Expression of the c-Harvey ras Oncogene Alters Peptide Synthesis in the ~eurosecretory Cell Line AtTZO*

Ras proteins are enriched in neurosecretory cells suggesting that ras may play an important role in the of proopiomelanocortin in AtT20 cells following ras oncogene

The protooncogene ras encodes a 21-kDa protein (p21"") that is highly enriched in terminally differentiated neurons, neurosecretory cells, and muscle (Swanson et al., 1986;Furth et al., 1987;Mizoguchi et al., 1989;Chera et al., 1987). ras has been implicated as a component of regulatory pathways controlling cell division and differentiation, including differentiation to a neuronal phenotype (Bar-Sagi and Feramisco, 1985;Noda et al., 1985;Hagag et al., 1986;Barbacid, 1987;Hanley and Jackson, 1987). Its high expression in mature neurons suggests additional regulatory roles in processes characteristic of these cells, such as electrical excitability and secretion. To address whether ~21"" is capable of regulating transmitter release from neurosecretory cells, we have transfected the human H-ras oncogene, EJ-ras (Shih and Weinberg, 1982), into a model neuropeptide synthesizing and secreting cell line, AtTXO. These cells are an excitable mouse anterior pituitary cell line that generate long duration, spontaneous calciumdependent action potentials (Suprenant, 1982;Adler et al., 1983). AtT20 cells have been used as a model system for the study of a number of processes underlying neuropeptide physiology. These include hormonal regulation of ACTH' and 8endorphin biosynthesis and release (Herbert et al., 1978), * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "~v e~~e m e n t~' in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

EXPERIMENTAL PROCEDURES
Culture Conditions and Transfections-All cells were maintained in Dulbecco's modified Eagle's medium with 10% fetal bovine serum at 37 "C in 8% COz and 92% air, except where noted. Transfections were performed by the method of CapoI coprecipitation as described (Gorman et aL, 1983). Plasmid EJ6.6 was a gift from Ciaho Shih (University of Pennsylvania). Plasmid pRSVNeo was a gift from Bruce Howard (National Cancer Institute).
Preparation of RNA and Northern Blots-RNA was prepared by acid guanidinium thiocyanate-phenol-chloroform extraction, essentially as described (Chomczynski and Sacchi, 1987). Five pg of total RNA was run on 0.66 M formaldehyde, 1% agaroae (wfv) gels in 1 X MOPS buffer (Davis et nl., 1986) with recirculation. Gels were transferred onto Genescreen Plus membranes (Du Pont-New England Nuclear) in 10 X SSC (1.5 M NaCI, 0.15 M sodium citrate). Membranes were then baked for 2 h at 80 "C and prehybridized in Hyb-N (Davis et al., 1986) containing 1% SDS and 200 pg/ml denatured salmon sperm DNA at 44 "C for >1 h. Hybridization was carried out at 44 "C for >16 h in Hyb-N, 1% SDS, 100 pg/ml denatured salmon sperm DNA, and 0.5 to 2 X lo6 cpm/ml of labeled probe. The highest stringency washes were at 0.1 X SSC, 1% SDS at 44 "C. Autoradiography was performed using Kodak X-Omat AR x-ray film to visualize the hybridized probe. The human y-actin cDNA probe was a gift from Y.-C. Cheng (Yale University). The rat cyclophilin cDNA was a gift from R. Handschumacher. The mouse proopiomelanocortin cDNA clone pMKSU16 (Uhler and ) was a gift from M. Uhler. Each cDNA fragment was gel-purified and labeled by random primer extension with DNA polymerase I large fragment and Primer Extension-The primer used in these experiments was a human Harvey ras antisense 20-mer corresponding to positions 137-156 of human H-ras exon 111. The oligonucleotide was end-labeled using T4 polynucleotide kinase (New England Biolabs) and [-y-32P] ATP. The assay conditions were as follows: 5 pg of total RNA, 20 ng of labeled oligomer, 1 mM dNTPs, 50 mM Tris-HC1 (pH 8.31, 75 mM KCl, 10 mM MgC12, 0.5 mM spermidine, 20 mM dithiothreitol, 40 pg/ ml actinomycin D, and 4000 units/ml avian myeloblastosis virus reverse transcriptase were added to tubes in a volume of 50 pl. RNAs were heated in the presence of the primer at 95 "C for 5 min and quickly chilled on ice. The other reaction components were then added, and reactions were carried out for 2 h at 44 "C. After terminating the reactions with 5 p1 of 0.25 M EDTA, samples were extracted with phenol/chlorofo~ and pr~ipitated with ethanol. Pellets were dried, resuspended in formamide loading buffer, boiled 5 min, and immediately chilled on ice. Reverse transcripts were then fractionated by size on an 8% acrylamide, 8 M urea sequencing gel in 1 X TBE (90 mM Tris-HC1,SO mM NaB04, 1 mM EDTA (pH 8.3)). pUC18 plasmid digested with MspI and end-labeled with [y-32P]ATP served as a molecular weight marker. The gel was dried and exposed to X-Omat AR film at -70 "C with an intensifying screen to visualize the products.
Allele-specific PCR-cDNAs were synthesized from total RNA from normal and ras oncogene-transfected AtT20 cells and were [cP~*P]~CTP. subjected to PCR amplification with a sense strand primer, specific either for the wild type H-ras sequence or for the mutation found in the human c-Harvey ras oncogene clone EJ6.6 (Stork et al., 1991). A sense strand 13-mer matching the sequence of the human Harvey rm oncogene EJ6.6 (positions 1686-1699) was incubated with a conserved antisense 20-mer derived from exon I11 (see "Primer Extension" above). The allele-specific sense strand probe differs only in the terminal 3'-nucleotide from the wild type gene, a thymidine residue substituting for a guanine residue in the normal H-ras gene. cDNA was subjected to 30 cycles of a temperature program including 1 min at 94 "C, 2 min at 58 "C, and 1 min at 72 "C for the extension reaction in the presence of 2 units of TaqI DNA polymerase (Perkin-Elmer Cetus), four dNTPs at 0.25 mM, and 10 mM Tris-HC1 (pH 8.3), 50 mM KCl, 1.5 mM MgC12, and 0.01% gelatin, in 100 pl. Fifteen pl of each reaction mixture was analyzed by gel electrophoresis in 1.2% agarose in a Tris acetate buffer. The PCR products were then visualized by ethidium bromide fluorescence.
Preparation of Cell Protein Extracts and Western Blots-Extracts containing synapsin I were prepared from wild type or ras-transfected cells as described (Shapiro et al., 1988). Briefly, cells were lysed in hypotonic buffer using a Dounce homogenizer, osmolarity was restored, and the homogenate was spun at 12,000 x g for 30 s at 4 "C. The pellet was resuspended and rocked for 30 min at 4 'C, followed by centrifugation at 150,000 X g for 1.5 h at 4 'C. Protein was precipitated by addition of ammonium sulfate to 0.33 g/ml and recovered by centrifugation at 85,000 X g for 20 min at 4 'C. After resuspension of the pellet and dialysis, the Bradford protein assay (Bio-Rad) was used to determine protein concentration. To prepare cell lysates containing p21, cells were harvested and lysed by Dounce homogenization in 50 mM Tris-HC1, 1% Triton X-100 (pH 7.4). Lysates were centrifuged 12,000 X g for 15 min at 4 "C. The supernatant was collected and stored overnight at -20 "C, thawed, and respun at 1200 rpm in a microcentrifuge for 10 min at 4 "C. Supernatants were then assayed for protein concentration by Bradford assay. Extracts were fractionated on a 13% acrylamide, 0.1% SDS gel using a 5% acrylamide stacking gel. The electroblot transfer to BioTrace NT nitrocellulose (Gelman) was carried out in 25 mM Tris, 192 mM glycine, 20% methanol, 0.075% SDS for 2 h at 250 mA with cooling. Membranes were air-dried and blocked for >1 h in 3% bovine serum albumin/TN buffer at 20 "C. TN buffer is 20 mM Tris, 500 mM NaCl (pH 7.5). After blocking, membranes were incubated with primary antibody for 2 h at 20 "C, with shaking in 1% bovine serum albumin/TN. Membranes were then washed twice more for 10 min in TN, once for 10 min in TN/0.05% Triton X-100, and twice for 10 min in TN. '*'I-Protein A (final concentration, 1 pCi/ml) in 1% bovine serum albumin/TN was then added as secondary antibody, and membranes were incubated for 1 h at 20 "C with shaking. Membranes were washed as described above, dried, and exposed to Kodak X-Omat AR film to visualize protein bands. The anti-H-ras monoclonal antibody was RAS 10 (Carney et al., 1988), a gift of Walter Carney, Du Pont.

Establishment of AtT20 Cell Lines Transformed with the c-
Harvey ras Oncogene Clone EJ6.6-Plasmid DNA from the ras oncogene clone EJ6.6 (Shih and Weinberg, 1982), derived from a human bladder carcinoma cell line, was cotransfected into AtT2O cells with the selective marker pRSVNeo by calcium phosphate coprecipitation. Cells were selected for their resistance to the antibiotic G418. Unlike sham-transfected controls or cells cotransfected with pRSVNeo and other plasmids, many EJ6.6-transfected cells were observed to survive more than 2 weeks of selection in G418 without undergoing cell division. Their morphologies changed to larger cell bodies with extended and branched processes (Hemmick et al., 1992). Three stable G418-resistant colonies with similar morphology to the surviving nondividing cells were cloned. The frequency of stable transformants of the ras oncogene clone EJ6.6 with pRSVNeo was much lower than that of transfections of the same cells by other genes or by the nontransforming ras gene. This suggests that transfection of the EJ-ras gene results in cells that are unable to continue dividing and may cause terminal differentiation. The frequencies of G418-resistant colony formation in a representative transfection experiment are reported in Table I. EJ-ras Is Expressed in Transfected AtT20 Cells-To determine whether expression of the introduced ras oncogene was achieved in an EJ-ras-transformed AtT2O line, we quantified the amount of H-ras mRNA and p21 protein present in normal and ras-transfected AtT2O cells by primer extension reaction and Western blot. An oligonucleotide 20-mer complementary to a region of H-ras mRNA in exon I11 in the human H-ras gene was used as a specific primer. In the presence of reverse transcriptase and deoxyribonucleotides, this oligonucleotide is predicted to prime transcripts of 330 and 450 nucleotides with human H-ras mRNA as template (Ishi et al., 1986). Total RNA was analyzed from normal AtT2O cells, ras-transfected clone R1, and two human cell lines known to express endogenous H-ras, T24 and A431 (Fig. lA). When the relative intensities of the predicted primer extension products in total RNA from normal AtT2O cells and from ras-transfected clone R1 were compared, a 3-&fold stronger signal was reproducibly detected in the RNA from the ras-transfected clone (Fig. l A , lane 1 versus lane 2 ) . Primer extension products were also detected using RNAs from T24 and A431 cells as previously reported (Ishi et al., 1986). We find multiple primer extension products from RNAs from the human cell lines whereas only one product is observed in AtT20 cell RNAs. This indicates that the choice of transcription start sites of the endogenous mouse H-ras gene or the transfected EJ-ras gene, containing the human H-ras promoter, expressed in AtT20 cells differs from that of H-ras expressed in the human cells.
We also analyzed p21'"" expression at the level of immunoreactive protein on a Western blot. The anti-H-ras monoclonal antibody Ras 10 (Carney et al., 1988) was used as a probe to detect the presence of ~21'"" protein, followed by lz5Ilabeled protein A to visualize the bands. As a positive control on the Western blot, we used protein extracted from the human bladder carcinoma cell line T24. The results of this analysis (Fig. 1B) are consistent with an increase in the amount of ~21'"" protein present in the ras-transfected AtT2O cell line.
To determine whether the oncogenic EJ-ras allele is uniquely expressed in the transfected AtT2O clone, we applied a variation of the PCR sensitive to point mutations in the middle base of H-ras codon 12 (Stork et al., 1991). Under the PCR conditions we used, amplification is sensitive to single base mismatches at the 3'-nucleotide of the sense strand 14mer used in the PCR. Two sense strand oligonucleotides were synthesized that differ only at their 3"nucleotide corresponding to the middle base of codon 12 of the human H-ras gene.  Gorman et al. (1983). Plates were treated with 600 pg/ml G418 (GIBCO, 43% pure) for 14 days. and 5, ras-transfected AtT20 cells. PCRs were performed using a common downstream primer derived from the antisense strand of human Harvey ras gene and one of two allele-specific primers derived from the sense strand of the human ras gene (positions 1686-1698) terminating with a guanidine residue (protooncogene specific) (lanes 2 and 4) or a thymidine residue (oncogene specific) (lanes 3 and 5). pUC19, cut with DdeI, is shown in lane 1 as a marker. Note that only the PCR product in lane 5 (ras-transfected cells, oncogene-specific primer) shows a band on a 1% agarose gel corresponding to a DNA fragment having the predicted size. bp, base pairs. cells (Phillips and Tashjian, 1982). In order to determine whether ras oncogene-transformed AtT2O cells are capable of releasing ACTH in response to a known stimulus, normal AtT20 cells and ras-transfected clone R1 were treated with 8-Br-CAMP (1 mM) for 4 h. Aliquots of media were collected and assayed for their ACTH content by radioimmunoassay. As displayed in Fig. 2, panel A, normal AtT2O cells release a readily detectable level of ACTH in response to 8-Br-CAMP treatment. No detectable ACTH was released from the rastransformed cells under identical stimulating conditions. The loss of the ability of AtT2O cells to undergo stimulated release of ACTH, a characteristic feature of normal AtT2O cells, could be explained by several mechanisms. We speculated POMC mRNA-

Kb
Each primer was used in a PCR with a common antisense 20mer 240 nucleotides downstream (see "Experimental Procedures"). Expression of the mutant H-ras allele was identified in cDNA from the EJ-ras transfected R1 cells by the efficient amplification of the predicted 240-base pair fragment by PCR in the presence of the mutant allele-specific primer. Under these conditions, no detectable amplified product was observed with the mutant allele-specific primer when cDNA from wild type human or normal AtT2O RNA was used as template (Fig. IC). The wild type H-ras sense strand primer failed to generate a specific PCR product after 30 cycles in either nontransfected or ras-transfected AtT20 cells. However, with additional cycles, a band of the predicted size was observed, indicating the presence of wild type H-ras in both cell types at lower levels (data not shown). We therefore conclude that only the EJ-transfected cell line expresses the oncogenic H-ras allele.
Down-Regulation of ACTH Release and POMC Gene Expression in ras-transfected AtT20 Cell Lines"AtT20 cells synthesize and release the polypeptide hormone ACTH (Herbert et al., 1978). Exposure of AtT20 cells to a number of agents, including corticotropin-releasing factor and cyclic AMP analogues, induces rapid secretion of ACTH from AtT20

ras Alters Peptide Synthesis in AtT20
Cells 15467 that the loss of ACTH release was due at least in part to a decreased synthesis of the ACTH precursor POMC. To address this possibility we measured mRNA levels encoding POMC in normal and ras-transformed cells. Total RNA from normal AtT2O cells and three ras-transfected clones were assayed by Northern blot analysis with a "P-labeled mouse POMC cDNA probe. The results (Fig. 2, panel B ) indicate that the level of steady state POMC mRNA in two of the three ras-transformed clones has been reduced below the limits of detectability in the assay, estimated to be 100-fold less than that in normal ras cells. When this experiment was repeated using poly(A) RNA from clone R1, no POMC mRNA signal was detectable, even with extended x-ray film exposure times, which allowed the lower abundance high molecular weight POMC mRNA precursors to be visualized in the lane corresponding to wild type AtTZO cells (Fig. 2,panel C ) . These results are consistent with the loss of stimulated ACTH release and the down-regulation of POMC mRNA occurring at the level of transcription of the POMC gene.
Comparison of Expression of Other Genes in Normal and ras-transformedAtT20 Cells-To determine whether the large change in expression of the POMC gene represents a generalized effect of ras oncogene transformation, we monitored the expression of several other endogenous genes. In particular, we assayed for the expression of two genes that are ubiquitously expressed in many cell types, the genes encoding y-actin and cyclophilin. Using the human cDNA clone for yactin and the cDNA for cyclophilin as hybridization probes, we found that the levels of y-actin and cyclophilin mRNA levels in normal and ras-transformed AtT20 cells differ only slightly (Fig. 3, A and B ) . The level of y-actin mRNA (Fig.  3A) appears to be increased 2-%fold in ras-transformed cells probed with cyclophilin cDNA, as in panel A. C, Western blot assay of synapsin I protein content in normal AtT20 cells and ras oncogenetransfected clone R1. Protein extracts (20 pg) from wild type (lane 1 ) and ras oncogene-transfected clone R1 (lane 2 ) were run on a Laemmli SDS gel and transferred to nitrocellulose. The primary antibody was a rabbit monoclonal antibody G-94 (a gift of Martin Bahlin, Rockefeller University) directed against synapsin I. '*'I-Protein A was used as the secondary antibody. Autoradiography revealed that both the wild type and rm-transfected cells exhibited a band at 75 kDa, as determined by molecular weight markers (not shown). over wild type AtT2O cells. The level of cyclophilin mRNA remains essentially unchanged (Fig. 3B).
We also measured levels of a differentiated gene product thought to be expressed only in neuronal and neurosecretory cell types, synapsin I. Protein extracts from wild type AtT20 cells and clone R1 were assayed by Western blot with a monoclonal antibody specific to synapsin I. As with y-actin and cyclophilin, there is only a small difference in expression between the two cell types, with synapsin I appearing slightly decreased in the EJ-ras-transformed cells (Fig. 3C). We conclude from these data that the ~21'"" oncoprotein specifically affects expression of the POMC gene in AtT2O cells transformed with the EJ-ras gene.

DISCUSSION
ras has been shown previously to mediate differentiation in one neuronal cell line, PC-12 (Noda et al., 1985). We investigated the effects of expression of the H-ras oncogene on the phenotype of the pituitary mouse cell line AtTZO. We cotransfected pEJ6.6, the EJ-ras gene, into AtT2O cells with the selectable marker pRSVNeo. A small number of G418-resistant clones were obtained. Compared with the frequency of stable transfection when the normal H-ras gene or other plasmids are used, a reduced frequency of appearance of G418resistant clones with the EJ-ras oncogene was observed (Table  I). This is consistent with expression of EJ-ras in transfected AtT2O cells causing terminal differentiation. In the three G418-resistant clones that were established, a number of distinguishing features were observed. The ras-transformed cells and nondividing transfected cells that survived in 600 pg/ml G418 for 2 weeks could be distinguished morphologically from nontransfected cells by their larger size and extensive process growth (Hemmick et al., 1992). A number of retrovirus-infected AtT20 clones expressing the v-H-ras gene also displayed this altered morphology. ' We show that the ras gene is overexpressed in transfected AtT2O cells at the mRNA level and at the level of p21'"" protein ( Fig. 1). With the application of the technique of allele-specific PCR, we provide evidence that only the rastransfected AtT20 cells express the point mutation in codon 12 found in the EJ-ras gene (Fig. IC).
Stimulated release of ACTH in response to 8-Br-cAMP, a hallmark of normal of AtT20 cells, was found to be missing in a ras-transfected morphologically transformed clone ( Fig.  2.4, for photomicrographs see Hemmick et al. (1992)). We hypothesized that expression of the gene encoding the ACTH precursor proopiomelanocortin was down-regulated. Results from Northern blot analysis support this hypothesis (Fig. 2,   B and C ) . A Northern blot of poly(A)+ RNA from wild type AtT2O and clone R1 cells was hybridized to POMC cDNA and exposed to x-ray film long enough to visualize the lower abundance POMC mRNA precursor (Fig. 2C). Under these conditions no hybridization signal was detected in the rastransfected clone. This suggests the block to POMC expression is at the level of transcription. A profound change in POMC mRNA half-life from 12-24 h (Birnberg et al., 1983) to minutes could give rise to significantly reduced steady state mRNA levels. However, the absence of any detectable POMC pre-mRNA species is more consistent with a block at the level of transcription. The virtual extinction of POMC mRNA levels in two of three ras-transformed clones analyzed was never observed in normal AtT2O cells or in clones transfected with genes other than EJ-ras. Three control genes, one neuralspecific, synapsin I, and two housekeeping, y-actin and cyclo-

Synthesis in AtT20 Cells
philin, assayed at the RNA or protein level, were found to be largely unchanged (Fig. 3). The large effect of activated rus on gene expression in AtT2O cells appears restricted to a subset of expressed genes. The down-regulation of expression of the POMC gene may be by modulation of the activity of sequence-specific transcription factors or an accessory protein that interacts with the transactivation domain of a DNAbinding protein.
The proopiomelanocortin promoter is composed of a large number of synergistically interacting DNA elements (Therrien and Drouin, 1991). It is possible that one or more transcription factor proteins that bind to the POMC gene are inactivated or reduced in their expression as a result of transformation by the EJ-ras gene. We are currently making ndclear extracts from normal and ras-transformed AtT20 cells to examine the complement of binding activities to different cis-acting elements from the POMC promoter.
In studies of the effects of activated rus on fibroblast cell lines, rus transformation results in de-differentiation and increased cell proliferation (Barbacid, 1987). In contrast, PC12 cells, an adrenal chromaffin cell line, are induced to terminal differentiation by nerve growth factor or following introduction of rus oncogenes or ~2 1 ' "~ oncoprotein. This has been demonstrated by retroviral infection, by transfection with inducible rus genes, and by microinjection (Thomson et al., 1990;Noda et al., 1985;Bar-Sagi and Feramisco, 1985;Guerrero et aZ., 1988). PC12 cells represent a well characterized model for a neurotrophin-responsive cell type, terminally differentiating to sympathetic neurons following nerve growth factor treatment. This differentiation can be inhibited or reversed by microinjection of the neutralizing anti-Ras monoclonal antibody Y13-259 (Hagag et al., 1986;Furth et al., 1982) thereby implicating a ras-dependent signaling pathway as a necessary component in the action of a neural differentiating agent.
We have reported here that a profound down-regulation in polypeptide hormone gene expression occurs in AtT2O cells stably transformed by EX-rus. We recently reported on alterations in electrophysiological properties and the induction of two potassium channel mRNA species in ras-transformed cells (Flamm et al., 1990;Hemmick et al., 1992). The prepro-grammed responses of PC12 cells and AtT2O cells to the intracellular second messenger pathway mediated by ras provide interesting models to study the regulation of differentiation by a neurotrophin-mediated intracellular signaling pathway common to neuroendocrine cells.