An inverse correlation between expression of a preprocathepsin B-related protein with cysteine-rich sequences and steroid 11beta -hydroxylase in adrenocortical cells.

A cDNA encoding a secretory protein hitherto unknown was cloned from mouse adrenocortical cells by subtractive hybridization between the cells without and with expressing steroid 11beta-hydroxylase (Cyp11b-1), a marker for the functional differentiation of cells in the zonae fasciculata reticularis (zFR). The deduced protein consisting of 466 amino acids contained a secretory signal, epidermal growth factor-like repeats, and a proteolytically inactive cathepsin B-related sequence. The amino acid sequence was 89% identical with that of human tubulointerstitial nephritis antigen-related protein. Among the mouse organs examined, adrenal glands prominently expressed its mRNA. The mRNA and its encoded protein were detected in the outer adrenocortical zones that do not express Cyp11b-1, i.e. the zona glomerulosa and the undifferentiated cell zone, while being undetectable in zFR that express Cyp11b-1. The new protein was designated as adrenocortical zonation factor 1 (AZ-1). Clonal lines with different levels of AZ-1 expression were established from Y-1 adrenocortical cells that originally express Cyp11b-1 but little AZ-1. Analyses of the clonal lines revealed that Cyp11b-1 is detected in the clonal lines maintaining little AZ-1 expression and becomes undetectable in those expressing AZ-1. On the other hand, irrespective of the AZ-1 expression, all clones expressed cholesterol side-chain cleavage enzyme, which occurs throughout the cortical zones. These results demonstrated that adrenocortical cells expressing AZ-1 do not express Cyp11b-1, whereas those with little AZ-1 express this zFR marker in vitro and in vivo, implying a putative role of AZ-1 in determining the zonal differentiation of adrenocortical cells.

The adrenal cortex of mammals consists of three major zones that contain both functionally and morphologically distinct cells; they are the zona glomerulosa (zG), 1 the zona fasciculata (zF), and the zona reticularis (1). The cells in zG, the outermost zone of the cortex, secrete aldosterone, the strongest mineralocorticoid; and those in zF, the middle zone, produce glucocorticoids such as corticosterone in most rodents and cortisol in other mammals including human. Finally, the cells in the zona reticularis, the innermost portion of the cortex, also secrete glucocorticoids in many mammals including rats and mice and produce adrenal androgens in human and some other mammals. In addition to these three, a zone composed of 3-4 layers of small round cells has been recognized between zG and zF in various animals (2)(3)(4)(5)(6)(7)(8)(9). By using rats we showed that the cells of the zone were devoid of the two enzymes (10) determining the distinct steroidogenic functions of zG and zFR, i.e. aldosterone synthase (Cyp11b-2) and steroid 11␤-hydroxylase (Cyp11b-1) responsible for the synthesis of corticosterone/cortisol, respectively (11). The zone thus has been named the undifferentiated cell zone (zU) (12) after the functionally undifferentiated nature of its component cells (13)(14)(15)(16)(17)(18).
Development and maintenance of the adrenocortical zones require many cellular processes including regulation of the steroidogenic gene expression and regulation of cell renewal and arrangements. Steroidogenic factor 1 (19) (SF-1, also referred to as Ad4BP (20)) is a transcription factor essential for embryonic development of steroidogenic organs including adrenal cortex and gonads (21). SF-1 also plays a pivotal role in the earlier steps of the adrenocortical steroidogenesis over the entire adrenal cortex by controlling expression of cholesterol side-chain cleavage enzyme (Cyp11a) and steroid 21-hydroxylase (22). Based on these features, SF-1 is unlikely to be a key regulator for the zonal differentiation of the steroidogenesis. Regarding factors regulating expression of the steroidogenic genes for the last steps of the syntheses, we previously suggested that AP-1 transcription factors were necessary for the spatially restricted expression of Cyp11b-1 in zFR (23,24). Other regulatory factors playing a crucial role in the zonespecific steroidogenesis of zFR have been unknown. Further-more, no regulatory factor for the functional differentiation of the rest of the cortex, zG and zU, has been identified so far. Therefore, molecular mechanisms underlying the zonal differentiation of the adrenocortical steroidogenesis remain to be solved.
The goal of this study was to explore unidentified factors that control the functional differentiation of adrenocortical cells. To this end, we used the mouse adrenocortical cell lines that we established recently (25). They are derived from the adrenal glands of transgenic mice (26,27) carrying a temperaturesensitive large T-antigen gene of simian virus 40, being at different degrees of differentiation from one another. In the present study, a subtractive cDNA cloning employing the adrenocortical cell lines, named AcA101 and AcA201, as well as a conventional cell line Y-1, was carried out. Among the cell lines, AcA101 is the most undifferentiated one which expresses neither Cyp11b-1 nor Cyp11b-2, whereas Y-1 is the most differentiated one that expresses Cyp11b-1 with a responsiveness to ACTH stimuli (28). A cDNA cloned from AcA101 cells by a subtractive hybridization encodes a protein termed adrenocortical zonation factor-1 (AZ-1), the subject of this paper. AZ-1 is a unique secretory protein with a preprocathepsin B-related structure carrying epidermal growth hormone (EGF) motifs. This study demonstrates that expression of AZ-1 in adrenocortical cells is inversely correlated with expression of Cyp11b-1 in vitro and in vivo.
Subtractive cDNA Cloning-Total RNA was extracted from AcA101, AcA201, and Y-1 cells with a modified single-step isolation method employing Trizol reagent (Invitrogen). Poly(A) ؉ RNA was prepared with an oligo(dT)-cellulose column (Amersham Biosciences) and treated with RNase-free DNase (Promega, Madison, WI). Poly(A) ؉ RNA (2 g) from AcA101 cells was converted into cDNA with a NotI-oligo(dT) primer. A cDNA library of AcA101 cells was prepared using the ZipLox system (Invitrogen) according to the instructions from the manufacturer. Subtractive probes were prepared with the chemical crosslinking subtraction method using an RNA subtraction kit (Amersham Biosciences). Single-stranded cDNA (0.4 g) of AcA101 cells was subtracted with poly(A) ϩ RNA (10 g) of AcA201 or Y-1 cells by hybridization and chemical cross-linking reaction. To reduce signals from SV40 T-antigen mRNA, RNA encoding T-antigen was synthesized in vitro by using cloned T-antigen genes (25) and was added to the hybridization reaction. The resulting subtracted cDNA was labeled with [␣-32 P]dCTP (3000 Ci/mmol, Amersham Biosciences) using Sequenase version 2.0 (Amersham Biosciences). The AcA101 cDNA library (2 ϫ 10 5 plaqueforming units) was screened with the subtractive probes under stringent conditions. Positive plaques were isolated through a second screening process using digoxigenin-labeled cDNA (Roche Diagnostics) probes (not subtracted) from AcA101, AcA201, and Y-1 cells. Plasmids carrying a cDNA insert were recovered from phage clones utilizing the cre-lox systems. DNA sequencing was performed by the dideoxy termination method using Thermo Sequenase TM (Amersham Biosciences).
Northern Blotting-Total RNA was prepared from adrenal glands, whole brains, hearts, kidneys, livers, skeletal muscles, spleens, and testes of C57BL/6 mice using Trizol reagent as described above. The RNA preparations (10 g) were subjected to Northern blot analysis as described previously (24). Before transfer to positively charged nylon membranes (Roche Diagnostics), ribosomal RNAs were visualized by ethidium bromide to verify that the amounts of RNA loaded were comparable with each other (Ͻ15%) and that degradation of the RNA preparations was undetectable under our experimental conditions. A cDNA fragment (position 1490 -1810 of Fig. 1A) was labeled with [␣-32 P]dCTP (3000 Ci/mmol, Amersham Biosciences) and High Prime (Roche Diagnostics). Hybridization signals were detected with a Kodak BioMax film with an intensifying screen.
Southern Blotting-Genomic DNA was prepared from the liver of CL57BL/6 mice and was digested with EcoRI, BamHI, and HindIII. The digests (10 g) were subjected to electrophoresis through 0.8% agarose and blotted on a positive charged nylon filter (Roche Diagnostics). A cDNA fragment of AZ-1 (position 1490 -1810) was isolated and labeled with DIG-dUTP (Roche Diagnostics). Hybridization, stringent washing, and detection with color development were carried out according to the manufacturer's instructions (Roche Diagnostics) and as described previously (29,30).
In Vitro Transcription-Translation-To construct a plasmid for expression of recombinant proteins, oligonucleotide primers for PCR were prepared. A forward primer including the putative translational initiation codon was designed to contain an artificial BstXI site at position 48 of the sequence shown in Fig. 1A: 5Ј-CGCCAGTGTGCTGGAG-GCACCATGTGGGGATGT-3Ј. A reverse PCR primer corresponding to the C terminus of the protein was designed to contain the sequence encoding FLAG epitope followed by an XbaI site: 5Ј-GCTCTAGACTA-CTTGTCATCGTCGTCCTTGTAGTCGTGGTGCCCCATGTCCTCC-3Ј. Another reverse PCR primer 5Ј-GCTCTAGATCAGTGGTGCCCCATG-TCCTCC-3Ј lacking the FLAG tag sequence was also designed. cDNA fragments (either the presence or absence of the FLAG sequence) that were amplified by PCR were inserted into pRc/CMV (Invitrogen). To avoid incorrect nucleotides incorporated in PCR, the DraIII fragment (position 69 -1280 of Fig. 1A) carrying the PCR-amplified region was replaced by the authentic fragment prepared from the original cDNA. The resulting pR/C11.13FD1 and pR/C11.13D1 plasmids were used for templates in in vitro transcription-translation reaction of T7 TNT-coupled Reticulocyte Lysate System (Promega) in the presence or absence of [ 35 S]methionine (1000 Ci/mmol, Amersham Biosciences). Reaction mixtures were subjected to SDS-PAGE. Detection was carried out by autoradiography or immunoblotting using anti-FLAG M2 monoclonal antibody (Sigma) as described below.
Antibody Preparation-Peptides P1 (NH 2 -Cys-Ser-Gln-Gly-Arg-Pro-Glu-Gln-Tyr-Arg-Arg-His-Gly-Thr-COOH) and P2 (NH 2 -Cys-Gly-Arg-Val-Gly-Met-Glu-Asp-Met-Gly-His-His-COOH), corresponding to amino acid residues 386 -398 and 456 -466, respectively, were synthesized and each conjugated at the additional Cys residue to keyhole limpet hemocyanin (Pierce) through a thioether bond using maleimideactivated keyhole limpet hemocyanin (Pierce) as described in the instruction manual. The peptide-keyhole limpet hemocyanin conjugates were mixed (1:1 (w/w)) and emulsified with Freund's complete adjuvant and injected into IsaBrown hens 6 times intradermally at 2-week intervals. Blood was collected on the 35th day or thereafter. Antibodies were purified from antisera (10 ml) with a peptide-conjugated agarose gel column. The agarose gel was prepared by coupling a 1:1 (w/w) mixture of the two peptides using a SulfoLink coupling gel (Pierce) according to the instructions. Based on titration by an immunoblot analysis using a recombinant AZ-1 protein, the recovery of the purified antibody after the chromatography was ϳ70%.
In Situ Hybridization-The EcoRI-BamHI fragment (position 1490 -1810 of Fig. 1A) of the isolated cDNA clone was subcloned into pZL1 (Invitrogen). Mouse cDNAs encoding Cyp11b-1 (position 761-950) or Cyp11b-2 (761-953) (31) were obtained by PCR with primer pairs as described previously (25) from the total RNA of Y-1 cells after reverse transcription, and they were cloned into pGEM4-Z (Promega). DIGlabeled antisense and sense RNA probes were synthesized with T7 and SP6 RNA polymerases, respectively, using DIG RNA labeling kit (Roche Diagnostics). Mouse adrenals were excised after transcardial perfusion with phosphate-buffered saline(Ϫ) containing 4% paraformaldehyde and were further fixed with the same fixative overnight at 4°C. The adrenals were embedded with paraffin, and 4-m sections were prepared using 3-aminopropyltriethoxysilane-coated glass slides. After deparaffinization with standard methods, in situ hybridization was carried out as described previously (17). Concentrations of the antisense and sense probes (400 ng/ml) in the hybridization solutions were adjusted based on analysis by the agarose gel electrophoresis of the synthesized RNA.
Immunohistochemistry-Paraffin sections of paraformaldehydefixed mouse adrenal glands were treated with affinity-purified chicken anti-AZ-1 antibody (10 g/ml) or control chicken IgY (10 g/ml, Sigma) overnight at 4°C. The concentrations of the antibodies were adjusted after protein determination with Coomassie Brilliant Blue G-250. The secondary antibody used was rabbit anti-chicken IgY antibody conjugated with horseradish peroxidase (1:300; Promega). The peroxidase activity was visualized with 3,3Ј-diaminobenzidine tetrahydrochloride and hydrogen peroxide as described previously (32).
Transfection-pR/C11.13FD1, which was the same DNA construct as used for in vitro synthesis of FLAG-tagged AZ-1 (AZ-1F), or a control vector pRc/CMV was linearized by digestion with BglII. Y-1 cells (3 ϫ 10 6 ) were transfected with the linearized DNA (10 g) by calcium phosphate precipitation method using the Profection kit (Promega). The cells were selected for resistance to antibiotics G418 (400 g/ml), and colonies were isolated and expanded for characterization as described below.
Immunoprecipitation and Immunoblotting-A mock transfectant (clone 9) and an AZ-1F vector-transfectant (clone 14; see Fig. 7), which showed the highest level of AZ-1F mRNA, were used for detection of the AZ-1 polypeptide. Cell extracts from them (3.6-cm 2 well) were prepared with an SDS sample buffer (120 l) consisting of 62 mM Tris-HCl, pH 6.8, 2% (w/v) SDS, 2% (v/v) 2-mercaptoethanol, and 0.01% bromphenol blue. The extracts were incubated at 100°C for 5 min before electrophoresis. For immunoprecipitation, purified chicken anti-AZ-1 antibody (150 l) was bound to anti-chicken IgY-agarose (150 l of 50% suspension; Promega) by incubation for 1 h at 4°C with gentle shaking. The beads were washed with 1 ml of Tris-buffered saline (TBS) 5 times. The supernatants of the used culture media (1 ml) were prepared by centrifugation to remove cells and cell debris and incubated with 30 l of the gel suspension for 18 h at 4°C with gentle shaking. The agarose gel beads were washed with TBS containing 0.1% Tween 20 (TBST) 4 times followed by a wash with TBS. They were resuspended in an SDS sample buffer and treated at 100°C for 5 min. The immunoprecipitates and the cell extracts were subjected to 10% PAGE in the presence of SDS, and polypeptides were electrophoretically blotted onto Immobilon-P membranes (Millipore, Bedford, MA) according to standard procedures. The membranes were treated with anti-FLAG monoclonal M2 antibody (0.88 g/ml) overnight at 4°C. They were washed with TBST and then incubated with a secondary antibody solution of rabbit anti-mouse IgG conjugated with horseradish peroxidase (1:25,000 dilution with TBST) for 3 h at room temperature. When the affinity-purified chicken anti-AZ-1 antibody (1:200 dilution) was used for immunoblotting, membranes were treated with rabbit anti-chicken IgY conjugated with horseradish peroxidase (1:1000 dilution). Bound secondary antibodies were detected by enhanced chemiluminescence (Amersham Biosciences).
Post-translational Modification-An AZ-1F-vector transfectant (clone 14) was cultured in the presence of 10 g/ml tunicamycin (Sigma) or vehicle (0.2% (v/v) methanol in the culture medium) for 18 h, and cell extracts were prepared as described above. For treatment with peptide N-glycosidase F, cell extracts of the transfectant containing 2% SDS were diluted by 10-fold with 10 mM Tris-HCl, pH 7.5, and incubated with 50 units/ml of N-glycosidase F (Roche Diagnostics) at 37°C for 18 h. After the treatment, the SDS sample buffer was added to the reaction mixtures. These samples were analyzed by immunoblotting as described above.
RT-PCR-RT-PCR analysis was performed by the methods as described in the previous paper (25)  see Fig. 1A) and FL1485r, 5Ј-GTCATCGTCGTCCTTGTAGTCG-3Ј that corresponds to the FLAG peptide encoding sequence); (ii) AZ-1: 420f, 5Ј-GGACAACTGCAATCGATGCACC-3Ј (420 -441); 1003r, 5Ј-GGCT-GTGCATCATACAACGAGG-3Ј (1003-982); this primer pair was used for detection of mRNA of both endogenous Az-1 gene and AZ-1F; (iii) Cyp11b-1 (25); (iv) Cyp11a (25); and (v) glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (25). Amplification conditions for PCR were 45 s at 94°C, 45 s at 56°C, and 2 min at 72°C for appropriate cycle numbers followed by 7 min at 72°C. Cycle numbers were 20 for AZ-1 (420f and 1003r), Cyp11a, and GAPDH and 25 for AZ-1F (64f and FL1485r), and Cyp11b-1. Experiments for comparison of relative amounts of mRNAs among the transfectants were performed within the exponential phase of the amplification reactions to obtain the linear response concerning the initial mRNA amounts. PCR products were analyzed by agarose gel electrophoresis followed by visualization with ethidium bromide. Intensities of the visualized products were determined by densitometric analysis and were normalized with that of GAPDH cDNA.

RESULTS
Cloning and Characterization of AZ-1 cDNA--For the subtractive cDNA cloning to identify genes whose transcripts are expressed in higher levels in undifferentiated adrenocortical cells than in differentiated cells, we prepared two probes as described under "Experimental Procedures"; one was obtained by a subtraction between AcA101 and AcA201 and the other was between AcA101 and Y-1. A cDNA library of AcA101 cells was prepared to represent the whole population of mRNA molecules. It was screened by hybridization of the two subtracted probes onto plaque lifts. We isolated cDNA clones whose hy-bridization signal was detected with both subtractive probes.
Among the cDNA clones isolated, one was composed of 1926 nucleotides except for a poly(A) tail as shown in Fig. 1A. The cDNA attracted our attention because the amino acid sequence of its encoded protein had a unique structure as described below. The protein was termed as AZ-1 since its expression level was inversely related to the degree of the functional differentiation of adrenocortical cells as described below. The cDNA contained an open reading frame encoding a polypeptide chain consisting of 466 amino acids, which was from a translational initiation codon at position 67 to a termination codon at position 1465. Several nucleotides preceding the methionine codon were consistent with the consensus sequence for translational initiation (33). A typical polyadenylation signal was found at position 1907. The N-terminal 17-amino acid sequence was predictable as a signal peptide for the secretory pathway (34). No hydrophobic region large enough to span a membrane was recognized. Possible N-glycosylation sites were present at Asn-77 and Asn-160. These features in the predicted amino acid sequence suggest that AZ-1 is a secretory protein.
When searched in the GenBank TM /EMBL/DDBJ data base, the predicted amino acid sequence of the open reading frame had an identity of 89% with human tubulointerstitial nephritis antigen-related protein (TIN-ag-RP) (35) (Fig. 1B), indicating that the identified cDNA encodes a mouse orthologue of the human protein. As Wex et al. (35) discussed the features of amino acid sequence of TIN-ag-RP, the N-terminal one-third of AZ-1 polypeptide was rich in Cys residues and contained EGFlike repeats. The C-terminal two-thirds of AZ-1 polypeptide contained procathepsin B-related sequence with a Ser residue at position 228, which replaced a conserved Cys residue in the active site of cysteine proteinases (Fig. 1A), suggesting that AZ-1 does not have a proteinase activity. These structural features were also found in tubulointerstitial nephritis antigen (TIN-ag) (36 -38) in mammals and also in non-mammalian homologues including Caenorhabditis elegans F26E4.3 protein (39).
To verify the size of the polypeptide encoded by the isolated cDNA, we carried out in vitro synthesis of AZ-1 using its cDNA as a template for a coupled transcription and translation reaction in the presence of [ 35 S]Met ( Fig. 2A, left). As seen, the reaction with the cDNA insert (lane 2) gave a major product with an electrophoretic mobility of 46 kDa, which matched with the predicted molecular mass of the encoded protein. We also constructed a DNA encoding the AZ-1 polypeptide tagged with a FLAG peptide at the C terminus. The resulting FLAG-tagged polypeptide, AZ-1F, showed a band with a mobility of 47 kDa (lane 3). The increase of 1 kDa in molecular mass was consistent with an addition of the FLAG peptide of 8 amino acid residues to the native polypeptide. Additional bands with higher mobilities might be polypeptides where internal Met residues were used for translational initiation. The reaction with the control vector gave no signal under the experimental conditions (lane 1).
Transcription-translation reaction products in the absence of radioactive amino acids were examined by immunoblotting employing an anti-FLAG antibody ( Fig. 2A, right). Because the control vector gave many bands under the current experimental conditions (lane 4), the experiment with an AZ-1encoding template without FLAG tag resulted in no meaningful signal (lane 5). When the template carrying the FLAGencoding sequence (lane 6) was used, the anti-FLAG antibody recognized a 47-kDa band in addition to the other bands common to lanes 4 and 5. Collectively, the results from in vitro synthesis indicated that the isolated cDNA encoded a 46-kDa polypeptide. To compare the AZ-1 cDNA sequence with mouse genome sequence, Ensembl Genome Browser (www.ensembl.org/) was used. The cDNA sequence was located on chromosome 4 and was encoded in 12 exons that spanned 9175 bases from positions 127,681,610 to 127,672,436 of the chromosome in a reverse direction (Fig. 3, A and B). The genomic sequences were identical to the cDNA sequence. Because the transcription start site was not determined, the 5Ј-end of the cDNA sequence was putatively assigned to the 5Ј-end of exon 1. The sizes of the hybridization signals obtained with the Southern blot analysis (Fig. 2B) agreed well with the restriction sites in the genomic sequence of the data base.
Distribution of AZ-1 mRNA-Expression levels of the AZ-1 mRNA in the cells used for the cloning of the AZ-1 cDNA were examined by Northern blot analysis (Fig. 4A). The greatest intensity of the mRNA signal (ϳ2 kb) was observed in RNA from AcA101 cells (lane 1), which were used as the source of the cDNA library, among the three cell lines. A signal with a lower intensity was detected in RNA from AcA201 cells (lane 2), whereas no signal was obtained in RNA from Y-1 cells (lane 3) under the experimental conditions. Several other cell lines Closed and open squares represent the exons and the introns, respectively. Their sizes (bp) are shown above (exons) and below (introns) the squares. Thick bar marked with P represents the hybridization probe used in the Southern blot analysis of Fig. 2B. Bars marked with E, B, or H represent hybridization fragments predicted by restriction sites for EcoRI, BamHI, and HindIII in the genomic sequence of the data base, being consistent with the results of the Southern analysis (Fig. 2B). B, DNA sequences of exonintron junctions in the Az-1 gene. Upper and lowercase letters represent the exon and the intron sequences, respectively. The encoded amino acid sequence is shown below the nucleotide sequence. The 5Ј-end of the cDNA is putatively assigned to the 5Ј-end of exon 1. were visualized by ethidium bromide to verify that amounts of RNA loaded were comparable with each other and that degradation of the RNA preparations was undetectable under the experimental conditions. An AZ-1 cDNA fragment (position 1490 -1810) was labeled using [␣-32 P]dCTP. Hybridization, stringent washing, and detection by autoradiography were carried out as described under "Experimental Procedures." established simultaneously with AcA101 and AcA201 were also analyzed for AZ-1 mRNA. Levels of the mRNA in such cell lines as AcE60 (25) were consistent with their phenotypes; cell lines displaying functionally undifferentiated phenotypes had higher levels of AZ-1 mRNA expression than those displaying differentiated ones (data not shown).
RNA preparations from various organs of mouse were then analyzed by Northern blot analysis using AZ-1 cDNA as the probe (Fig. 4B) Subsequently, in situ hybridization analysis was performed to locate expression of AZ-1 mRNA in the adrenal glands. As shown in Fig. 5A, hybridization signals of an antisense probe for AZ-1 mRNA were detected in the outer cortical regions composed of zG and zU. The medulla also exhibited weak hybridization signals. Nuclei of the zF cells appeared to give weak nonspecific signals. The faint signals seen in the FR were nonspecific background comparable with those detected with the sense probe (Fig. 5B). To compare the location of AZ-1 mRNA with those of Cyp11b-2 mRNA, in situ hybridization employing an antisense probe for Cyp11b-2 was performed. Cyp11b-2 mRNA was observed sporadically in zG (Fig. 5C). It should thus be noted that AZ-1 mRNA signals were detectable irrespective of the hybridization signals for Cyp11b-2 mRNA in zG. When an antisense probe for Cyp11b-1 was used, hybrid-ization signals were detected in zFR (Fig. 5D). As shown, the hybridization signals for AZ-1 mRNA (Fig. 5A) and those for Cyp11b-1 mRNA (Fig. 5D) did not overlap each other. Therefore, AZ-1 mRNA signals were present in zG and zU but not in zFR, which express Cyp11b-1 mRNA. Fig. 5E illustrates immunohistochemical localization of the AZ-1 protein in mouse adrenal glands by using an anti-AZ-1 antibody raised against peptide antigens (see "Experimental Procedures"). The immunoreactivities were detected in the outer cortical regions corresponding to zG and zU. The weak signals in zFR seemed to be nonspecific because the control IgY gave a comparable level of background (Fig. 5F). The medulla was also immunoreactive. Thus, distribution of the immunoreactivities to anti-AZ-1 antibody and that of the hybridization signals of the AZ-1 antisense probe were indistinguishable from each other. Both AZ-1 mRNA and protein were thus expressed in the outer regions of the adrenal cortex, i.e. the outside of zFR.
Detection of FLAG-tagged AZ-1 Protein with Stably Transfected Y-1 Cells-We then produced stable transfectants of Y-1 cells with an expression vector encoding AZ-1F protein. Among the clones obtained, a clone exhibiting the highest expression level of AZ-1F mRNA was characterized in detail. When the cell extracts prepared from the clone were analyzed by immunoblotting, two bands with electrophoretic mobilities of 51 and 54 kDa were detected by the use of an anti-AZ-1 antibody (Fig. 6A) and an anti-FLAG antibody (Fig. 6B). Cell extracts from a clone obtained with a control vector gave no significant band in either case. We also prepared immunoprecipitates from the cultured medium of the transfected cells by using anti-AZ-1 antibody. As shown in Fig. 6C, immunoblotting of the precipi-FIG. 5. Localization of mouse AZ-1 in the adrenal glands. A, in situ hybridization analysis of mouse adrenal sections using DIG-labeled AZ-1 antisense RNA probe shows dark purple staining in the outer cortical regions composed of zG and zU and the medulla. Note that AZ-1 mRNA signals are detectable in zG irrespective of Cyp11b-2 mRNA expression and in zU (compare with the hybridization signals in C and D). B, in situ hybridization using labeled AZ-1 sense RNA probe gives a weak background signal comparable with that detectable with the antisense probe. C, in situ hybridization using the labeled Cyp11b-2 antisense RNA probe shows hybridization signals sporadically in zG. D, in situ hybridization using the labeled Cyp11b-1 antisense RNA probe shows hybridization signals in zFR. E, immunohistochemistry using the affinity-purified chicken anti-AZ-1 antibody shows reddish brown staining in the outer cortical regions composed of zG and zU and the medulla. Distribution of AZ-1 immunoreactivity is comparable with that of AZ-1 mRNA (compare with the signals in A). F, immunohistochemistry using a control chicken IgY gives a weak background signal comparable with that seen with anti-AZ-1 antibody. Bar, 50 m. tates using anti-FLAG antibody showed 51-and 54-kDa polypeptides. A faint 52-kDa band seen with the medium of the control clone (Fig. 6C, lane 5) was likely to result from a cross-reactivity of the secondary antibody to the heavy chain of bovine or horse IgG in the culture medium. These results were consistent with the presence of the signal peptide sequence at the N terminus of the deduced amino acid sequence of AZ-1.
The difference in the molecular mass of the AZ-1F protein detected with transfected cells (54 and 51 kDa) from that of the in vitro synthesized polypeptide (47 kDa) may be attributable to post-translational modification. To examine N-glycosylation of the polypeptide produced in the transfected cells, the cells were incubated with tunicamycin, and the cell extracts were analyzed by immunoblotting with anti-AZ-1 antibody. As shown in Fig. 6D, a 48-kDa band was detected after the tunicamycin treatment in addition to the 54-and 51-kDa bands, and N-glycanase treatment of the cell extracts from the transfected cells also produced a 48-kDa band (Fig. 6E). These results suggested that AZ-1 protein was post-translationally Nglycosylated in the cells.
Inverse Correlation between AZ-1 and Cyp11b-1 Expression in Clonal Lines Derived from Y-1 Cells-The expression levels of AZ-1 mRNA in the cell lines used for the subtractive cloning were apparently in an inverse correlation to the levels of Cyp11b-1 mRNA (25). The effects of expression of the AZ-1 protein on the phenotypes of Y-1 cells then were examined by analyzing mRNA levels of the steroidogenic genes in the transfectants (Fig. 7). We analyzed 10 clones of mock transfectants obtained with a control DNA and 11 clones obtained with an AZ-1F expression vector. Fig. 7A shows that mRNA encoding AZ-1F was detected in 4 clones (lanes 12, 13, 15, and 18) among the 11 clones isolated after transfection with the AZ-1F expression vector, whereas no AZ-1F mRNA was detectable among the 10 clones transfected with the control vector (lanes 1-10). A PCR primer pair that was able to detect mRNA of both the endogenous Az-1 gene and AZ-1F was used. Unexpectedly, 7 of 10 control clones were found to express significant levels of the endogenous Az-1 gene mRNA (Fig. 7B, left). Among the 11 clones obtained with transfection of the AZ-1F vector, 6 clones (lanes 12, 13, 15, 18, 20, and 21) expressed mRNA of either the endogenous or exogenous Az-1 genes, indicating that the endogenous Az-1 gene was expressed in at least two clones (lanes 20 and 21). The reason for the induction of the endogenous Az-1 gene through the selection procedure for the transfectants was unknown. Fig. 7C shows that Cyp11b-1 mRNA was undetectable only in the clones expressing either AZ-1 from the endogenous gene or AZ-1F. On the other hand, Cyp11a mRNA was detectable in all of the 21 clones with various levels (Fig. 7D). The experiments with a GAPDH primer pair monitored efficiencies of reverse transcription and amplification from the RNA preparations (Fig. 7E). Fig. 7F depicts the inverse correlation of the mRNA levels between AZ-1 and Cyp11b-1 (closed circles); the clones expressing a very low or undetectable level of the endogenous Az-1 gene mRNA expressed Cyp11b-1 mRNA, whereas those expressing higher levels of the AZ-1 mRNA did not express Cyp11b-1 mRNA. The same correlation was observed for the 4 lines expressing AZ-1F (closed triangles). On the other hand, Cyp11a mRNA was expressed irrespective of AZ-1 expression (open circles and open triangles). These results indicated that Cyp11b-1 was expressed in the clonal lines maintaining little AZ-1 expression and became undetectable in those expressing AZ-1. DISCUSSION AZ-1 cloned in the present study turned out to be a putative factor involved in the zonal differentiation of adrenocortical cells, as judged by the solid inverse correlation between expres-sion of AZ-1 and Cyp11b-1 both in vitro and in vivo. Although it has not been established if AZ-1 plays a causal role as a direct signal in repressing the Cyp11b-1 gene, the current results led us to conclude that expression of AZ-1 is coupled with mechanisms determining the functional differentiation of the cortex. As compared with known factors that determine differentiation of adrenocortical cells, AZ-1 showed distinct features. First, AZ-1 is an extracellular protein with unique EGF domains and a catalytically inactive proteinase domain related to cathepsin B. Second, different from SF-1 that is essential for development of the tissue and occurs throughout the adrenocortical zones (19,20), AZ-1 is expressed in a spatially restricted manner and is implicated in the adrenocortical zonation. Third, in contrast to AP-1 which we have previously identified as an activator for the Cyp11b-1 gene expression (23,24), AZ-1 is apparently involved in mechanisms for repressing the gene. Topographic distribution of AP-1 in vivo also agrees well with the function of selectively activating the Cyp11b-1 gene expression (23,24). Collectively with our previous studies, the present results suggest that AP-1 and AZ-1 are expressed in a mutually exclusive manner that topographically corresponds to the actual functions of the cells to execute the zone-specific steroidogenesis.
Recent studies from other investigators have suggested the presence of factors occurring in a zone-specific manner that could simultaneously regulate the zonal differences in the adrenocortical steroidogenesis. Preadipocyte factor 1 (40) (also called zona glomerulosa-specific protein) was reported as such a candidate that is expressed in cells of zG of rats (41). Based on an inhibitory function of preadipocyte factor 1/zona glo- merulosa-specific protein in the differentiation of preadipocytes (40) together with its expression patterns through the development of the adrenal cortex, this factor is presumed to have a role in the functional differentiation of the adrenocortical zones (42). Inner zone antigen is another candidate occurring in zFR in rat adrenal cortex that was identified to be a putative membrane progesterone receptor (43). Because antiinner zone antigen antibody inhibited activities of Cyp11b-1 in addition to steroid 21-hydroxylase in vitro (44), the factor seemed to be implicated in the steroidogenesis of zFR. For identification of these factors, functionally differentiated cells in vivo such as cells of zG and zFR of rats were utilized as the resource materials. Although such methodologies used in these previous studies turned out to be useful in obtaining the specific molecular markers for the adrenocortical zones (41,45), their functions in the zone-specific steroidogenesis and/or in the development of the zones still remain to be elucidated. In order to overcome difficulties to search for factor(s) involved in the functional differentiation, we herein used the recently established adrenocortical cell line AcA101 at an undifferentiated state as the alternative resource material. In this context, the present study employing a different strategy has revealed that the novel factor whose expression was accompanied by the lack of expression of Cyp11b-1 was responsible for the zone-specific steroidogenesis. Thus, regulation of the Az-1 gene is intercon-nected with determining the zonal structures of the adrenal cortex.
At the same time, the current results raise further questions as to the function of AZ-1 in the zonal differentiation of adrenocortical cells. Because expression of the endogenous Az-1 gene was induced through the selection processes for stably transfected cells, direct evidence that cells expressing AZ-1 are devoid of Cyp11b-1 expression was obtained. Although AZ-1 was shown to be secreted in the culture medium by using the transfected cells, it is unknown at present whether the secreted form of AZ-1 could repress expression of the Cyp11b-1 gene by a signaling through the cell surface. Another important issue for the adrenocortical zonation is identification of molecules that determine expression of Cyp11b-2 in cells of zG. Based on the present findings that Cyp11b-2 is expressed in the AZ-1positive cells in vivo and that the Cyp11b-1 is expressed in AZ-1-negative cells in vivo, we speculate that AZ-1 is a prerequisite for cells of zG to express Cyp11b-2 without expression of Cyp11b-1. Identification of such factors occurring in cells of zG that determine Cyp11b-2 expression deserves further studies provided that a combination of AZ-1 with them could actually determine the functional differentiation of cells in zG.
The primary structure of AZ-1 protein suggests that it carries the potential to play a role in formation of the cell arrangements in the adrenal cortex. Among the novel protein family to FIG. 7. Inverse correlation between AZ-1 and Cyp11b-1 expression in clonal lines derived from Y-1 cells. RNAs from 10 stable transfectants obtained with pRc/CMV (control DNA) and 11 obtained with pR/C11.13FD1 (AZ-1F vector DNA) were analyzed by RT-PCR. cDNAs synthesized with oligo(dT) primer using total RNA (1 g) were amplified by PCR using specific primer pairs for FLAG-tagged AZ-1 (AZ-1F) (A), both the endogenous AZ-1 and AZ-1F (B), steroid 11␤-hydroxylase (Cyp11b-1) (C), cholesterol side-chain cleavage enzyme (Cyp11a) (D), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (E) as described under "Experimental Procedures." PCR products (15 l) were analyzed through 2% agarose gels and visualized with ethidium bromide. Numbers with arrowheads indicate sizes of the products. ns, nonspecific band. Amplified conditions were within the exponential phase of the reactions to obtain the linear dose response concerning the initial mRNA amounts. which AZ-1 belongs, functions of TIN-ag have well been characterized (37,46). TIN-ag has been shown to be essential for development of glomeruli in embryonic kidney (37), suggesting that TIN-ag is involved in the epithelial-mesenchymal interactions. Furthermore, despite the replacement of the cysteine residue essential for the proteolytic activity of cathepsin B (47)(48)(49), it was shown that purified TIN-ag binds to laminin and type IV collagen and inhibits polymerization of laminin (46). Because AZ-1 shares essentially the same structural features as TIN-ag, AZ-1 could interact with extracellular matrix proteins. These molecules are generally involved in arrangements of cells in multicellular systems including tumor invasion. It may not be unreasonable to postulate that AZ-1 functions to regulate the arrangement of adrenocortical cells through such cell-to-cell and/or cell-to-matrix interactions.
Previous studies using chimeric (50) and transgenic (51) animals have suggested that adrenocortical cells arranging centripetally are clonal cells; a progenitor cell differentiates into two functionally distinct cell species, cells of zG and those of zFR. Thus, two patterns of the cell arrangement are likely to be present in the adrenal cortex: the centripetal arrangements of clonal cells and the concentric arrangements of functionally differentiated cells. By analogy with the structure of TIN-ag, it can be hypothesized that AZ-1 plays a role in development and maintenance of the cell arrangements. Further characterization of such an effect of AZ-1 on the cell arrangements through the interaction with matrix proteins and its functional link to steroidogenic gene regulation is now underway in this laboratory.