Immediate early genes induced by H-Ras in thyroid cells

Abstract

Expression of oncogenic v-H-Ras in the thyroid cell line FRTL-5 (FRTL-5Ras) results in uncontrolled proliferation, loss of thyroid-specific gene expression and tumorigenicity. Concomitant expression of constitutively activated MEK and Rac, two major H-Ras downstream effectors, in FRTL-5 (FRTL-5MEK/Rac) recapitulates H-Ras effects on proliferation and morphology. In contrast to FRTL-5Ras, however, FRTL-5MEK/Rac cells remain differentiated and are not tumorigenic. To find H-Ras induced genes potentially responsible for tumorigenicity and loss of differentiation, we have used subtractive suppression hybridization (SSH), a PCR-based cDNA subtraction technique, between de-differentiated and tumorigenic FRTL-5Ras cells and differentiated and non-tumorigenic FRTL-5MEK/Rac cells. We examined 800 of the cDNA clones obtained after subtraction and verified their levels of expression in the two cell lines by reverse northern, identifying 337 H-Ras induced genes. By sequence analysis, we clustered 57 different genes. Among these, 39 were known genes (involved in diverse signal transduction processes regulating mitogenic activity, cell survival, cytoskeletal reorganization, stress response and invasion) while the remaining 18 clones were novel genes. Among the 57 H-Ras specific clones, we identified those genes whose expression is induced early by H-Ras. We suggest that these immediate-early genes may play a crucial role in H-Ras-mediated transformation in thyroid epithelial cells.

Introduction

Expression of an activated ras oncogene in rodent fibroblast cell lines induces a pleiotropic response, including alterations in cell morphology, loss of contact inhibition, stable changes in gene expression, decreased dependence on growth factors and the ability to proliferate in the absence of adhesion to a substratum (Shields et al., 2000).

The molecular responses elicited by activated H-Ras have been extensively studied and a number of H-Ras downstream effectors have been identified. The Raf/MEK/MAPK cascade is thought to mediate the growth-promoting activity of H-Ras (Shields et al., 2000). The Rho/Rac family of GTP-binding proteins plays crucial roles in the reorganization of the actin cytoskeleton and cell adhesion (Ridley and Hall, 1992; Ridley et al., 1992; Nobes and Hall, 1995). The phosphatidylinositol 3-kinase (PI3-K) is another downstream effector of H-Ras, and is important in controlling apoptosis (Rodriguez-Viciana et al., 1994, 1997; Datta et al., 1999).

The signaling process by which H-Ras activation elicits different responses has been studied mostly in fibroblasts, while little is known about similar processes in epithelial cells, where H-Ras mutations are most frequently found (Bos, 1989). Several studies have reported that H-Ras activation is an early event also in thyroid tumorigenesis (Farid et al., 1994). Expression of oncogenic v-H-Ras in the thyroid follicular cell line FRTL-5 (FRTL-5Ras) induces growth-factor independent proliferation, morphological transformation and anchorage-independent growth (Francis-Lang et al., 1992). Furthermore, subcutaneous injection of FRTL-5Ras in nude mice results in tumor formation (Fusco et al., 1987). Another outstanding consequence of v-H-Ras expression in FRTL-5 cells is the down-regulation of thyroid-specific genes such as thyroglobulin (TG), thyroperoxidase (TPO) and the sodium-iodide symporter (NIS) (Francis-Lang et al., 1992).

We have previously characterized the contribution of two major pathways activated by oncogenic v-H-Ras: the Raf/MEK/MAPK cascade and the Rac cascade, since it has been demonstrated that concomitant activation of these pathways is sufficient to recapitulate the H-Ras effects in fibroblasts (Cowley et al., 1994; Mansour et al., 1994; Qiu et al., 1995). We have demonstrated that concomitant expression of constitutively activated MEK-1 and Rac-1 in FRTL-5 cells (FRTL-5MEK/Rac) is able to reproduce the H-Ras effectson cell proliferation and morphology. In fact,FRTL-5MEK/Rac cells are growth factor- and anchorage-independent. However, at variance from FRTL-5Ras, the FRTL-5MEK/Rac cells they are still differentiated and do not form tumors in nude mice (Cobellis et al., 1998).

The molecular mechanisms leading to anchorage-independent growth in vitro are closely related to the progressive altered growth properties of naturally occurring tumors in vivo. The persistent expression of differentiated phenotype in MEK/Rac-expressing cells correlates with the non-tumorigenic phenotype. Thus, our previous studies suggest that the FRTL-5MEK/Rac cells may represent a ‘pre-tumorigenic’ step, which still requires additional signals before developing into a full malignant transformation.

The molecular mechanisms by which H-Ras causes tumors and blocks the thyroid differentiation are still poorly understood. Thyroid-specific gene expression is, in large part, due to the combinatorial contribution of three major transcription factors, TTF-1, TTF-2 and Pax8 (Damante and Di Lauro, 1994). In FRTL-5Ras cells, both Pax8 and TTF-2 mRNA are down-regulated; in contrast, TTF-1 protein is present at nearly wild-type levels and its DNA binding activity is normal, although the protein is not competent of transcriptional activation (Francis-Lang et al., 1992). We recently demonstrated that H-Ras interferes directly with the activation of TTF-1 through two complementary mechanisms, one involving activation of the Raf/MEK/ERK cascade, and the other involving an independent and yet uncharacterized pathway triggered by the V12N38 H-Ras mutant (Missero et al., 2000). Thus, it appears that oncogenic v-H-Ras is likely to require several pathways to lead to full malignant transformation in thyroid cells and that one or more of such pathways may represent novel ones.

The genetic complexity of cellular transformation at the level of gene expression was first described more than 10 years ago (Groundine and Weintraub, 1980; Augenlicht, 1987). The activation of different signaling pathways results in the transcription of a set of genes, whose induction may require de novo protein synthesis or activation of latent transcriptional activators.

Several gene regulators, such as the ETS-domain transcription factor Elk1, the serum-responsive factor (SRF), the leucine zipper protein JUN, the c-fos transcription factor, the activation transcription factor 2 (ATF2) and nuclear factor-κB (NF-κB) are stimulated by signaling pathways downstream of H-Ras, indicating that oncogenic H-Ras affects a complex set of transcriptional targets (Khosravi-Far et al., 1998; Malumbres and Pellicer, 1998).

To identify downstream genes in the H-Ras signaling pathway, and especially those that could play a role in thyroid tumor development and progression, we performed a suppression subtractive hybridization (Diatchenko et al., 1996) between FRTL-5Ras and FRTL-5MEK/Rac mRNAs. These cells were chosen because they differ in tumorigenicity and in differentiation, even though they have a similar altered growth potential (Fusco et al., 1987; Francis-Lang et al., 1992; Cobellis et al., 1998). In this study, using SSH approach, we isolated 57 different genes, 18 of which represent novel sequences, which are expressed at higher level in FRTL-5Ras than in FRTL-5MEK/Rac cells. Among these genes, we identified those that are induced early by oncogenic H-ras. This represents the first attempt to identify the first wave of gene expression induced by oncogenic Ras in epithelial cells.

Results

Generation of a subtractive library and screening of differentially expressed clones

We used the SSH technique to search for genes differentially expressed between FRTL-5Ras and FRTL-5MEK/Rac cells (Cobellis et al., 1998).

The mRNA samples from these two cell populations were reverse transcribed and subjected to the SSH procedure as described by Diatchenko et al. (1996). Expression of the GAPDH gene, present in both cell populations, and of the GFP gene, transcribed in vitro and added only to the FRTL-5Ras population, was evaluated to monitor the subtraction efficiency. Southern blot hybridization analysis with a GAPDH probe showed that this transcript was detectable in the cDNA mixture before the subtraction step and disappeared after the subtraction, indicating efficient removal of the common sequences (Figure 1a). Furthermore, Figure 1b shows that the GFP gene was highly enriched in the subtracted cDNA population. This analysis clearly demonstrated that the subtraction between these two cDNA populations occurred properly.

Figure 1
figure1

Evaluation of subtraction efficiency: Southern blot analysis of equal amount of unsubtracted and subtracted cDNA tester PCR products (100 ng) were fractionated on 1.4% agarose gels, blotted and hybridized with α-32P-CTP-labeled GAPDH and GFP probes, as described in Materials and methods. (a) the expression of the housekeeping GAPDH cDNA is drastically reduced between unsubtracted and subtracted cDNA. (b) the abundance of the GFP cDNA, used as a tracer and present only in FRTL-5Ras population, is enriched after the subtraction step

A total of 800 individual clones were considered in this study. To verify the differential expression of these clones, an initial screening was performed by ‘reverse Northern’, a technique that allows rapid identification of clones harboring differentially expressed cDNAs. Colony PCRs were performed for the 800 clones and the PCR products were immobilized on nylon filters. The filters were then hybridized with radiolabeled cDNA probes from either tester (FRTL-5Ras) or driver (FRTL-5MEK/Rac) mRNA populations. Representative signals obtained on 96 clones hybridized with cDNA probes from either of the two cells populations are shown in Figure 2. Filters were exposed to films for up to 4 days and the signals of identical clones were compared. Quantitative analysis was obtained by Phosphor Imager. A clone was considered a candidate H-Ras target if it was induced at least twofold in the filter hybridized with FRTL-5Ras specific cDNA, compared to the filter hybridized with FRTL-5MEK/Rac cDNA. Of the 800 clones that were analysed, 337 (42%) showed a stronger hybridization signal with radiolabeled cDNA derived from FRTL-5Ras cDNA, suggesting that the cognate mRNAs are induced by H-Ras. 160 (21%) showed a hybridization signal with both cDNA probes, indicating that some common cDNAs are still present after the subtraction. Clones that failed to hybridize (37–302%) may represent extremely rare transcripts and thereby lie below the sensitivity limit of reverse Northern. Twenty-four clones (3%) did not contain any inserts (Table 1).

Figure 2
figure2

Analysis of subtracted clones by reverse Northern: Colony PCRs were performed as described in Materials and methods and the products resolved on 1.5% agarose gels in parallel, blotted and hybridized with tester and driver radiolabeled single-stranded cDNA of equal specific activity. The hybridization was performed under stringent conditions (see Materials and methods for details) and after the washing, the filters were first exposed over night to the Phosphor Imager to quantify the signals (see Materials and methods) and then exposed to autoradiography film at −80°C for up to 4 days. (a) Hybridization with Ras cDNA probe of 96 representative clones is shown; (b) the same clones hybridized with the MEK/Rac cDNA probe. The arrows indicate those clones that show differential hybridization between the two cDNA probes, while the arrowheads indicate common clones that hybridized with both cDNA probes. This analysis was performed for all 800 clones

Table 1 Summary of the Ras-subtracted library

Comparison of output sequences to the databases

To characterize the 337 differentially expressed clones, the inserts were automatically sequenced and the output sequences were compared to EMBL, Swisspro and dbEST databases in order to find homologies with already known genes. After sequence comparison and clustering analysis, we found that the 337 H-Ras-induced clones clustered into 57 unique genes. As shown in Table 2, most of them were identified as known genes, while others did not show any homology to currently known sequences (18 clones–5%). Among the known genes, we found some that were previously reported as H-Ras targets, such as the cDNAs encoding the metastasis-related glicoprotein gp55, the metalloproteinase Mmp-1 and the myosin regulatory light chain. Furthermore, we found genes involved in stress response (pro-apoptotic protein SIVA, thymosin beta-4, superoxide dismutase and ERCC-1), in cytoskeletal reorganization (vimentin, actin and fibronectin), in mitogenic activity (p15 ink4), signaling molecules (CDC42 or Ca2+ binding protein), transcriptional effectors (E2F-6) and cytoplasmic regulatory proteins (pyruvate kinase, MKP-5).

Table 2 H-Ras-induced genes in thyroid cells

Thus, the repertoire of genes revealed that H-Ras affect many aspects of cell physiology altered in transformed cells, such as cell growth, cell cycle, adhesion and cytoskeletal organization.

Kinetic study of differentially expressed clones

To determine the expression kinetic of these clones, we developed a strategy to follow the time of induction of these clones upon H-Ras expression. To this end, we generated an adenoviral vector carrying H-Ras gene in order to achieve an efficient and acute expression of H-Ras in thyroid cells upon viral infection. Attempts to obtain an adenovirus carrying V12 H-Ras failed because elevated expression of V12 H-Ras was cytotoxic in HEK293 cells. Thus, we prepared two distinct adenoviruses carrying the mutations S35 and C40 in the V12 H-Ras background. The V12S35 mutant has been reported to efficiently activate the Raf/MEK/MAP kinase cascade but not other effectors, whereas V12C40 activates only the PI3K/Rac cascade (White et al., 1995).

FRTL-5 cells were infected with H-Ras adenoviruses at a m.o.i. of about 10 p.f.u./cells. As control, we used in a separate infection an adenovirus encoding for the GFP gene (AdGFP). Cells were incubated for different times (0, 4, 8, 12, 16, 20, 24 and 48 h) upon infection and the H-Ras expression levels were evaluated by Western blotting analysis.

As shown in Figure 3A (panels a and d), high levels of H-Ras protein were achieved within 8 h upon AdS35Ras and AdC40Ras infection, and H-Ras expression remained constant up to 48 h. The H-Ras protein was undetectable in FRTL-5 cells infected with AdGFP for 48 h or in uninfected FRTL-5 cells (Figure 3A, panels a and d).

Figure 3
figure3

Kinetic analysis of AdRas expression and activity upon viral infection. (A) FRTL-5 cells were incubated upon infection for different times (4, 8, 12, 16, 20, 24 and 48 h) with AdS35Ras and in parallel with AdGFP. Cells were then lysed in sample buffer and equal amount of proteins were run in parallel on a 15% and on 10% SDS–PAGE gels, and blotted on a PVDF membrane. After transferring, the membranes were used to perform Western blotting analysis. Immunoblotting analysis was performed with the anti-H-Ras (panel a; F235-Santa Cruz Biotechnology), with the anti phospho-specific ERK antibodies (panel b; E4-Santa Cruz Biotechnology) and with anti-total ERK (C14, C16-panel c) to normalize the protein amount. Similarly, FRTL-5 cells were incubated upon infection for different times (4, 8, 12, 16, 20, 24 and 48 h) with AdC40Ras. Protein samples were subjected to immunoblotting analysis with the anti-H-Ras (panel d; F235-Santa Cruz Biotechnology), with the anti phospho-specific ERK antibodies (panel e; E4-Santa Cruz Biotechnology) and with anti-total ERK (C14, C16-panel f) to normalize the protein amount. (B) Total cell extracts from wild-type, chronically H-Ras expressing cells and S35, C40 and GFP-infected cells were analysed by a 4–15% gradient SDS-polyacrilamide gels and immunoblotted with thyroglobulin and Pax8-specific polyclonal antibodies (TG and Pax8). The300-kDa thyroglobulin band is detected in FRTL-5 and in GFP-infected cells, while a drastic reduction was observed in H-Ras expressing cells and in S35 and C40 H-Ras -infected cells. The filter was probed with antibodies against Pax8: the protein is detectable in wild-type and GFP-infected cells, while it is reduced in H-Ras and S35 C40-infected cells

To examine the effect of S35Ras expression, we evaluated the activation of the MAPK cascade using antibodies against the phosphorylated isoforms of ERK1/2. As shown in Figure 3A (panel b), a sustained ERK1 and ERK2 phosphorylation was induced at 8 h upon infection at the same time in which H-Ras started to be expressed. A similar induction was observed with the AdC40Ras (Figure 3A, panel e). The C40 mutant is known to predominantly activate the PI3K/Rac cascade, so little ERK activation should be obtained with this mutant. It is likely that the high levels of Ras expression obtained by adenoviral infection results in loss of specificity of these two mutants.

Then we evaluated the effects of the adenoviruses on thyroid-specific gene expression. Cell extracts from chronically H-Ras expressing cells and S35 and C40 Ras infected cells were compared for the expression of TG and Pax8. As shown in Figure 3B (middle panel), very low levels of TG protein were achieved in S35 and in C40 infected cells, similarly to the H-Ras expressing cells. Moreover, a similar down-regulation of Pax8 protein was observed in S35 and C40 infected cells as in H-Ras expressing cells. (Figure 3B, lower panel). No decreases in expression of these genes were detected in control GFP-infected cells.

These data suggest that cells infected with S35 and C40 H-Ras adenoviruses behave similarly to the stable transfected ones, at least as judged by ERK activation and thyroid-specific gene expression.

Thus, we decided to use AdS35Ras and AdC40Ras as tools to investigate changes in gene expression induced by activated H-Ras. RNAs were collected from FRTL-5 cells infected with AdS35Ras and AdGFP at different times after infection (8, 16, 24 and 48 h). From each time point, we purified mRNA and prepared a radiolabeled cDNA probe. These probes were used in a reverse Northern analysis for the 57 isolated clones. Signal intensity of the various clones was quantified by Phospho-Imager and normalized against that of GAPDH, which remained constant in AdS35Ras and AdGFP infected cells. Expression changes of at least threefold were scored as positive. Figure 4 shows the results of the reverse Northern analysis performed with cDNA probes from AdRas and AdGFP infected cells at 8 h. The differentially expressed genes classified with respect to their time of induction are listed in Table 3. Similar experiments were performed using the C40 mutant as a probe, which gave analogous results (data not shown).

Figure 4
figure4

Time course of induction of the H-Ras-induced clones by reverse Northern analysis. Colony PCRs were performed as described in Materials and methods and the products resolved on 1.5% agarose gels in parallel, blotted and hybridized with probes prepared from poly-(A)+ RNA isolated from FRTL-5-infected cells at different times (8, 16, 24 and 48 h). Hybridization performed with probes prepared from infected cells with AdRas (a) and AdGFP (b) at 8 h after the infection. Signal intensity was quantified with the Instant Imager software and the ratio of the net counts calculated (AdRas vs AdGFP) after GAPDH normalization. The arrows indicate those clones that show differential hybridization between the two cDNA probes, while the arrowheads indicate common clones that hybridized with both cDNA probes. The asterisk indicates the signal intensity for the GAPDH band, used for normalization. This analysis was performed for the 57 H-Ras-induced clones at different times of infection, namely at 8, 16, 24 and 48 h

Table 3 Time course of induction of Ras-induced clones

Northern analysis of early induced clones

A role for H-Ras in transcriptional regulation has been well documented (Zuber et al., 2000). Although we sought to explore the contribution of H-Ras expression from a more global view, we focused our attention on those clones that are early induced by H-Ras, thinking that these genes may be capable of late events responsible of transformation. H-Ras expression caused induction of 12 clones out of 57 within 48 h of infection (Table 3). Six of them met the criteria of being immediate early genes as judged by reverse Northern, since the induction was observed as early as 8 h after infection, at the same time when the Ras protein is produced. We performed conventional Northern analysis to validate the results of the reverse Northern since in many studies it has been clearly shown that the extent of differential expression often varies with these two diverse experimental approaches. Furthermore, we wanted to examine again the time course of induction and to confirm that these genes were differentially expressed in the two original populations, i.e. H-Ras- and MEK/Rac-expressing cells (Figure 5). All the six clones showed differential expression between the two original cell populations. Concerning their time of induction, three genes, cl. 25, 41 and 55, were induced at 8 h in Northern analysis, while the others were induced at 16 h upon infection (Figure 5). Clone 25 corresponded to ranp-1, a gene previously identified in neuronal cells involved in stress response, being induced during stress conditions (axotomy) (Uwabe et al., 1997). The clone 41 encoded for the chemoattractant protein-1 that it is known to be overexpressed in tumoral cells (Yoshimura et al., 1991). Clone 55 encoded for E2F-6, a protein that belongs to the E2F family of transcription factors and that has been proposed to function as a transcriptional repressor (Cartwright et al., 1998; Gaubatz et al., 1998; Trimarchi et al., 1998). This clone showed elevated expression levels in H-Ras-expressing cells, compared to the wild-type cells, while no signal was detectable in FRTL-5Mek/Rac cells (Figure 5).

Figure 5
figure5

Northern analysis of early induced clones: Two micrograms of poly(A)+ RNA were resolved on formaldehyde-agarose gel and then transferred to a nylon filter. The hybridizations were performed in ULTRAhyb buffer (Ambion Inc.). The poly(A)+ RNA was isolated from the original cell populations FRTL-5, FRTL-5Ras and FRTL-5MEK/Rac cell lines, to confirm the differential expression. Poly(A)+ RNA were isolated from AdRas infected cells at the times indicated in the top panels and from AdGFP infected cells after 48 h. The probes were prepared as described in Materials and methods. Filter's hybridizations performed with labeled probes of the indicated clones (cl. 17, 25, 37, 41, 42 and 55) are shown in the top panels; filter's hybridizations performed with the GAPDH probe are shown in the lower panels

Thus, the kinetic analysis revealed, as early clones, genes involved either in the regulation of the transcriptional machinery and in the reorganization of the cytoskeleton.

Discussion

Mutations in the ras proto-oncogenes are frequently found in epithelial cell tumorigenesis and result in uncontrolled cell proliferation, morphological transformation and loss of the differentiated phenotype.

In the thyroid cell line FRTL-5, expression of activated H-Ras results in suppression of thyroid-specific gene expression and tumorigenesis (Fusco et al., 1987; Francis-Lang et al., 1992). Little is known about the molecular mechanisms responsible for these effects in epithelial cells. We have previously reported that stable expression of an activated form of MEK-1 and Rac-1 had no effect on thyroid-specific gene expression. Concomitant expression of MEK and Rac, however, recapitulates the H-Ras effects on growth potential, morphology and anchorage-independent proliferation (Cobellis et al., 1998). The expression of oncogenic H-Ras results in the inactivation or reduced expression of the thyroid-specific transcription factors, TTF-1, TTF-2 and Pax8 (Missero et al., 1998). Recently, we have demonstrated that TTF-1 is phosphorylated by ERK proteins upon H-Ras activation (Missero et al., 2000). However, it has not been demonstrated that this modification is responsible for TTF-1 inactivation. Others have investigated the effects of the PI3K pathway in the same thyroid cell line, although the activation of this pathway is important in controlling proliferation of FRTL-5 cells, no effects on thyroid specific gene expression have been reported (Cass and Meinkoth, 2000).

Thus, the identification of H-Ras downstream targets could allow isolation of genes responsible for loss of differentiation and tumorigenicity in thyroid cells.

Using SSH technique, we have compared gene expression differences between tumorigenic FRTL-5Ras cells and non-tumorigenic FRTL-5Mek/Rac cells in order to isolate genes that could be induced by H-Ras and responsible for tumorigenicity in thyroid cells. We have analysed 800 clones and isolated by reverse Northern analysis 337 differentially expressed clones, corresponding to 57 different genes by sequence analysis. We postulated that putative candidates responsible for thyroid tumorigenicity might be present among the identified clones. Furthermore, we have focused on those genes that are early induced by H-Ras, reasoning that early targets of H-Ras should interfere with the transcriptional machinery of the host cells inducing other genes, which in turn may be responsible for cellular transformation. To this purpose, we performed a kinetic analysis using an adenoviral vector expressing H-Ras to define the time of induction of the selected clones. From such an analysis, we recovered 12 clones that are induced by H-Ras within the first 48 h after H-Ras expression. The identification of such clones indicates that H-Ras induces the expression of genes involved in a variety of cellular processes, such as cytoskeletal organization, signal transduction and transcription. Among these, we have isolated cDNAs encoding the transmembrane proteins gp55, a membrane glicoprotein reported as a cell-surface molecule that is considered as a tumoral antigen (Langnaese et al., 1997). We also found cDNA encoding for secreted proteins, such as MCP-1 (chemoattractant protein-1) (Yoshimura et al., 1991) and MMP-1 (metalloproteinase-1). All these proteins have been reported to be involved in the tumoral invasion. Other genes, found among those that are early induced by H-Ras, are involved in the cellular stress response (Ranp-1, p58NK receptor and ERCC-1). No information is available concerning the clone 37; the output comparison in the databases scores a match with a human PAC clone from chromosome 7q11.2-p12, while the full-length sequence comparison matches with a protein containing a kinesin domain. It is known that mutations in this domain lead to constitutive activation of these proteins resulting in many pathologies, including cellular transformation (Ciccarelli et al., 2000). Another clone induced at 8 h (twofold higher) encodes the rat homolog of human E2F-6 (clone 55). This protein belongs to the E2F family of transcription factors. These proteins play an important role in the regulation of cell cycle progression. In comparison with all other E2F species, E2F-6 possesses an unusual biological activity, behaving as a strong transcriptional repressor (Cartwright et al., 1998; Gaubatz et al., 1998; Trimarchi et al., 1998). It is not yet known if this protein could have an effect on differentiation, but its biological function may represent a link between cellular growth and differentiation control.

Our study is the first attempt to identify H-Ras targets in a differentiated epithelial cell type. It is clear that the H-Ras-mediated effects could be different in other cell types. Nevertheless, among the genes found in this subtraction analysis, we recovered several genes common to a similar analysis performed with SSH technique in fibroblasts (Zuber et al., 2000), suggesting that at least some of the genes induced by H-Ras are common in different cell types. Furthermore, at difference of Zuber et al. (2000) we further characterized the H-Ras targets using an experimental approach that led to the identification of immediate early H-Ras-induced genes. The kinetic analysis revealed 12 out of 57 mRNA induced at early times by H-Ras in our cells.

In the identification of differentially induced genes performed by reverse Northern, we found 270 clones that failed to hybridize. It is conceivable that these clones correspond to mRNAs that are too rare to be detecting by the reverse Northern. Further sequence analysis of these clones will possibly reveal other potentially interesting cDNAs. Overall, the data presented here represent a valuable effort to define the complex network that is behind the pleiotropic effects exerted by H-Ras oncogene in a differentiated cell model system, such as the FRTL-5.

Materials and methods

Cell culture

FRTL-5, FRTL-5Ras and FRTL-5MEK/Rac cells were grown in Coon's modified F-12 medium (Sigma) supplemented with 5% calf serum (Life technologies, Inc.) and six growth factors including TSH (1 mu/ml) and insulin (10 mg/L), as previously described (Ambesi-Impiombato et al., 1982).

Immunoblotting

Whole-cell lysates of stably transfected cells were prepared in sample buffer, normalized for equal concentration of protein by the Bradford assay (BioRad). Thirty micrograms of protein samples were resolved on SDS–PAGE at the concentrations indicated in the figure legends, and transferred on a PVDF membrane (Millipore). Anti-H-Ras (F235), anti-ERK1 (C-16), anti-ERK2 (C-14) and anti p-ERK (E-4) were purchased from Santa Cruz Biotechnology, Inc. and used at a 1 : 100 dilution. Affinity purified polyclonal Ab anti-thyroglobulin and anti-Pax8 were obtained in our laboratory and used approximately at 1 μg/ml. Immune complexes were detected by enhanced chemiluminescence as instructed by the manufacturer (Amersham Life Sciences, PLC).

RNA isolation and Northern blot analysis

Total RNA was isolated from cultured cells using the Trizol reagent as instructed by the manufacturer (Life Technologies, Inc.). Purification of poly-(A)+ RNA was performed using Dynabeads Oligo(dT)-25 according to the manufacturer's instructions. Two micrograms of poly-(A)+ were run on a 1% formaldehyde-agarose gel and transferred to Hybond N+ nylon (Amersham Life Sciences) and hybridized in the ULTRAhyb buffer (Ambion Inc.) according to the manufacturer's instructions. Nylon blots were probed with a 300 bp cDNA fragment in the 3′ region of the rat GAPDH, and with 700 bp fragment encoding the entire GFP gene, labeled by random priming with α-32P-dCTP (Amersham Life Science).

Probes from each clone were prepared from PCR amplification using M13 and T7 primers present in the vector and labeled by random priming.

Generation of a subtracted library by SSH

SSH was performed between FRTL-5Ras and FRTL-5MEK/Rac cells (tester and driver cDNAs, respectively) using the PCR-Select cDNA Subtraction Kit (Clontech). The cDNA was synthesized from 2 μg of poly-(A)+ RNA from the two cells type. The tester and driver cDNAs were digested with RsaI, a four-base-cutting restriction enzyme that yielded blunt ends. The tester cDNA was then subdivided into two portions, and each was ligated with a different cDNA adaptor (Adaptor 1 and 2R). The ends of the adaptor did not have a phosphate group, so only one strand of each adaptor attached to the 5′ ends of the cDNA. The two adaptors had stretches of identical sequence to allow annealing of the PCR primers.

Two hybridizations were then performed. First, an excess of driver was added to each sample tester. The samples were then heat denatured and allowed to anneal. The concentration of high- and low-abundance sequences was equalized because reannealing is faster for the more abundant molecules due to the second order kinetic of hybridizations. During the second hybridization, the two primary hybridization samples were mixed together without denaturing. Fresh denatured driver cDNA was added to further enrich fraction of differentially expressed molecules. The entire population of molecules was then subjected to PCR to selectively amplify the desired differentially expressed sequences, which had the two different adaptors, suppressing the amplification of the other molecules. Next, a secondary PCR amplification was performed using nested primers to further reduce any background PCR products. All PCR and hybridization steps were performed on a Perkin Elmer 9600 thermal cycler.

Cloning of cDNA mixture

Approximately 100 ng PCR-amplified cDNA were ligated without further purification into 10 ng of PCR 2.1 TOPO cloning vector (Invitrogen). The ligation mixture was introduced into electrocompetent Top 10 F′ cells by electroporation (1.7 kV) using an E. coli pulser (BioRad). The library was plated onto 150 mm2 ampicillin agar plates containing 100 mM IPTG and 50 mg/l X-Gal and bacteria were grown overnight at 37°C. Plates were then incubated further at 4°C until blue/white staining could be clearly distinguished.

Reverse Northern blots and screening

A total of 800 individual recombinant clones were picked and used to inoculate eight sterile 96-well microtitre plates containing LB medium and ampicillin at 100 mg/L. After incubation of bacteria on a shaker for 4 h at room temperature, 5 μl of bacterial culture were transferred into PCR tubes and lysed by heating to 95°C for 5′. Bacterial lysates were used to PCR amplify cloned inserts in a 50 μl reaction in standard conditions with M13 and T7 primers, which flank the multiple cloning site of PCR 2.1 topo cloning vector, using the following conditions: 30 cycles each of 94°C for 40 s, 50°C for 40 s and 72°C for 60 s. After amplification, 10 microliters were loaded on 1.5% agarose gels and transferred onto Hybond N+ nylon (Amersham Life Sciences). The filters were hybridized under stringent conditions in Church and Gilbert buffer (Church and Gilbert, 1984) at 65°C with equivalent amounts of 32P-labeled single-stranded cDNA of approximately equal specific activity derived from driver and tester mRNA respectively. Filters were washed 2×20 h in 2×SSC, 0.1% SDS and 2×20 h in 0.2×SSC, 0.1% SDS at 65°C. Filters were first exposed to the Phosphorimager to quantify signal intensities, then were exposed to Biomax films (Kodak) for up to 4 days and the signal of clones were compared. The folds of induction were calculated with the Instant Imager software (Canberra Packard, Inc.). Each signal was quantified and the ratio between the signals obtained with the two hybridizations (Ras vs MEK/Rac or AdRas vs AdGFP) was calculated.

Preparation of cDNA probes

One microgram of poly-(A)+ RNA was converted into radiolabeled first-strand cDNA by a reverse transcriptase reaction in the presence of α-32P-dATP. All labeling reactions were obtained combining 1 μg poly-(A)+ RNA with 50 U of MMLV reverse transcriptase, 5 mM DTT, 10 mM dNTP mix (dATP 2.5 mM, dCTP 10 mM, dGTP 10 mM and dTTP 10 mM), 3.5 μCi of α-32P-dATP, with oligo-dT and random primers in the appropriate enzyme buffer. The reaction was incubated at 70°C for 2 min in a preheated PCR thermal cycler, then at 50°C for 20 min. The reaction was terminated with 100 mM EDTA. The 32P-labeled cDNA was purified from unincorporated 32P-labeled nucleotides and small cDNA fragments using a G-25 column (Boehringer Mannheim, Inc.).

Generation of recombinant adenoviral plasmids

To express H-Ras mutants and GFP as control in an adenoviral plasmid, we followed the protocol by He et al. (1998). Briefly, the H-Ras mutants were first cloned into the shuttle vector, pAdTrack-CMV in BglI site. The resultant plasmid was linearized by digestion with PmeI and subsequently cotrasformed into E. coli BJ5183 cells in the presence of the adenoviral backbone plasmid, pAdEasy-1. Recombinants were selected for kanamycin resistance and recombination confirmed by BamHI, SpeI and EcoRI digestion. The linearized recombinant plasmids were transfected into adenovirus packaging cell lines, HEK293 cells. Transfected cells were monitored for GFP expression and collected 7–10 days after transfection by scraping cells off plates and pelletting them along with any other floating cells in the culture. The harvested cells were then used for serial infection of HEK293 cells after three rapid freezing/thawing steps (He et al., 1998). The resultant viruses were purified by CsCl banding, quantified and used for FRTL-5 infection experiments.

Infection experiments

FRTL-5 cells were split 72 h before the infection at a density of 1.2×106 cells/100 mm2 plate. The infection was performed in F12 medium in the absence of growth factors and serum with 15 μl of purified virus. Cells were incubated in the presence of virus for 1 h and then complete medium was added. The infection was monitored by the GFP expression and after 6 h 90% of the cells were GFP-positive. RNA and proteins were collected at the time indicated to perform immunoblotting analysis and single-stranded cDNA probes (see Results).

Sequence and searching in the databases

PCR samples to be sequenced were transferred and collected in a 384-well plate, using a Biomek 2000 instrument (Beckman). Each PCR template (200 ng) was enzimatically purified using a mixture of Shrimp Alkaline Phosphatase (1 U) and Exonuclease I (5 U) Shrimp Alkaline Phosphatase, 37°C for 15 min, 80°C for 15 min. Sequence reactions were performed using Big Dye Terminator Cycle Sequencing Ready Reaction Kit from Perkin Elmer. Samples were loaded on an ABI 377 Automated Sequencer. Samples management, from PCR to the final sequence, was performed by the NEW-BioLIMS software developed at CRIBI by Dr Nicola Cannata. Sequences were clustered using a program based on the Tiger Assembler optimized by Dr Rosario Dioguardi. Clusters were then used for homology search against EMBL, dbEST (release 57), and Swissprot (release 37) databases (Sutton et al., 1995).

References

  1. Ambesi-Impiombato FS, Picone R, Tramontano D . 1982 Cold Spring Harbour Conferences on Cell Proliferation (9: Growth of Cells in Hormonally Defined Media) pp. 483–492

  2. Augenlicht LH . 1987 Cancer Res. 47: 6017–6021

  3. Bos JL . 1989 Cancer Res. 49: 4682–4689

  4. Cartwright P, Muller H, Wagener C, Holm K, Helin K . 1998 Oncogene 17: 611–623

  5. Cass LA, Meinkoth JL . 2000 Oncogene 19: 924–932

  6. Church GM, Gilbert W . 1984 Proc. Natl. Acad. Sci. USA 81: 1991–1995

  7. Ciccarelli FD, Acciarito A, Alberti S . 2000 Hum. Mol. Genet. 12: 1001–1007

  8. Cobellis G, Missero C, Di Lauro R . 1998 Oncogene 17: 2047–2057

  9. Cowley S, Paterson H, Kemp P, Marshall CJ . 1994 Cell 77: 841–852

  10. Damante G, Di Lauro R . 1994 BBA 1218: 255–266

  11. Datta SR, Brunet A, Greenberg ME . 1999 Genes Dev. 13: 2905–2927

  12. Diatchenko L, Chris Lau Y-F, Campbell AP, Chenchik A, Moqadam F, Huang B, Lukyanov S, Lukianov K, Gurskaya N, Sverdlov ED, Siebert PD . 1996 Proc. Natl. Acad. Sci. USA 93: 6025–6030

  13. Farid NR, Shi Y, Zou M . 1994 Endocr. Rev. 15: 202–232

  14. Francis-Lang H, Zannini MS, De Felice M, Berlingieri MT, Fusco A, Di Lauro R . 1992 Mol. Cell. Biol. 12: 5793–5800

  15. Fusco A, Berlingieri MT, Di Fiore PP, Portella G, Grieco M, Vecchio G . 1987 Mol. Cell. Biol. 7: 3365–3370

  16. Gaubatz S, Wood JG, Livingston DM . 1998 Proc. Natl. Acad. Sci. USA 95: 9190–9195

  17. Groundine M, Weintraub H . 1980 Proc. Acad. Natl. Sci. USA 77: 5351–5354

  18. He T-C, Zhou S, da Costa LT, Yu J, Kinzler KW, Vogelstein B . 1998 Proc. Natl. Acad. Sci. USA 95: 2509–2514

  19. Khosravi-Far R, Campbell S, Rossman KL, Der CJ . 1998 Adv. Cancer Res. 72: 57–107

  20. Langnaese K, Beesley PW, Gundelfinger ED . 1997 J. Biol. Chem. 272: 821–827

  21. Malumbres M, Pellicer A . 1998 Front. Biosci. 3: 887–912

  22. Mansour SJ, Matten WT, Hermann AS, Candia JM, Rong S, Fukasawa K, Vande Woude GF, Ahn NG . 1994 Science 265: 966–970

  23. Missero C, Cobellis G, De Felice M, Di Lauro R . 1998 Mol. Cell. Endocr. 140: 37–43

  24. Missero C, Pirro MT, Di Lauro R . 2000 Mol. Cell. Biol. 20: 2783–2793

  25. Nobes CD, Hall A . 1995 Cell 81: 53–62

  26. Qiu R-G, Chen J, Kirn D, McCormick F, Symons M . 1995 Nature 374: 457–459

  27. Ridley AJ, Hall A . 1992 Cell 70: 389–399

  28. Ridley AJ, Paterson HF, Johnston CL, Diekmann D, Hall A . 1992 Cell 70: 401–410

  29. Rodriguez-Viciana P, Warne PH, Dhand R, Vanhaesebroeck B, Gout I, Fry MJ, Waterfield MD, Downward J . 1994 Nature 370: 527–532

  30. Rodriguez-Viciana P, Warne P, Khwaja A, Marte BM, Pappin D, Das P, Waterfield MD, Ridley A, Downward J . 1997 Cell 89: 457–467

  31. Shields JM, Pruitt K, McFall A, Shaub A, Der CJ . 2000 Trends Cell Biol. 10: 147–154

  32. Sutton G, White O, Adams M, Kerlavage A . 1995 Genome Sci. Technol. 1: 9–19

  33. Trimarchi JM, Fairchild B, Verona R, Moberg K, Andon N, Lees JA . 1998 Proc. Natl. Acad. Sci. USA 95: 2850–2855

  34. Uwabe K, Gahara Y, Yamada H, Miyake T, Kitamura T . 1997 Neuroscience 80: 501–509

  35. White MA, Nicolette C, Minden A, Polverino A, Van Aelst L, Karin M, Wigler MH . 1995 Cell 80: 533–541

  36. Yoshimura T, Takeya M, Takahashi K . 1991 Biochem. Biophys. Res. Commun. 174: 504–509

  37. Zuber J, Tchernitsa OI, Hinzmann B, Schmitz A-C, Grips M, Hellriegel M, Sers C, Rosenthal A, Schafer R . 2000 Nature Genetics 24: 144–152

Download references

Acknowledgements

We are grateful to Dr Mario De Felice, Dr Jong-Tai Chun and Dr Gabriella De Vita for critical reading of the manuscript. We are indebted to Dr Elio Biffali, Raimondo Pannone and Michele Pischetola for technical assistance. This work was supported by grants from the Associazione Italiana per la Ricerca sul Cancro (AIRC), the Consiglio Nazionale delle Ricerche ‘Target Project on Biotechnology’, the Ministero per I'Universita’ e la Ricerca Scientifica Project ‘Meccanismi molecolari responsabili del differenziamento delle cellule tiroidee e loro utilizzazione in diagnostica e terapia', and the Programma Biotecnologie legge 95/95 (MURST 5%).

Author information

Correspondence to Roberto Di Lauro.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Cobellis, G., Missero, C., Simionati, B. et al. Immediate early genes induced by H-Ras in thyroid cells. Oncogene 20, 2281–2290 (2001). https://doi.org/10.1038/sj.onc.1204320

Download citation

Keywords

  • thyroid
  • tumorigenicity
  • differentiation
  • subtraction library
  • adenovirus

Further reading