immunological and cotransfection techniques. The studies were aided by the development of two rabbit polyclonal antibodies, designated AR32 and AR52, directed against epitopes within the N-terminal region of AR. Each antibody recognizes native AR by sucrose gradient analysis and detects a 114-kilodalton protein in COS cells transfected with human or rat AR cDNA. Covalent binding of the synthetic androgen [3H]methyltrienolone (R1881) to the 114kDa protein was saturable. The endogenous native AR was similarly 114 kDa on immunoblots of a human prostate adenocarcinoma cell line, LNCaP, and rat sex accessory gland extracts. AR was localized in nuclei of transfected COS cells and in LNCaP cells by immunocytochemical staining. Androgen induction of CAT activity was dose dependent in CV1 cells cotransfected with the AR expression vector and a reporter plasmid containing the mouse mammary tumor virus promoter linked to the chloramphenicol acetyltransferase gene. It is concluded that antipeptide antibodies are useful reagents in characterizing both native and denatured forms of the AR protein. The 114-kDa protein expressed transiently in cultured cells represents the full-length AR protein, has a molecular size equivalent to that of endogenous AR, and mediates androgen-dependent transcriptional activation in CV1 cells. (Molecular Endocrinology 4: 1399-1407, 1990)
molecular techniques. Studies with DNA probes revealed a 10-kilobase (kb) AR mRNA in reproductive organs of the rat (3, 7) and approximately 10- and 7-kb mRNAs in human prostate (4, 5). AR mRNA is acutely regulated by androgen, i.e. androgen causes a decrease in the amount of AR mRNA both in the LNCaP human prostate cancer cell line and in vivo in several tissues of the rat known to express AR (7). The level of AR mRNA varies among tissues and generally reflects the degree of androgen responsiveness of a tissue. This applies as well to sublines of the Dunning prostate tumor, where the androgen-dependent tumors have greater amounts of AR mRNA (8). Sequences of 15-20 amino acids within the AR Nterminal domain were synthesized for use as peptide immunogens in raising anti-AR antibodies. One of these antibodies, AR52, was shown previously to be effective in immunocytochemical analysis of AR and revealed AR nuclear localization in epithelial cells of rat (3) and human (4) prostate and in a variety of other tissues, including brain and pituitary (9). In the present report the full-length AR cDNA was expressed in cultured mammalian cells for analysis of the AR protein by immunocytochemistry, ligand binding, and immunoblotting. The studies indicate that recombinant AR appears identical to native AR, and transient cotransfection is well suited for structure/function analyses.
RESULTS Antipeptide Antibodies that React with AR An antibody that reacts selectively with the native AR protein was previously described by this laboratory (3) and is referred to as AR52. The epitope for AR52 lies on the N-terminal side of the DNA-binding domain, as shown in Fig. 1. A second antipeptide antibody was raised in a rabbit to a region beginning nine amino acids from the N-terminus. This AR32 antibody was directed against a 21-amino acid sequence in common with the human and rat receptor, shown schematically in Fig. 1.
INTRODUCTION
Knowledge of the full-length cDNA sequence of the androgen receptor (AR) (1-6) offers the opportunity to analyze this low abundance protein using a variety of 0888-8809/90/1399-1407$02.00/0 Molecular Endocrinology Copyright © 1990 by The Endocrine Society
1399
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 05:30 For personal use only. No other uses without permission. . All rights reserved.
Vol 4 No. 9
MOL ENDO-1990 1400
N-terminal
DNA
Hormone
919
AspHisValLeuProlleAspTyrTyrPheProProGlnLysThrGlyGlyCys
ArgValTyrProArgProProSerLysThrTyrArgGlyAlaPheGlnAsnLeuPheGlnCys Fig. 1. Synthetic Peptide Immunogens for Antibodies AR32 and AR52 Human AR is 919 amino acids in length, and its functional domains include the N-terminal-, DNA-, and hormone-binding domains as shown. The peptide for AR32 begins at amino acid 9 from the N-terminus. It contains 20 amino acids derived from the human and rat AR sequences and a carboxyl-terminal cysteine for use in coupling to keyhole limpet hemocyanin. The peptide for AR52 consists of amino acids 544-558 of human AR and 527-541 for rat AR plus a Gly-Gly-Cys linker. It occurs on the N-terminal side of the DNA-binding domain.
I400
AR32 recognizes the free peptide by enzyme-linked
immunosorbant assay at a serum dilution greater than 1:10,000 (data not shown) and reacts with native rat AR by increasing receptor sedimentation on a sucrose gradient. As shown in Fig. 2, approximately 40% of the [3H]R1881-labeled AR in a rat epididymal cytosol fraction shifted to 9S on a sucrose gradient after incubation with AR32 immunoglobulin G (IgG). The remaining portion of AR not recognized by the antibody may represent proteolytically cleaved fragments that have severed the high affinity androgen binding domain from the N-terminus containing the antibody epitope. Thus, two antipeptide antibodies were found to react with native endogenous AR protein.
1200
I000
§ 800 o ro o. 600 o
Immunoblot Analysis of AR Transiently Expressed in COS Cells
400
200
10 20 Fraction Number
30
Fig. 2. Sucrose Gradient Analysis of [3H]R1881-Labeled Rat Epididymal AR Analyzed in the Presence and Absence of AR32 IgG A 45% ammonium sulfate fraction of epididymal cytosol prepared from a 24-h castrate rat was obtained as previously described (12) and labeled for 1 h at 4 C with 20 nM [3H] R1881. Samples were incubated in the presence (O) or absence (•) of 270 Mg AR32 IgG for 2 h at 4 C. A control sample was incubated in the presence of excess unlabeled R1881 (A). A 1000-fold molar excess of unlabeled R1881 was added to all samples 30 min before charcoal adsorption to promote dissociation of labeled hormone from any androgen-binding protein present in the sample. Samples were analyzed on 2 20% sucrose gradients containing 0.15 M KCI, 10% glycerol, 5 mM mercaptoethanol, and 50 IDM Tris, pH 7.2. Sedimentation markers were analyzed on a parallel gradient and included ovalbumin (3.6S) and 7-globulin (7S).
Mammalian expression vectors, pCMVrAR and pCMVhAR (p5HBhAR-A), containing full-length coding sequences of rat and human AR cDNAs, respectively, were used in transient transfection of COS-7 cells and were predicted to express proteins of approximately 100 kilodaltons (kDa). Immunoblot analysis with AR52 of cell lysates containing human AR showed three predominant bands at 114, 88, and 71 kDa and a less prominent band at 48 kDa (Fig. 3A, lane 2). These bands were absent from mock-transfected cells (Fig. 3A, lane 1) and from cells transfected with the parent pCMV plasmid lacking the AR sequence (data not shown), suggesting that they represent AR protein. There were no detectable bands when AR52 immune IgG was preadsorbed with the peptide antigen before immunoblot analysis (Fig. 3A, lanes 3 and 4). A comparison of human and rat AR revealed a common major 114-kDa immunoreactive band (Fig. 3B). The smaller bands in Fig. 3, A and B, may result from alternative translational initiation within the AR sequence or from proteolytic degradation. The N-terminal domain of human and rat AR contains sequences that could serve as alternative initiation signals for truncated receptors (4). Two of these might initiate AR peptides of 88 and 48 kDa if electrophoretic mobilities correspond to true mol wt. On the other hand, the varied intensity of the
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 05:30 For personal use only. No other uses without permission. . All rights reserved.
1401
AR Expression in Mammalian Cells
lower mol wt bands suggests the possibility of proteolytic degradation. Further evidence that the 114-kDa immunoreactive band represents AR was obtained by photoaffinity labeling. COS cells transiently expressing human AR were incubated for 2 h with a synthetic androgen, [3H] methyltrienolone (R1881), in the presence or absence of excess unlabeled R1881 to assess nonspecific binding. After brief exposure to UV light at 4 C, cells were lysed and analyzed on sodium dodecyl sulfate (SDS)polyacrylamide gels by fluorography and immunoblotting. Fluorography revealed a predominant radioactive band of 114 kDa which was absent when cells were labeled in the presence of excess unlabeled R1881, suggesting that binding of this ligand was saturable (Fig. 4, lanes 4 and 5). The [3H]R1881-labeled band corresponds in size to the 114-kDa protein observed in a parallel track immunoblotted with the antipeptide antibody AR52 (Fig 4, lane 2). Pretreatment of AR52 with the peptide antigen resulted in loss of the immunoreactive 114-kDa band (Fig. 4, lane 3). An additional protein species at 245 kDa was saturably labeled with [3H]R1881 and reacted with the antipeptide antibody. This 245-kDa band may represent receptor dimerization with itself or association with another protein, possibly resulting from its overexpression. Although AR appears as intense bands both immunologically and by photoaffinity labeling in COS cells, no predominant protein was detected when a parallel sample was stained with Coomassie blue (Fig. 4, lane 6), suggesting that even in transfected COS cells, AR represents a relatively small proportion of total cellular protein. The binding affinity of AR expressed in COS cells is similar to that of native AR (10). Immunoblot Analysis of Endogenous Native AR The metastatic human prostate carcinoma cell line LNCaP and several androgen-responsive reproductive tissues of the rat were analyzed using the immunoblot technique. In the LNCaP human prostate tumor cell line, two immunoreactive bands were detected with AR52. One corresponded in size to the 114-kDa band observed in transfected COS cells (Fig. 5, lane 3), and a second was of similar intensity, migrating at 167 kDa. Like the 114-kDa band, the 167-kDa band was not observed after preadsorption of antibody with the peptide (Fig. 5, lane 4). Exposure time for the autoradiographs from LNCaP cells was approximately 6 times longer than that for COS cells transfected with AR cDNA, indicating that expression of endogenous AR in LNCaP cells is significantly lower than that in the transient transfection system. Since the same two bands of 167 and 114 kDa were observed with endogenous AR in rat epididymis, ventral prostate, and the R3327G Dunning tumor (data not shown), the second antibody, AR32, was tested in immunoblots of endogenous AR. Extracts were analyzed from rat ventral prostate, epididymis, and spleen, a tissue shown to lack androgen-binding activity (11)
peptide competed A
-AR+AR
-AR+AR
(T
I
MW
MW
kDa -200
kDa
if
II4-I
-114 _92
7I-I
48-1
-46
.-30
I 2
3
Fig. 3. Immunoblot Analysis Using AR52 IgG against Proteins Solubilized from COS-7 Cells Transiently Transfected with the Human and Rat AR Expression Vectors A, Cell lysates of mock-transfected cells to which no plasmid DNA was added (-AR) and cells transfected with pCMVhAR (+AR) were separated on an 8% SDS-polyacrylamide gel and transferred to an Immobilon membrane as described in Materials and Methods. The membrane was reacted with AR52 IgG, either untreated or pretreated with the free peptide immunogen (peptide competed). Cells were harvested, washed in PBS, and resuspended by sonication in 25 n\ SDS-sample buffer. Aliquots of 10 n\ containing approximately 45 M9 total protein of cell lysates were applied to each gel lane. Lane 1, Mock-transfected cell extract; lane 2, COS cells transfected with pCMVhAR. Lanes 1 and 2 were reacted with AR52 IgG. Lanes 3 and 4, same as lanes 1 and 2, except exposed to AR52 IgG, which was preadsorbed with peptide. Lane 5, 14 CRadiolabeled mol wt (MW) markers, with sizes shown on the right. Calculated mol wt of immunoreactive bands are shown on the left. B, Immunoblot comparison of rat and human AR expressed in COS cells. COS cells transfected with full-length rat pCMVrAR (lane 1) and human pCMVhAR (lane 2) were lysed as described above, and equivalent aliquots containing approximately 15 M9 total cell protein were analyzed on immunoblots, as described in Materials and Methods. 14C-Labeled mol wt markers are shown in lane 3. The calculated mol wt of immunoreactive bands are shown on the left.
and AR mRNA by Northern blot analysis (3). The 114kDa protein species recognized by AR52 was observed with AR32 in ventral prostate and epididymis (Fig. 6, lanes 1 and 2) and in transfected COS cells (Fig. 6, lane 4). However, the 167-kDa protein species was not observed. Detection of the 114-kDa AR band was eliminated by preadsorption of antibody with its peptide antigen (Fig. 6, lanes 5-8) and was absent in spleen (Fig. 6, lane 3). Two other major reactive proteins of approximately 120 and 74 kDa were recognized by AR32 in tissue extracts and were partially competed by preadsorption of the IgG fraction with peptide. These additional protein species are not AR because they were absent in cells transfected with the AR expression vector and were present in spleen, a tissue that lacks measurable amounts of the AR gene product. An additional AR fragment with a mol wt of about 66 kDa was detected
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 05:30 For personal use only. No other uses without permission. . All rights reserved.
Vol 4 No. 9
MOL ENDO-1990 1402
MW
IMMUNOBLOT
FLUORO6RAPH
peptide compete
MW COS
STAIN
LNCaP
RI88I compete
peptide compete
245200-
kDa
Mf: kDa
200-
II4-
-I67 -II4
64-
9346-
69-
32-
Valerie E. Quarmby, Jon A. Kemppainen, Madhabananda Sar, Dennis B. Lubahn, Frank S. French, and Elizabeth M. Wilson Laboratories for Reproductive Biology Department of Pediatrics, Cell Biology, and Anatomy (M.S.), Pathology (D.B.L.), and Biochemistry (E.M.W.) University of North Carolina Chapel Hill, North Carolina 27599
Full-length rat and human androgen receptor (AR) cDNA clones were expressed in COS-7 and CV1 monkey kidney cells to analyze the AR protein using immunological and cotransfection techniques. The studies were aided by the development of two rabbit polyclonal antibodies, designated AR32 and AR52, directed against epitopes within the N-terminal region of AR. Each antibody recognizes native AR by sucrose gradient analysis and detects a 114-kilodalton protein in COS cells transfected with human or rat AR cDNA. Covalent binding of the synthetic androgen [3H]methyltrienolone (R1881) to the 114kDa protein was saturable. The endogenous native AR was similarly 114 kDa on immunoblots of a human prostate adenocarcinoma cell line, LNCaP, and rat sex accessory gland extracts. AR was localized in nuclei of transfected COS cells and in LNCaP cells by immunocytochemical staining. Androgen induction of CAT activity was dose dependent in CV1 cells cotransfected with the AR expression vector and a reporter plasmid containing the mouse mammary tumor virus promoter linked to the chloramphenicol acetyltransferase gene. It is concluded that antipeptide antibodies are useful reagents in characterizing both native and denatured forms of the AR protein. The 114-kDa protein expressed transiently in cultured cells represents the full-length AR protein, has a molecular size equivalent to that of endogenous AR, and mediates androgen-dependent transcriptional activation in CV1 cells. (Molecular Endocrinology 4: 1399-1407, 1990)
molecular techniques. Studies with DNA probes revealed a 10-kilobase (kb) AR mRNA in reproductive organs of the rat (3, 7) and approximately 10- and 7-kb mRNAs in human prostate (4, 5). AR mRNA is acutely regulated by androgen, i.e. androgen causes a decrease in the amount of AR mRNA both in the LNCaP human prostate cancer cell line and in vivo in several tissues of the rat known to express AR (7). The level of AR mRNA varies among tissues and generally reflects the degree of androgen responsiveness of a tissue. This applies as well to sublines of the Dunning prostate tumor, where the androgen-dependent tumors have greater amounts of AR mRNA (8). Sequences of 15-20 amino acids within the AR Nterminal domain were synthesized for use as peptide immunogens in raising anti-AR antibodies. One of these antibodies, AR52, was shown previously to be effective in immunocytochemical analysis of AR and revealed AR nuclear localization in epithelial cells of rat (3) and human (4) prostate and in a variety of other tissues, including brain and pituitary (9). In the present report the full-length AR cDNA was expressed in cultured mammalian cells for analysis of the AR protein by immunocytochemistry, ligand binding, and immunoblotting. The studies indicate that recombinant AR appears identical to native AR, and transient cotransfection is well suited for structure/function analyses.
RESULTS Antipeptide Antibodies that React with AR An antibody that reacts selectively with the native AR protein was previously described by this laboratory (3) and is referred to as AR52. The epitope for AR52 lies on the N-terminal side of the DNA-binding domain, as shown in Fig. 1. A second antipeptide antibody was raised in a rabbit to a region beginning nine amino acids from the N-terminus. This AR32 antibody was directed against a 21-amino acid sequence in common with the human and rat receptor, shown schematically in Fig. 1.
INTRODUCTION
Knowledge of the full-length cDNA sequence of the androgen receptor (AR) (1-6) offers the opportunity to analyze this low abundance protein using a variety of 0888-8809/90/1399-1407$02.00/0 Molecular Endocrinology Copyright © 1990 by The Endocrine Society
1399
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 05:30 For personal use only. No other uses without permission. . All rights reserved.
Vol 4 No. 9
MOL ENDO-1990 1400
N-terminal
DNA
Hormone
919
AspHisValLeuProlleAspTyrTyrPheProProGlnLysThrGlyGlyCys
ArgValTyrProArgProProSerLysThrTyrArgGlyAlaPheGlnAsnLeuPheGlnCys Fig. 1. Synthetic Peptide Immunogens for Antibodies AR32 and AR52 Human AR is 919 amino acids in length, and its functional domains include the N-terminal-, DNA-, and hormone-binding domains as shown. The peptide for AR32 begins at amino acid 9 from the N-terminus. It contains 20 amino acids derived from the human and rat AR sequences and a carboxyl-terminal cysteine for use in coupling to keyhole limpet hemocyanin. The peptide for AR52 consists of amino acids 544-558 of human AR and 527-541 for rat AR plus a Gly-Gly-Cys linker. It occurs on the N-terminal side of the DNA-binding domain.
I400
AR32 recognizes the free peptide by enzyme-linked
immunosorbant assay at a serum dilution greater than 1:10,000 (data not shown) and reacts with native rat AR by increasing receptor sedimentation on a sucrose gradient. As shown in Fig. 2, approximately 40% of the [3H]R1881-labeled AR in a rat epididymal cytosol fraction shifted to 9S on a sucrose gradient after incubation with AR32 immunoglobulin G (IgG). The remaining portion of AR not recognized by the antibody may represent proteolytically cleaved fragments that have severed the high affinity androgen binding domain from the N-terminus containing the antibody epitope. Thus, two antipeptide antibodies were found to react with native endogenous AR protein.
1200
I000
§ 800 o ro o. 600 o
Immunoblot Analysis of AR Transiently Expressed in COS Cells
400
200
10 20 Fraction Number
30
Fig. 2. Sucrose Gradient Analysis of [3H]R1881-Labeled Rat Epididymal AR Analyzed in the Presence and Absence of AR32 IgG A 45% ammonium sulfate fraction of epididymal cytosol prepared from a 24-h castrate rat was obtained as previously described (12) and labeled for 1 h at 4 C with 20 nM [3H] R1881. Samples were incubated in the presence (O) or absence (•) of 270 Mg AR32 IgG for 2 h at 4 C. A control sample was incubated in the presence of excess unlabeled R1881 (A). A 1000-fold molar excess of unlabeled R1881 was added to all samples 30 min before charcoal adsorption to promote dissociation of labeled hormone from any androgen-binding protein present in the sample. Samples were analyzed on 2 20% sucrose gradients containing 0.15 M KCI, 10% glycerol, 5 mM mercaptoethanol, and 50 IDM Tris, pH 7.2. Sedimentation markers were analyzed on a parallel gradient and included ovalbumin (3.6S) and 7-globulin (7S).
Mammalian expression vectors, pCMVrAR and pCMVhAR (p5HBhAR-A), containing full-length coding sequences of rat and human AR cDNAs, respectively, were used in transient transfection of COS-7 cells and were predicted to express proteins of approximately 100 kilodaltons (kDa). Immunoblot analysis with AR52 of cell lysates containing human AR showed three predominant bands at 114, 88, and 71 kDa and a less prominent band at 48 kDa (Fig. 3A, lane 2). These bands were absent from mock-transfected cells (Fig. 3A, lane 1) and from cells transfected with the parent pCMV plasmid lacking the AR sequence (data not shown), suggesting that they represent AR protein. There were no detectable bands when AR52 immune IgG was preadsorbed with the peptide antigen before immunoblot analysis (Fig. 3A, lanes 3 and 4). A comparison of human and rat AR revealed a common major 114-kDa immunoreactive band (Fig. 3B). The smaller bands in Fig. 3, A and B, may result from alternative translational initiation within the AR sequence or from proteolytic degradation. The N-terminal domain of human and rat AR contains sequences that could serve as alternative initiation signals for truncated receptors (4). Two of these might initiate AR peptides of 88 and 48 kDa if electrophoretic mobilities correspond to true mol wt. On the other hand, the varied intensity of the
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 05:30 For personal use only. No other uses without permission. . All rights reserved.
1401
AR Expression in Mammalian Cells
lower mol wt bands suggests the possibility of proteolytic degradation. Further evidence that the 114-kDa immunoreactive band represents AR was obtained by photoaffinity labeling. COS cells transiently expressing human AR were incubated for 2 h with a synthetic androgen, [3H] methyltrienolone (R1881), in the presence or absence of excess unlabeled R1881 to assess nonspecific binding. After brief exposure to UV light at 4 C, cells were lysed and analyzed on sodium dodecyl sulfate (SDS)polyacrylamide gels by fluorography and immunoblotting. Fluorography revealed a predominant radioactive band of 114 kDa which was absent when cells were labeled in the presence of excess unlabeled R1881, suggesting that binding of this ligand was saturable (Fig. 4, lanes 4 and 5). The [3H]R1881-labeled band corresponds in size to the 114-kDa protein observed in a parallel track immunoblotted with the antipeptide antibody AR52 (Fig 4, lane 2). Pretreatment of AR52 with the peptide antigen resulted in loss of the immunoreactive 114-kDa band (Fig. 4, lane 3). An additional protein species at 245 kDa was saturably labeled with [3H]R1881 and reacted with the antipeptide antibody. This 245-kDa band may represent receptor dimerization with itself or association with another protein, possibly resulting from its overexpression. Although AR appears as intense bands both immunologically and by photoaffinity labeling in COS cells, no predominant protein was detected when a parallel sample was stained with Coomassie blue (Fig. 4, lane 6), suggesting that even in transfected COS cells, AR represents a relatively small proportion of total cellular protein. The binding affinity of AR expressed in COS cells is similar to that of native AR (10). Immunoblot Analysis of Endogenous Native AR The metastatic human prostate carcinoma cell line LNCaP and several androgen-responsive reproductive tissues of the rat were analyzed using the immunoblot technique. In the LNCaP human prostate tumor cell line, two immunoreactive bands were detected with AR52. One corresponded in size to the 114-kDa band observed in transfected COS cells (Fig. 5, lane 3), and a second was of similar intensity, migrating at 167 kDa. Like the 114-kDa band, the 167-kDa band was not observed after preadsorption of antibody with the peptide (Fig. 5, lane 4). Exposure time for the autoradiographs from LNCaP cells was approximately 6 times longer than that for COS cells transfected with AR cDNA, indicating that expression of endogenous AR in LNCaP cells is significantly lower than that in the transient transfection system. Since the same two bands of 167 and 114 kDa were observed with endogenous AR in rat epididymis, ventral prostate, and the R3327G Dunning tumor (data not shown), the second antibody, AR32, was tested in immunoblots of endogenous AR. Extracts were analyzed from rat ventral prostate, epididymis, and spleen, a tissue shown to lack androgen-binding activity (11)
peptide competed A
-AR+AR
-AR+AR
(T
I
MW
MW
kDa -200
kDa
if
II4-I
-114 _92
7I-I
48-1
-46
.-30
I 2
3
Fig. 3. Immunoblot Analysis Using AR52 IgG against Proteins Solubilized from COS-7 Cells Transiently Transfected with the Human and Rat AR Expression Vectors A, Cell lysates of mock-transfected cells to which no plasmid DNA was added (-AR) and cells transfected with pCMVhAR (+AR) were separated on an 8% SDS-polyacrylamide gel and transferred to an Immobilon membrane as described in Materials and Methods. The membrane was reacted with AR52 IgG, either untreated or pretreated with the free peptide immunogen (peptide competed). Cells were harvested, washed in PBS, and resuspended by sonication in 25 n\ SDS-sample buffer. Aliquots of 10 n\ containing approximately 45 M9 total protein of cell lysates were applied to each gel lane. Lane 1, Mock-transfected cell extract; lane 2, COS cells transfected with pCMVhAR. Lanes 1 and 2 were reacted with AR52 IgG. Lanes 3 and 4, same as lanes 1 and 2, except exposed to AR52 IgG, which was preadsorbed with peptide. Lane 5, 14 CRadiolabeled mol wt (MW) markers, with sizes shown on the right. Calculated mol wt of immunoreactive bands are shown on the left. B, Immunoblot comparison of rat and human AR expressed in COS cells. COS cells transfected with full-length rat pCMVrAR (lane 1) and human pCMVhAR (lane 2) were lysed as described above, and equivalent aliquots containing approximately 15 M9 total cell protein were analyzed on immunoblots, as described in Materials and Methods. 14C-Labeled mol wt markers are shown in lane 3. The calculated mol wt of immunoreactive bands are shown on the left.
and AR mRNA by Northern blot analysis (3). The 114kDa protein species recognized by AR52 was observed with AR32 in ventral prostate and epididymis (Fig. 6, lanes 1 and 2) and in transfected COS cells (Fig. 6, lane 4). However, the 167-kDa protein species was not observed. Detection of the 114-kDa AR band was eliminated by preadsorption of antibody with its peptide antigen (Fig. 6, lanes 5-8) and was absent in spleen (Fig. 6, lane 3). Two other major reactive proteins of approximately 120 and 74 kDa were recognized by AR32 in tissue extracts and were partially competed by preadsorption of the IgG fraction with peptide. These additional protein species are not AR because they were absent in cells transfected with the AR expression vector and were present in spleen, a tissue that lacks measurable amounts of the AR gene product. An additional AR fragment with a mol wt of about 66 kDa was detected
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 05:30 For personal use only. No other uses without permission. . All rights reserved.
Vol 4 No. 9
MOL ENDO-1990 1402
MW
IMMUNOBLOT
FLUORO6RAPH
peptide compete
MW COS
STAIN
LNCaP
RI88I compete
peptide compete
245200-
kDa
Mf: kDa
200-
II4-
-I67 -II4
64-
9346-
69-
32-
