Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Transcriptional networks of knockout cell lines identify functional specificities of H-Ras and N-Ras: significant involvement of N-Ras in biotic and defense responses

Abstract

We characterized differential gene expression profiles of fibroblast cell lines harboring single or double-homozygous null mutations in H-ras and N-ras. Whereas the expression level of the individual H-, N- and K-ras genes appeared unaffected by the presence or absence of the other ras loci, significant differences were observed between the expression profiles of cells missing N-ras and/or H-ras. Absence of N-ras produced much stronger effects than absence of H-ras over the profile of the cellular transcriptome. N-ras−/− and H-ras−/− fibroblasts displayed rather antagonistic expression profiles and the transcriptome of H-ras−/− cells was significantly closer to that of wild-type fibroblasts than to that of N-ras−/− cells. Classifying all differentially expressed genes into functional categories suggested specific roles for H-Ras and N-Ras. It was particularly striking in N-ras−/− cells the upregulation of a remarkable number of immunity-related genes, as well as of several loci involved in apoptosis. Reverse-phase protein array assays demonstrated in the same N-ras−/− cells the overexpression and nuclear migration of tyrosine phosphorylated signal transducer and activator of transcription 1 (Stat1) which was concomitant with transcriptional activation mediated by interferon-stimulated response elements. Significantly enhanced numbers of apoptotic cells were also detected in cultures of N-ras−/− cells. Our data support the notion that different Ras isoforms play functionally distinct cellular roles and indicate that N-Ras is significantly involved in immune modulation/host defense and apoptotic responses

Introduction

The mammalian H-ras, N-ras and K-ras genes code for closely related small guanine triphosphate (GTP)ases acting as critical components of signaling pathways that control cellular proliferation, survival or differentiation. The Ras proteins encoded by these genes (H-Ras, N-Ras, K-Ras4B/4A) work as molecular switches cycling between inactive (guanine diphosphate-bound) and active (GTP-bound) states, in a process modulated by a variety of regulatory proteins including GAPs (GTPase activiating proteins) and GEFs (guanine nucleotide exchange factors) (Reuther and Der, 2000; Colicelli, 2004). The three ras genes appear to be ubiquitously expressed in mammalians, although there are reported differences of expression levels for each gene depending on the tissue and/or developmental stage under consideration (Leon et al., 1987; Su et al., 2004).

It is still unclear whether the different Ras proteins play specific or overlapping functional roles in physiological and pathological processes. Their high degree of sequence homology, and the observation that all Ras isoforms share common sets of downstream effectors and upstream activators (Reuther and Der, 2000; Rojas and Santos, 2002; Colicelli, 2004), suggested initially that they were mostly redundant functionally. In contrast, the preferential activation of specific Ras isoforms in particular malignancies (Bos, 1989; Rojas and Santos, 2002), the different transforming potential of transfected ras genes depending on the recipient cell line (Maher et al., 1995; Oliva et al., 2004), or the distinct sensitivities exhibited by different Ras family members for inhibition by GAPs (Bollag and McCormick, 1991), activation by GEFs (Jones and Jackson, 1998; Clyde-Smith et al., 2000), or for interaction with various downstream effectors (Yan et al., 1998; Walsh and Bar-Sagi, 2001; Hancock, 2003; Liao et al., 2003; Plowman and Hancock, 2005), suggested otherwise. Reports indicating that different Ras isoforms follow different intracellular processing pathways and their final, mature products locate to different membrane microdomains or subcellular compartments also support the notion of distinct cellular roles for the different Ras isoforms (Hamilton and Wolfman, 1998; Roy et al., 1999; Voice et al., 1999; Hancock, 2003; Liao et al., 2003; Ehrhardt et al., 2004; Plowman and Hancock, 2005; Rocks et al., 2005; Matallanas et al., 2006).

Gene targeting offers another valid approach to probe for functional specificity of Ras isoforms. Although K-ras-deficient embryos die in utero (Johnson et al., 1997; Koera et al., 1997), H-ras, N-ras and K-ras4A knockout mice are viable and do not show any obvious abnormalities (Umanoff et al., 1995; Johnson et al., 1997; Esteban et al., 2001; Plowman et al., 2003). Simultaneous removal of H-ras and N-ras results also in viable mice (Esteban et al., 2001) indicating that, among the different ras genes, only K-ras is necessary and sufficient for development of mice to the adult stage and suggesting that K-Ras performs specific function(s) that cannot be carried out by either H-Ras or N-Ras. However, a recent study describing that the knockin of H-ras at the K-ras locus results in viable adult mice with cardiomyopathy, suggests that H-Ras may succesfully mimic K-Ras function during embriogenesis but not during adult life (Potenza et al., 2005).

Multiple strategies have been used in the past to check the functionality of Ras proteins in cells. The development of genomic and proteomic analysis tools opens the way to more exhaustive, genome-wide studies aimed at characterizing transcriptional networks associated to the function of specific ras genes or proteins. So far, most studies on Ras-related genomic profiling were concerned with characterization of cell lines transformed by various oncogenic Ras forms (Zuber et al., 2000; Brem et al., 2001; Croonquist et al., 2003; Ohnami et al., 2003; Vasseur et al., 2003; Sweet-Cordero et al., 2005).

In this study, we undertook detailed genomic and proteomic analyses of mouse embryo fibroblasts derived from single- and double-knockout mice for the H-ras and N-ras loci that were generated in our laboratory. We reasoned that comparing the transcriptomic profiles linked to deficiency of H-Ras and/or N-Ras in fibroblasts to those of their respective, wild-type, counterparts could provide significant clues on the functional specificity or redundancy of the H-ras and N-ras gene products. Our experimental data, generated from commercial oligonucleotide microarrays, reverse-phase protein microarrays and complementary functional assays documented the existence of distinct, specific transcriptional networks in the various Ras-deficient cells analysed, and supported the notion of functionally distinct cellular roles for the H-Ras and N-Ras protein isoforms.

Results

Expression analysis of ras knockout cell lines: level of expression of individual ras genes

To determine whether the different Ras family members are associated to specific gene expression programs, we compared the genomic expression profiles of immortalized fibroblasts derived from knockout mice harboring single- or double-null mutations in H-ras and N-ras with those of their corresponding, normal control counterparts. Pre-confluent cultures of at least two separate cell lines belonging to each of the ras-related genotype(s) under study (WT, H-ras−/−, N-ras−/− and H-ras−/−/N-ras−/−) were harvested and their RNA extracted for subsequent analysis using commercial high-density oligonucleotide microarrays. At least three independent microarray hybridizations were performed with RNA corresponding to each of the null-mutant ras genotypes under study. In all, this study encompassed analysis of a total of 15 separate chip microarray hybridizations (six for controls and three for each of the three null-mutant genotypes). Global analysis was possible by using robust microarray analysis (RMA) (Irizarry et al., 2003b) as a tool allowing simultaneous background correction, multichip normalization and quantitation of probe set expression level signals in all separate microarray hybridizations performed.

Using this methodology, we first compared the levels of expression of the individual ras genes in all control and knockout cell lines analysed. Figure 1 presents the normalized, absolute expression values of the H-ras, N-ras and K-ras probe sets in different chip hybridizations of RNA from knockout and wild-type cell lines. We observed that the level of expression of each H-ras, N-ras and K-ras gene was not affected significantly by the presence or absence of the other ras loci, suggesting that single- or double-deficiency of H-ras and N-ras does not result in significant compensatory changes in the expression levels of the other Ras isoforms. Thus, the observation that the expression level of K-ras did not change significantly among wild-type and knockout- N-ras and H-ras-mutant samples (Figure 1) confirms previous reports indicating that overexpression of K-ras is not needed to compensate the absence of H-ras and N-ras (Esteban et al., 2001). We also observed that the expression level of the H-ras probe set presented a relatively small decrease when comparing H-ras knockout samples to wild-type and N-ras knockout samples (Figure 1). This behavior is consistent with the low level of basal expression reported for this gene (http://symatlas.gnf.org/SymAtlas/; Leon et al., 1987; Su et al., 2004) and with a low difference between the absolute hybridization signals generated by the perfect match and mismatch oligonucleotide probe set for H-ras that are present in this particular microarray. In contrast, the two probe sets corresponding to the N-ras locus showed significant decreases in expression value when comparing the corresponding control and knockout samples (Figure 1). The absence of ras RNA and protein was routinely checked in all knockout cell lines analysed in this study (not shown).

Figure 1
figure1

Expression values of probe sets for the H-ras, N-ras and K-ras genes in microarray hybridizations with RNA from wild-type, H-ras−/− and N-ras−/− cell lines. Graphical representation of the absolute expression values calculated for the probe sets corresponding to H-ras (red), N-ras (blue) and K-ras (green) in the murine MGU74Av2 GeneChip microarrays used in this study. ID numbers of the different ras probe sets present in the chips were as follows: H-ras, 160536_at; N-ras, 94362_at and 160925_at; K-ras, 97991_at. The points in the plot represent normalized absolute expression values computed from a collection of independent microarray hybridizations performed with RNA from wild-type (four microarrays), H-ras−/− (three microarrays); N-ras −/− (three microarrays) immortalized fibroblastic cell lines.

Global expression analysis: differential gene expression in ras knockout cell lines

Statistically significant gene expression changes occurring in single and double H-ras and N-ras knockout cell lines were identified using the significance analysis of microarrays (SAM) algorithm (Tusher et al., 2001). Setting stringent false discovery rate (FDR) values for significance, the plots in Figure 2 allow identification of the differentially expressed gene probe sets in the H-ras−/− (Figure 2a), N-ras−/− (Figure 2b) and H-ras−/−/N-ras−/− (Figure 2c) cell lines, based on their distance to the diagonal encompassing the probe sets whose expression level is unchanged between knockout and wild-type samples.

Figure 2
figure2

Statistical identification of differentially expressed genes in ras knockout cell lines. Graphical display of statistical analysis performed to identify genes undergoing significant changes of expression in ras knockout cell lines as compared to normal, control cell lines. Individual plots compare the expression profile of wild-type, normal mouse fibroblasts to that of knockout, H-ras−/− (a), N-ras−/− (b) and H-ras−/−/N-ras−/− (c) cells. They were generated by SAM-contrasting three independent microarray hybridizations, performed with RNA of cell lines belonging to each of the ras knockout genotypes, and the results of six separate hybridizations with RNA from corresponding wild-type fibroblasts. Differential expression for a given gene probe set is quantitated by Δ(i), measuring the distance of the spot representing its expression value to the no-change diagonal. Green dots identify probe sets presenting significant alterations of expression. Black dots remaining close to the diagonal represent probe sets whose expression level does not show significant change in ras-mutants relative to their controls.

Analysis of the contrast graphs documented that, for similar statistical significance rates, the N-ras−/− knockout cells presented the highest number of genes showing altered expression levels relative to wild-type control fibroblasts. Thus, 12 genes showed consistent changes of expression level in H-ras-deficient cell lines (five upregulated) (Figure 2a; Table 1), whereas altered expression of 96 genes (82 upregulated) was consistently detected in the N-ras-deficient fibroblasts (Figure 2b; Table 2). The H-ras−/−/N-ras−/− double-knockout cell lines allowed consistent identification of altered expression levels in 15 genes (six upregulated) (Figure 2c; Table 3). All in all, the three contrasts in Figure 2 identified a joint total of 114 differentially expressed genes (detected by 123 distinct probe sets in the microarrays).

Table 1 Differentially expressed genes identified in H-ras knockout cells
Table 2 Differentially expressed genes identified in N-ras knockout cells
Table 3 Differentially expressed genes identified in H-ras−/−/N-ras−/− knockout cells

Hierarchical clustering of the differentially expressed genes

Figure 3 depicts a dendrogram generated by unsupervised hierarchical clustering of microarray hybridization data sets corresponding to the 123 distinct probe sets showing statistically significant differential expression levels in our ras knockout cells.

Figure 3
figure3

Hierarchical clustering of genes showing differential expression in ras knockout cell lines. Dendrogram generated by cluster analysis of absolute expression values of gene probe sets in H-ras−/−, N-ras−/−, H-ras−/−/N-ras−/− and wild-Type control fibroblasts. Genes shown are those whose transcript levels showed statistically significant variations in Figure 2. Horizontal rows represent single-gene probe sets. Vertical columns represent results from single-microarray hybridizations. Lettering on top of columns identifies independent biological replicas analysed for each of the four relevant ras genotypes studied. Each box represents the hybridization signal value of a gene probe set in the corresponding microarray assay. The intensity of color saturation in each probe set box (ranging from 2 to 14 in a log 2 scale) provides a quantitative estimation of its expression level. Red color denotes overexpression, increasing in brightness with higher values. Green color denotes repression, increasing in brightness with lower values. Black color denotes unchanged expression signals relative to controls.

Two main vertical branches were identified. One encompassed the hybridizations corresponding to H-ras−/− cells and wild-type controls, whereas the second branch contained the remainder sets of hybridization data, corresponding to the N-ras−/− and H-ras−/−/N-ras−/− knockout cells (Figure 3, columns). Such column distribution indicates that the transcriptomic pattern of H-ras−/− cells is closest to that of wild-type fibroblasts, whereas the expression profile of N-ras−/− cells is significantly closer to that of double-knockout H-ras−/−/N-ras−/− cells than to that of H-ras−/− cells.

Focusing on comparing the transcriptional patterns specific to the different ras genotypes, we observed rather antagonistic expression profiles between N-ras−/− and H-ras−/− fibroblasts (Figure 3, see clusters 1–8). For example, clusters 1 and 7 defined gene groups whose expression was, respectively, increased or decreased in H-ras−/− and wild-type fibroblasts in comparison with N-ras−/− and H-ras−/−/N-ras−/− cells. These observations are clearly consistent with the notion of functional specificity for H-Ras and N-Ras in fibroblasts.

Functional GO (gene ontology) analysis of the clusters defined among the horizontal expression profiles of the dendrogram also uncovered statistically significant associations linking specific cellular functions to some of the individual ras genotypes under study. For example, most genes belonging to cluster 6 in Figure 3 were assigned, with high degree of statistical probability (P-value=0.00000865), to functional categories concerned with immune and defense responses or response to biotic stimuli, particularly response to interferon (IFN) (GO:0006955, GO:0006952, GO:0009607). As the genes in that cluster are also specifically overexpressed in cells lacking N-ras, they provide a clear functional signature associated to the deficiency of N-Ras in those cells (see also Table 2). Similarly, cluster 1 (Figure 3), containing the two N-ras probe sets, presented statistically significant accumulation (P-values: 0.0062–0.0025) of genes identified by GO annotations such as cell growth (GO: 0016049), cell cycle (GO: 0007049) or phosphorus metabolism (GO: 0006793). Functional annotation of other genes included in Figure 3 also suggested the significant accumulation of genes concerned with cation transport (GO: 0006812) or energy and mitochondrial electron transport (GO: 0006118) in cluster 5; with apoptosis (GO: 0006915, 0042981) and cell death (GO: 0043067, 0043068) in cluster 7; and with organogenesis (GO: 0009887) and cyclic adenine 3′,5′monophosphate (cAMP) signaling (GO: 0007186) in cluster 8.

Further analysis of functional annotations assigned to specific gene clusters of the dendrogram reinforced the notion of non-overlapping functional roles for H-Ras and N-Ras regarding their effect on the transcriptome. Thus, the groups of genes clustering next to the N-ras and H-ras loci in the dendrogram (clusters 1 and 4, respectively) were very distinct from each other from the functional point of view. We also noticed that separate clusters 2 and 4 contained, respectively, genes that were exclusively upregulated or exclusively downregulated in cells lacking H-Ras (Figure 3). It is also worth mentioning that the two different N-ras probe sets of the microarrays shared expression profile with, and clustered next to, Cdkn2a (Figure 3, cluster 1), raising the possibility of shared regulatory mechanisms for the expression of both loci. In addition, many other genes also undergoing strong reduction of expression in N-ras−/− fibroblasts were also localized to proximal sub-branches in the dendrogram (Figure 3, cluster 1). Finally, most genes that showed simultaneous downregulation in the N-ras−/− and in double-knockout H-ras−/−/N-ras−/− cells were also localized to this cluster. On the other hand, the genes that were exclusively or preferentially downregulated in H-ras−/−/N-ras−/− cells defined separate, although proximal branches of the dendrogram (cluster 3). In contrast, the genes upregulated in the H-ras−/−/N-ras−/− cells were more widely scattered throughout the dendrogram, although a large fraction of them concentrated to its lower region (Figure 3, cluster 7).

Gene expression signatures linked to deficiency of H-ras and/or N-ras

To facilitate the detailed analysis of our microarray expression data, the differentially expressed genes detected in H-ras−/−, N-ras−/− and H-ras−/−/N-ras−/− cells were tabulated according to their degree of overexpression/repression and functional category (Tables 1, 2 and 3).

Analysis of the H-ras−/− fibroblasts identified 13 differentially expressed probe sets (12 distinct genes) showing statistically significant changes of expression. Five of those genes displayed increased expression and seven were significantly repressed (Table 1). Most of these genes were concerned with basic cellular functions such as growth or proliferation. Thus, loci upregulated in H-ras−/− cells code for proteins participating in biological processes such as G protein-mediated signal transduction (Gnb1), cell proliferation and spermatogenesis (Gfer), organ development (Nnat, Hoxc8) or protection from senescence and oxidative stress (Prdx2). The loci downregulated in the same knockout cells include genes coding for proteins involved in glutamate transport (Slc1a7), protein biosynthesis (Rps11), potential phospholipid phosphatase activity (Mtmr13) or still unknown functions (2610204K14Rik, 1110035L05Rik).

The N-ras−/− immortalized fibroblasts displayed significantly higher numbers of differentially expressed genes than cells belonging to the other ras genotypes. One hundred and four different probe sets (corresponding to 96 genes) showed statistically significant alterations of expression level in comparison with controls (Table 2). A majority (82) of them were upregulated, suggesting that N-ras may have a preferential role as negative modulator of gene expression.

The most striking observation in N-ras−/− cells was the exceedingly high number of upregulated genes known to be involved in immune defense responses (Table 2; Figure 3 clusters 6–8). A major part of them corresponds to a block of genes involved in responses to IFN (Table 2) that were mostly localized to cluster 6 of Figure 3. Another important group of these upregulated genes is concerned with other immunity/defense-related functions (Table 2) and its members (H2-Q7, H2-D1, Cx3cl1, Trim11, Tmem43, Anx8, Ephx1, Ilr3ra1, Lcn2, Ilrn1, Cfh, etc) are also proximally clustered in the dendrogram (clusters 6–8, Figure 3). Interestingly, none of the genes downregulated in N-ras−/− cells appeared to be functionally related to immunological processes (Figure 3, Table 2). These observations are consistent with a possible role of N-Ras as a negative transcriptional modulator of immunological response genes.

The products of other groups of differentially expressed genes in N-ras−/− cells appear to be functionally concerned with the regulation of cell proliferation and survival (Table 2). The simultaneous upregulation of loci such as Ccng1, Ak1, Ppp1r7, Tubb2 and downregulation of Cdkn2a, which are well-known modulators of the mammalian cell cycle (Table 2) suggest that the deficiency of N-ras may result in altered responses to proliferative, tumorigenic or senescence signals.

We also observed in N-ras−/− cells the occurrence of changes of expression level of genes involved in apoptosis such as Phlda3, Perp, Lsp1 and Bax (Table 2). The overexpression of signal transducer and activator of transcription (Stat1) (Townsend et al., 2004), together with the above-mentioned gene expression changes of cell cycle modulators and, in particular, the overexpression of Perp, a direct target of p53/p63 (Ihrie et al., 2005), and Bax, a direct regulator of Blc2 (Ruiz-Vela et al., 2005), suggest the possibility of enhanced apoptotic responses in the N-ras-deficient cells.

Another collection of differentially expressed genes in N-ras−/− cells code for proteins known to affect various overlapping signaling pathways involved in processes such as cAMP-dependent signaling, cell–cell adhesion or development (Table 2). The simultaneous upregulation of Adcy2 (directly involved in the cAMP biosynthesis) and Lcn2 (reported to enhance the cAMP-dependent protein kinase activity in vitro), together with the downregulation of Pkia (involved in inhibition of the cAMP-dependent protein kinase activity) may result in enhanced cellular responses in cAMP-dependent signaling, which would be consistent with the reduced cAMP-dependent β-adrenergic responses reported previously in cells overexpressing N-Ras (Davies et al., 1989; Dumaz and Marais, 2005; Takahashi et al., 2006).

Finally, other groups of differentially expressed genes detected in N-ras−/− cells are known to participate in various developmental processes (Table 2). Thus, upregulated genes, such as Dlx5, Mest, Mglap, Mmp13, Crip2, Gja1, Anxa4 are known to be involved in various developmental processes affecting organs of neural, skeletal, vascular and other origins. Two downregulated genes (Ncam1 and Fbln2c) known to affect development are also well-known cell-adhesion molecules. Table 2 includes additional groups of genes that were classified according to the cellular roles suggested by their similar functional annotations.

Analysis of the H-ras−/−/N-ras−/− cell lines allowed identification of 16 differentially expressed probe sets (15 different genes) (Figure 3, Table 3). Six of them showed increased expression levels and nine were repressed, including the N-ras gene itself. Notice that H-ras (absent in the double-knockout cells) is not listed in Table 3 because of the highly stringent parameters of significance used to generate our microarray hybridization data. This observation underscores the biological significance of the transcriptional data reported here (Figure 3, Tables 1, 2 and 3), and indicates that these data represent an underestimation of the actual number of differentially expressed genes occurring in the ras knockout cells.

Whereas the lists of genes repressed in both H-ras−/− and H-ras−/−/N-ras−/− knockout cells were not coincident, a large fraction of the differentially expressed genes detected in H-ras−/−/N-ras−/− cells displayed also altered expression in the N-ras−/− knockout cells (Tables 1, 2 and 3). Thus, Anxa8, Ccng1, Cfh and Crabp2 (overexpressed in H-ras−/−/N-ras−/− cells) presented also increased expression in N-ras−/− cells, whereas Crap and Rbpms were repressed in both N-ras−/− and H-ras−/−/N-ras−/− knockout cells (Tables 2 and 3). Such behavior suggests an exclusive or preferential role of N-Ras in regulation of the expression of those loci.

Lgtn was the only locus upregulated in H-ras−/−/N-ras−/− cells that was not detected as differentially expressed in any of the single ras knockouts. In contrast, most downregulated genes identified in H-ras−/−/N-ras−/− cells appeared to be exclusive to such double-knockout cells (Table 3), suggesting cooperative or overlapping roles of H-Ras and N-Ras in control of their expression, as it was necessary the absence of both gene products to detect their impaired expression. The list of loci exclusively downregulated in double-knockout cell lines includes genes coding for a zinc-finger protein (Cxxc5), as well as other proteins involved in development (Fhl1, Npn2, Irx3) or signaling (Calcrl, Grinl1a) (Table 3). No single gene was identified that showed altered expression in all three different ras knockout genotypes studied here.

Functional verification of microarray-based expression data

We wished to verify the microarray-based expression data by means of alternative experimental approaches. Thus, we first carried out quantitative, real-time polymerase chain reaction (PCR) of a randomly selected collection of 48 of the 123 differentially expressed genes previously identified. The reverse transcription (RT)-PCR signals of those 48 loci (relative to the signal of gliceraldehyde-3-phosphate dehydrogenase (GAPDH), used as internal control) are included as appropriate in Tables 1, 2 and 3. In general, qualitative agreement between the microarray and quantitative PCR results was quite good, although some differences were observed in the quantitative extent of the gene expression alterations measured by either technology (Tables 1, 2 and 3). Furthermore, immunoblotting of cellular extracts from the same knockout cell lines analysed above, using a variety of commercially available antibodies, confirmed the overexpression or repression of a series of protein products encoded by the differentially expressed genes identified above.

Direct experimental support for our genomic expression data was also obtained by means of antibody staining of protein array layouts (Figure 4) prepared with lysate extracts of the same ras knockout cell lines previously analysed at the genomic level. As our transcriptomic findings pointed to differential expression of significant number of immunity-related genes in the N-ras knockout cells (Table 2), we initiated these studies by testing lysates of N-ras−/− cells with antibodies directed against some protein forms known to participate in immune-related processes. Figure 4 quantifies the amounts of phosphorylated and unphosphorylated forms of members of the Stat protein family in ras knockout cells. Consistent with the genomic expression results described above, these data fully confirmed the presence of elevated levels of the protein Stat1 in N-ras knockout cells (Figure 4a and b). Furthermore, it was particularly interesting to detect a parallel increase, in N-ras knockout cells, of tyrosine phosphorylated (Y701) Stat1, whereas the levels of the serine phosphorylated (S727) form of this protein remained unchanged when comparing N-ras knockout cells to their normal, wild-type counterparts (Figure 4b).

Figure 4
figure4

Reverse-phase protein array layout and antibody staining. (a) Example of protein array layout used for Stat1 detection, indicating cell sample distribution. Protein lysate samples were printed, as indicated in Materials and methods, onto two groups of columns corresponding, respectively, to control, wild-type (left side) and knockout samples (right side). In all cases, the samples corresponding to each ras genotype were printed in duplicate, using five-point, twofold dilution curves (starting at 2 μg/μl). The sixth point was always a negative control consisting of lysis buffer alone. Similar slides were prepared and used for staining with various other antibodies recognizing the phosphorylated form of Stat1 as well as other Stat family proteins. (b) After antibody staining and development, the slides were scanned and the ratios of the normalized signals of the knockout samples relative to their respective controls were depicted in graphic form. Similar results were obtained in two separate experiments.

The remarkable elevation of cellular Stat1 protein levels caused by deficiency of N-Ras was further confirmed by Western immunoblotting and subcellular fractionation studies showing a clear migration of tyrosine phosphorylated Stat1 from the cytosol to the nucleus of N-ras knockout cells (Figure 5a). As nuclear migration of tyrosine-phosphorylated Stat1 is an important step of the participation of Stat1 in signaling processes mediating response to IFN and involving transcriptional regulation of specific gene targets (Kisseleva et al., 2002; Platanias, 2005), these observations strongly suggest that the changes of gene expression observed in the N-ras KO cells indeed reflect a functionally relevant role of N-Ras in modulation of immunological processes in vivo. Indeed, strong experimental support for such a notion was provided by our luciferase assays detection of significant transcriptional activation of ISRE (IFN-stimulated response element), but not GAS (γ-IFN activation sequence), sites in the N-ras−/− cells in comparison to their wild-type counterparts (Figure 5b).

Figure 5
figure5

Overexpression of Stat1 in N-ras−/− cells. (a) Quantitation and subcellular distribution of Stat1 protein forms. Western immunoblots showing the distribution of Stat1, pStat1 (Y701), Mek1 and Cyclin A in subcellular fractions of the N-ras+/+ and N-ras−/− fibroblasts. Mek1 and Cyclin A were used as controls for their corresponding cytosolic and nuclear localizations. (b) Transcriptional activation mediated by IFN-responsive elements GAS and ISRE. Relative luciferase activity of reporter GAS and ISRE constructs versus their empty vector controls was measured in transfected N-ras+/+ and N-ras−/− cells as described in Materials and methods. The assays were carried out in triplicate, with error bars indicating s.d. (P<0.05).

Finally, the possibility of enhanced apoptotic responses suggested by the transcriptional alterations detected in N-ras−/− cells (Figure 3, Table 3) was experimentally confirmed by means of flow cytometric analysis showing the presence of almost ten-fold higher numbers of apoptotic cells in growing cultures of N-ras−/− cells, as compared with their wild-type controls (Figure 6a and b).

Figure 6
figure6

Enhanced apoptosis in N-ras−/− cultures. (a) Immunoblot assays for Bax and actin protein in lysates of N-ras+/+ and N-ras−/− fibroblasts. (b) Flow cytometric analysis of preconfluent cultures of N-ras−/− and wild-type control (WT) cells. Percentage of apoptotic cells in knockout and WT cultures in this figure was 12.81 and 1.64%, respectively. Results are representative of two separate determinations.

Discussion

This study was aimed at ascertaining whether the different members of the mammalian Ras protein family control specific or overlapping transcriptional networks. Our experimental approach involved using Affymetrix, Ltd, High Wycombe, UK oligonucletide microarrays to characterize the transcriptomic patterns of a variety of single- or double-knockout cell lines generated in our laboratory that carried homozygous null mutations for H-ras or N-ras. We expected that subsequent analysis and comparison of the gene sets/networks identified in association with the deficiency of specific ras isoforms would potentially help in answerig questions about the functional specificity and/or redundancy of the different Ras isoforms.

The validity of our conclusions regarding transcriptional networks controlled by either H-Ras or N-Ras, is strongly supported by (i) the complementary nature of the collection of different ras genotypes analysed (H-ras−/−; N-ras−/−; H-ras−/−/N-ras−/−; and their corresponding wild-type controls) and (ii) the fact that all cell lines studied here were generated and grown in parallel under similar conditions. The simultaneous consideration and study of the transcriptomic patterns identified for all those different genotypes ensured that the list of differentially expressed genes presented in this study includes only loci whose altered expression level is specifically linked to the deficiency of H-Ras and/or N-Ras, and not to any other variable of the cell lines under study, such as genetic background, culture conditions, immortalization, etc.

On the other hand, the reproducibility and biological significance of our genomic expression data are also firmly based on (i) the fact that different cell lines of the same genotype yielded similar results and on (ii) the highly stringent, restrictive cutoff parameters of significance applied for processing and selection of the transcriptional data reported. The advantages of the computational methods used here for handling and analysis of the RNA microarray hybridization data are described in detail in Materials and methods. In this regard, it is pertinent to mention that the use of those algorithms allowed to minimize background signal noise and to maximize statistical significance of the differential gene expression data obtained while using all the RNA microarray hybridization data without any byass or arbitrary pre-filtering. Furthermore, in contrast to pairwise comparison software packages (i.e., MAS5), these methodologies allowed for simultaneous normalization and quantitation of the signals of multiple microarrays, thus eliminating inter-chip variability and establishing a common background signal level for all samples analysed in separate hybridization experiments.

The overall comparison of the different transcriptional patterns reported in this study is strongly supportive of the notion that the different Ras family members play significantly different functional roles in eukaryotic cell. We observed initially that the expression of each individual ras gene family member appears to be independently regulated, as we documented that the absence of N-Ras and/or H-Ras did not cause any obvious compensatory changes of the level of expression of the other Ras isoforms in the knockout cells. On the other hand, the qualitative and quantitative analysis of the transcriptional patterns identified in N-ras and H-ras-deficient cells clearly indicated that N-ras exerts a significantly higher influence than H-ras on the profile of the cell transcriptome. Thus, the expression profile of H-ras−/− cells was much closer to that of wild-type fibroblasts than to that of N-ras−/− cells, whereas the transcriptomic pattern of double H-ras−/−/N-ras−/− knockout cells was closer to that of N-ras−/− than to H-ras−/− cells. In addition, the N-ras−/− cells presented a much higher (eightfold higher) number of differentially expressed genes than the H-ras knockout cells. Finally, the rather antagonistic transcriptional profiles exhibited by H-ras−/− and N-ras−/− cells may account, at least in part, for the relatively small number of differentially expressed genes identified in double-knockout H-ras−/−/N-ras−/− cells.

The comparison of the lists of differentially expressed genes identified in the various knockout cells analysed confirmed the presence of specific transcriptional signatures associated to the absence of H-ras or N-ras, suggesting that H-Ras and N-Ras control different transcriptional networks. It also documented the occurrence of rather antagonistic transcriptional profiles between the N-ras and the H-ras knockout fibroblasts, further supporting the notion that H-ras and N-ras play distinct, independent functional roles in eukaryotic cells.

The study of the transcriptional patterns uncovered in this work centered our attention on the significant role of N-Ras controlling important parts of the cellular transcriptomic profile. The observation that a majority of the differentially expressed genes identified in N-ras−/− cells were upregulated suggests that N-Ras may have a preferential role as a negative regulator of gene expression, although an alternative explanation could be that the absence of N-Ras needs to be compensated by the overexpression of all those genes.

It was particularly striking the observation that the expression of a great number of genes involved in the immune response, particularly response to IFN, was amplified in cells lacking N-Ras. These observations suggest a possible contribution of this protein as a transcriptional modulator of the immune response. It was also remarkable in N-ras−/− cells the observation of several transcriptional alterations suggesting that N-Ras plays a significant role in apoptotic processes. In particular, the overexpression of Stat1 (Townsend et al., 2004), Perp (a direct target of p53/p63) (Ihrie et al., 2005) and Bax (a direct regulator of Blc2) (Ruiz-Vela et al., 2005) suggest that the N-ras-deficient cells possess enhanced ability to display apoptotic responses. Finally, the observed transcriptional alterations of a number of genes involved in cell cycle control or cAMP responses (Ccng1, Cdkn2A, Adcy2, Pkia, etc) suggest possible roles of N-Ras as modulator of various cellular responses involved in proliferative, tumorigenic or senescence signals. In this regard, our observations at the transcriptional level are clearly consistent with reports indicating that loss of N-Ras promotes tumor progression or that its overexpression results in reduced cAMP-mediated responses (Davies et al., 1989; Dumaz and Marais, 2005; Takahashi et al., 2006).

Although our array-based RNA expression data may provide significant functional clues, it is apparent that alternative experimental approaches and parallel protein expression analysis of the same cellular samples may be needed to confirm the data and start ascertaining the functional biological implications of the observed transcriptomic changes. In this regard, although an extensive proteomic analysis of the same cellular samples analysed here was out of the scope of this article, basically all RT–PCR and immunoblotting studies performed in parallel to the microarray assays yielded confirmatory evidence for the transcriptional changes reported, both at level of RNA and protein.

It is also worth mentioning that at present it is not yet possible to determine whether the differential gene expression observed in knockout cells was caused exclusively by the absence of direct regulatory mechanisms mediated by the missing Ras isoforms, or by other compensatory mechanisms. Further, extensive analysis at the protein level and specific functional studies will be needed to ascertain the exact mechanisms involved in, and the biological relevance of, transcriptional changes potentially affecting a variety of cellular processes such as proliferation, apoptosis or defense mechanisms in the ras knockout cell lines analysed here. Such confirmatory studies were limited in this paper to functional and mechanistic analyses confirming the involvement of N-Ras in host defense and apoptotic responses.

Thus, consistent with the transcriptional data, immunological characterization of protein array layouts prepared from cell line extracts confirmed the overexpression of Stat1 protein in N-ras−/− cells. It is well established that latent Stat1 resides primarily in the cytoplasm, where it responds to different stimuli through tyrosine phosphorylation. After phosphorylation, Stat1 may dimerize and migrate to the nucleus, where it binds to specific DNA target sites (ISRE or GAS) in the promoters of regulated genes (Kisseleva et al., 2002; Platanias, 2005). Consistent with this model, our cell fractionation analysis demonstrated a significant migration and accumulation of tyrosine phosphorylated Stat1 to the nucleus of the N-ras−/− knockout cells, which coincided with a significant transcriptional activation of ISRE elements, known to be important mediators of a number of immunity-related cellular responses. These observations clearly confirm the involvement of N-Ras in immune and biotic cellular responses and are also consistent with recent reports of altered immunological responses and T-cell function in N-ras-deficient mice (de Castro et al., 2003; Perez de Castro et al., 2004). Further in vivo studies will be needed to fully characterize the functional relation between N-ras and the immune defense system in the knockout mice. In this regard, it is worth mentioning that the transcriptional data reported here are fully derived from analysis of fibroblasts rather than lymphocytes, and that the available literature indicates that IFN/STAT signaling is mostly independent of T-cell receptor-dependent gene expression or signal-transduction processes (Beadling et al., 1994; Petricoin et al., 1997).

Likewise, the flow cytometric characterization of the same cell lines previously analysed with microarrays demonstrated the presence of significantly higher numbers of apotoptic cells in N-ras−/− cultures than in N-ras+/+ cultures growing under similar conditions. This observation confirms a functional link between absence of N-Ras and enhanced apoptotic abilities and is consistent with reports from other laboratories suggesting that N-Ras promotes cell survival and avoidance of apoptosis (Wolfman et al., 2002; Eskandarpour et al., 2005).

Further cellular and molecular studies will be needed to obtain confirmatory evidence and additional information on the basic mechanisms involved in participation of H-Ras or N-Ras in a variety of other functional processes suggested by the list of specific transcriptional changes observed in the collection of normal and knockout ras cell lines studied in this report.

Materials and methods

Immortalized mouse embryonic fibroblast lines and cell culture procedures

Mouse embryos of the appropriate ras genotypes were recovered at DPC 12–14, mechanically minced and treated with trypsin-ethylenediaminetetraacetic acid (EDTA) 0.25% (Gibco-BRL, Cheshire, UK) for 30 min before plating on Dulbecco's modified Eagle's medium (DMEM; Gibco) supplemented with fetal bovine serum (10% FBS; Hyclone, Logan, Utah, USA), glutamine (2 mM), penicillin (100 U/ml) and streptomycin (100 μg/ml). Cultures were grown following a 3T3 protocol in a humidified CO2 (5%) atmosphere at 37°C. Immortalized cultures that survived crisis after 15–20 passages were identified and cloned and their genotypes reconfirmed by PCR analysis. Expression of Ras protein isoforms was monitored by immunoblotting with specific antibodies directed against the C-terminus of H-Ras (sc-520) or N-Ras (sc-519) (Santa Cruz Biotechnologies, Santa Cruz, CA, USA). At least two sets of independently generated, immortalized fibroblast cell lines (originated from different embryos) were used for analysis of each of the genotypes under study.

Subcellular fractionation of N-ras+/+ and N-ras−/− cells growing up to 70–80% confluence in 10 cm dishes was carried out using the Nuclear/Cytosol Fractionation Kit (MBL International Corporation, Woburn, MA, USA) following the manufacturer's instructions. Flow cytometry of these cultures was performed on permeabilized, propidium iodide-stained cells using a fluorescence-activated cell sorter Calibur (Becton & Dickinson, Franklin Lakes, NJ, USA) machine and the WinMDI program for results analysis.

RNA isolation, cDNA synthesis and microarray hybridization

RNA was purified from five 10 cm culture dishes per cell line using a commercial kit (RNeasy, Qiagen, Hilden, Germany). Its concentration was measured at 260 nm (Ultrospec 2000, Pharmacia Biotech) and its purity and quality was determined using RNA 6000 Nanochips (Agilent Technologies, Germany). RNA was then used to synthesize complementary RNA (cRNA) probes for hybridization to Affymetrix MGU74Av2 GeneChip high-density oligonucleotide microarrays (including 12488 oligonucleotide probe sets corresponding to about 10 000 mouse genes) according to protocols described in the Gene Expression Analysis Technical Manual (http://www.affymetrix.com). In brief, total RNA was reverse transcribed into double-stranded complementary DNA (cDNA) using an oligo(dT)24 primer containing a T7 polymerase promoter-binding site (Genset, LaJolla, CA, USA). cDNA was then used as a template to synthesize cRNA by in vitro transcription (Ambion T7 Megascript, Austin, TX, USA), with incorporation of biotinylated nucleotides (Enzo Diagnostics, Farmingdale, NY, USA). Labeled cRNAs were fragmented and hybridized to Affymetrix GeneChip MGU74Av2 using the Genechip Fluidics Station 450 (Affymetrix). Hybridized arrays were stained with streptavidin–phycoerythrin, rewashed, treated with biotinylated antistreptavidin–phycoerythrin antibodies and restained with streptavidin–phycoerythrin, according to the manufacturer's protocols. The stained arrays were finally scanned in a GeneArray Scanner (Hewlett Packard, Palo Alto, CA, USA).

Microarray hybridization data analysis: normalization, differential gene expression and clustering

The RMA algorithm (Irizarry et al., 2003b) was used for background correction and normalization of fluorescent hybridization signals of the microarrays, both at internal (intra-microarrays) and comparative (inter-microarrays) levels. This algorithm was selected over others available (MAS5, (Affymetrix 2001); MBEI, a model-based algorithm) (Li and Wong, 2001) because it was deemed to provide the best precision in signal detection to achieve adequate multiple-chip normalization (Bolstad et al., 2003), especially in cases of low-level gene expression (Irizarry et al., 2003a, 2003b; Barash et al., 2004) by producing efficient quantile normalization of the distribution of probe intensities from each array in the context of a complete set of arrays.

We used Bioconductor and R as computational tools (www.bioconductor.org), to apply RMA to the data set of 15 microarray hybridizations including three or more different biological replicas corresponding to each of the different ras genotypes under study (WT, H-ras−/−, N-ras−/− and H-ras−/−/N-ras−/−).

After quantitation of expression level of each probe set in all microarrays analysed, the SAM algorithm (Tusher et al., 2001) was used to identify probe sets displaying significant differential expression when comparing the knockout samples to their respective controls. This algorithm performs statistical discrimination analysis using permutations to check the stability of variables fulfilling the ‘alternative hypothesis’. The method calculates the type I error, or number of expected false positives, using the calculation of the FDR parameter (Benjamini et al., 2001). In this report, changes of probe set expression level were identified as significant using a FDR cutoff value of 0.04 for comparisons involving the H-ras and N-ras knockout samples and FDR=0.06 for comparisons of the double-knockout H-ras−/−/N-ras−/− to their controls.

Following identification of the differentially expressed probe sets, the corresponding matrix of expression values for all microarray hybridizations performed were analysed using the hclust clustering algorithm implemented in R (Murtagh, 1985). This algorithm performs hierarchical cluster analysis with complete linkage to find similarity between probe sets based on their expression values in the diferent chip microarrays analysed. The algorithm classifies the probe sets in correlated groups presenting similar expression profiles or expression signatures. The statistical significance of functional GO annotations was estimated by means of P-values of confidence calculated by running Fisher's exact tests to contrast the number of genes assigned to the various functional categories within each cluster of the dendrogram.

Real-time PCR

Real-time PCR was carried out using a commercial kit (Qiagen, Hilden, Germany), following the protocol described in the kit. Each reaction contained 100 ng RNA and 0.5 μ M primer in 20 μl final reaction volume. The PCR primer sets used in this study are shown in Table 4 and their sequences correspond to those used by Affymetrix in the MGU74Av2 Genechip used. The housekeeping GAPDH gene was used as internal control. Tm values were calculated using Oligo Calculator (http://mbcf.dfci.harvard.edu/docs/oligocalc.html).

Table 4 Primers used in PCRs reactions

Western blot analysis of cellular extracts

Protein extracts were routinely obtained by adding 250 μl of lysis buffer (1% sodium dodecyl sulfate (SDS), 1 mM phenylmethylsulfonyl fluoride, 20 mM Tris HCl (pH 7.5)], 150 mM NaCl, 1 mM EDTA, 1 mM ethyleneglycoltetraacetate, 1%Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM β-glycerolphosphate, 1 mM Na3VO4 and 1 μg/ml leupeptine) per 10 cm culture dish of each of the cell lines studied. After scraping, the content of the dishes was transferred to eppendorf tubes, vortexed for 1 min and centrifuged at 10 000 g and 4°C for 15 min. Protein concentration of the supernatant was determined using the Pierce reagent method (Pierce, Rockford, IL, USA) as described by the manufacturer.

Lysates (40 μg/lane) were loaded onto SDS polyacrylamide gels and the electrophoresed proteins transferred to polyvinylidene difluoride membranes (Millipore Immobilon-P, Billerica, MA, USA) by electroblotting. Membranes blocked in Tween 20-tris-buffered saline (TTBS) (10 mM Tris-HCl (pH 8.0), 150 mM NaCl, 0.05% Tween 20) plus 1% bovine serum albumin (BSA) were incubated, as appropriate, with dilutions of 0.2 μg/μl of commercial antibodies from Santa Cruz Biotechnologies (Santa Cruz, CA, USA) in 2% BSA and 1 × TTBS. The immunoblots were developed using the commercial kits enhanced chemiluminescence (ECL) and ECL plus (Amershan Pharmacia Biotech, Piscataway, NJ, USA) following procedures recommended by the supplier.

Reverse-phase protein lysate array layout and antibody staining

Cell pellets of the specified cell lines were resuspended in lysis buffer (containing TPER Reagent (Pierce, Rockford, IL, USA), 300 mM NaCl, 1 mM orthovanadate, 200 mM PEFABLOC (AEBSF) (Roche Ltd, Basel, Switzerland), 5 μg/ml aprotinin, 5 μg/ml pepstatin A and 5 μg/ml leupeptin), vortexed for 15 min, centrifuged briefly and incubated on ice for 20 min. Samples were centrifuged at 10 000 r.p.m. during 5 min, and protein concentration in the supernatant was measured. Immediately before arraying, 2 μg/μl of each sample were prepared using a buffer containing a 1:1 mixture of 2 × Tris-Glycine SDS sample buffer (Invitrogen Life Technologies, Carlsbad, CA, USA) and Tissue Protein Extraction Reagent (Pierce, Rockford, IL, USA) plus 2.5% β-mercaptoethanol. Serial dilutions made in lysis buffer were spotted onto nitrocellulose-coated glass slides and the resulting array layouts submitted to staining with primary and secondary antibodies following procedures described previously in detail (Espina et al., 2003; Wulfkuhle et al., 2003). Origin and dilution of the antibodies used was as follows: rabbit anti-Stat1, Cell Signaling, 1:100; rabbit anti-p-Stat1 (S727), Upstate, Charlottesville, VA, USA 1:500; rabbit anti-p-Stat1 (Y701), Upstate, 1:100; mouse anti-Stat2, Transduction Labs, Franklin Lakes, NJ, USA 1:100; rabbit anti-Stat3, Cell Signaling Technology, Inc. Danvers, MA, USA 1:1000; mouse anti-Stat5, Transduction Labs, 1:100). Development of antibody-stained arrays and quantitation of the signal data obtained after scanning the arrays were carried out as described (Herrmann et al., 2003; Wulfkuhle et al., 2003).

Luciferase reporter assays

Transcriptional activity of N-ras+/+ and N-ras−/− cells was assayed using luciferase reporter constructs (8 × ISRE-tkLuc, 8 × GAS-tkLuc and control tkLuc) kindly provided by Dr R Pine (The Public Health Research Institute, Newark, NJ, USA). The 8 × ISRE-tkLuc and the 8 × GAS-tkLuc plasmids contain eight copies of the ISRE or GAS promoter sequences, respectively (Pine et al, 1994). Cell seeded in six-well plates (5 × 105 cells/well) and cultured for 12 h were transfected with reporter plasmids (5.0 μg) using JetPEI (Polyplus transfection, Illkirch, France). phRL-tk plasmid (Promega, Madison, WI, USA) (50 ng) was co-transfected as an internal control. After further culture for 34 h in DMEM with 10% FBS serum, cell extracts were assayed for luciferase activity (Improta and Pine, 1997). Luciferase assays were performed using a dual luciferase reporter kit (Promega, Madison, WI, USA). Luminescence was determined with a MiniLumat LB9506 luminometer (Berthold, Bad Wildbad, Germany).

References

  1. Barash Y, Dehan E, Krupsky M, Franklin W, Geraci M, Friedman N et al. (2004). Comparative analysis of algorithms for signal quantitation from oligonucleotide microarrays. Bioinformatics 20: 839–846.

    CAS  Article  Google Scholar 

  2. Beadling C, Guschin D, Witthuhn BA, Ziemiecki A, Ihle JN, Kerr IM et al. (1994). Activation of JAK kinases and STAT proteins by interleukin-2 and interferon alpha, but not the T cell antigen receptor, in human T lymphocytes. EMBO J 13: 5605–5615.

    CAS  Article  Google Scholar 

  3. Benjamini Y, Drai D, Elmer G, Kafkafi N, Golani I . (2001). Controlling the false discovery rate in behavior genetics research. Behav Brain Res 125: 279–284.

    CAS  Article  Google Scholar 

  4. Bollag G, McCormick F . (1991). Differential regulation of rasGAP and neurofibromatosis gene product activities. Nature 351: 576–579.

    CAS  Article  Google Scholar 

  5. Bolstad BM, Irizarry RA, Astrand M, Speed TP . (2003). A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics 19: 185–193.

    CAS  Article  Google Scholar 

  6. Bos JL . (1989). ras oncogenes in human cancer: a review. Cancer Res 49: 4682–4689.

    CAS  PubMed  Google Scholar 

  7. Brem R, Certa U, Neeb M, Nair AP, Moroni C . (2001). Global analysis of differential gene expression after transformation with the v-H-ras oncogene in a murine tumor model. Oncogene 20: 2854–2858.

    CAS  Article  Google Scholar 

  8. Clyde-Smith J, Silins G, Gartside M, Grimmond S, Etheridge M, Apolloni A et al. (2000). Characterization of RasGRP2, a plasma membrane-targeted, dual specificity Ras/Rap exchange factor. J Biol Chem 275: 32260–32267.

    CAS  Article  Google Scholar 

  9. Colicelli J . (2004). Human RAS superfamily proteins and related GTPases. Sci STKE 2004: RE13.

    PubMed  PubMed Central  Google Scholar 

  10. Croonquist PA, Linden MA, Zhao F, Van Ness BG . (2003). Gene profiling of a myeloma cell line reveals similarities and unique signatures among IL-6 response, N-ras-activating mutations, and coculture with bone marrow stromal cells. Blood 102: 2581–2592.

    CAS  Article  Google Scholar 

  11. Davies SA, Houslay MD, Wakelam MJ . (1989). The effects of p21N-ras expression in NIH-3T3 cells upon cyclic AMP metabolism. Biochim Biophys Acta 1013: 173–179.

    CAS  Article  Google Scholar 

  12. de Castro IP, Diaz R, Malumbres M, Hernandez MI, Jagirdar J, Jimenez M et al. (2003). Mice deficient for N-ras: impaired antiviral immune response and T-cell function. Cancer Res 63: 1615–1622.

    Google Scholar 

  13. Dumaz N, Marais R . (2005). Integrating signals between cAMP and the RAS/RAF/MEK/ERK signalling pathways. Based on the anniversary prize of the Gesellschaft fur Biochemie und Molekularbiologie Lecture delivered on 5 July 2003 at the Special FEBS Meeting in Brussels. FEBS J 272: 3491–3504.

    CAS  Article  Google Scholar 

  14. Ehrhardt A, David MD, Ehrhardt GR, Schrader JW . (2004). Distinct mechanisms determine the patterns of differential activation of H-Ras, N-Ras, K-Ras 4B, and M-Ras by receptors for growth factors or antigen. Mol Cell Biol 24: 6311–6323.

    CAS  Article  Google Scholar 

  15. Eskandarpour M, Kiaii S, Zhu C, Castro J, Sakko AJ, Hansson J . (2005). Suppression of oncogenic NRAS by RNA interference induces apoptosis of human melanoma cells. Int J Cancer 115: 65–73.

    CAS  Article  Google Scholar 

  16. Espina V, Mehta AI, Winters ME, Calvert V, Wulfkuhle J, Petricoin EF et al. (2003). Protein microarrays: molecular profiling technologies for clinical specimens. Proteomics 3: 2091–2100.

    CAS  Article  Google Scholar 

  17. Esteban LM, Vicario-Abejon C, Fernandez-Salguero P, Fernandez-Medarde A, Swaminathan N, Yienger K et al. (2001). Targeted genomic disruption of H-ras and N-ras, individually or in combination, reveals the dispensability of both loci for mouse growth and development. Mol Cell Biol 21: 1444–1452.

    CAS  Article  Google Scholar 

  18. Hamilton M, Wolfman A . (1998). Oncogenic Ha-Ras-dependent mitogen-activated protein kinase activity requires signaling through the epidermal growth factor receptor. J Biol Chem 273: 28155–28162.

    CAS  Article  Google Scholar 

  19. Hancock JF . (2003). Ras proteins: different signals from different locations. Nat Rev Mol Cell Biol 4: 373–384.

    CAS  Article  Google Scholar 

  20. Herrmann PC, Gillespie JW, Charboneau L, Bichsel VE, Paweletz CP, Calvert VS et al. (2003). Mitochondrial proteome: altered cytochrome c oxidase subunit levels in prostate cancer. Proteomics 3: 1801–1810.

    CAS  Article  Google Scholar 

  21. Ihrie RA, Marques MR, Nguyen BT, Horner JS, Papazoglu C, Bronson RT et al. (2005). Perp is a p63-regulated gene essential for epithelial integrity. Cell 120: 843–856.

    CAS  Article  Google Scholar 

  22. Improta T, Pine R . (1997). Susceptibility to virus infection is determined by a Stat-mediated response to the autocrine effect of virus-induced type I interferon. Cytokine 9: 383–393.

    CAS  Article  Google Scholar 

  23. Irizarry RA, Bolstad BM, Collin F, Cope LM, Hobbs B, Speed TP . (2003a). Summaries of Affymetrix GeneChip probe level data. Nucleic Acids Res 31: e15.

    Article  Google Scholar 

  24. Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U . (2003b). Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 4: 249–264.

    Article  Google Scholar 

  25. Johnson L, Greenbaum D, Cichowski K, Mercer K, Murphy E, Schmitt E et al. (1997). K-ras is an essential gene in the mouse with partial functional overlap with N-ras. Genes Dev 11: 2468–2481.

    CAS  Article  Google Scholar 

  26. Jones MK, Jackson JH . (1998). Ras-GRF activates Ha-Ras, but not N-Ras or K-Ras 4B, protein in vivo. J Biol Chem 273: 1782–1787.

    CAS  Article  Google Scholar 

  27. Kisseleva T, Bhattacharya S, Braunstein J, Schindler CW . (2002). Signaling through the JAK/STAT pathway, recent advances and future challenges. Gene 285: 1–24.

    CAS  Article  Google Scholar 

  28. Koera K, Nakamura K, Nakao K, Miyoshi J, Toyoshima K, Hatta T et al. (1997). K-ras is essential for the development of the mouse embryo. Oncogene 15: 1151–1159.

    CAS  Article  Google Scholar 

  29. Leon J, Guerrero I, Pellicer A . (1987). Differential expression of the ras gene family in mice. Mol Cell Biol 7: 1535–1540.

    CAS  Article  Google Scholar 

  30. Li C, Wong WH . (2001). Model-based analysis of oligonucleotide arrays: expression index computation and outlier detection. Proc Natl Acad Sci USA 98: 31–36.

    CAS  Article  Google Scholar 

  31. Liao J, Wolfman JC, Wolfman A . (2003). K-ras regulates the steady-state expression of matrix metalloproteinase 2 in fibroblasts. J Biol Chem 278: 31871–31878.

    CAS  Article  Google Scholar 

  32. Maher J, Baker DA, Manning M, Dibb NJ, Roberts IA . (1995). Evidence for cell-specific differences in transformation by N-, H- and K-ras. Oncogene 11: 1639–1647.

    CAS  PubMed  Google Scholar 

  33. Matallanas D, Sanz-Moreno V, Arozarena I, Calvo F, Agudo-Ibanez L, Santos E et al. (2006). Distinct utilization of effectors and biological outcomes resulting from site-specific Ras activation: Ras functions in lipid rafts and Golgi complex are dispensable for proliferation and transformation. Mol Cell Biol 26: 100–116.

    CAS  Article  Google Scholar 

  34. Murtagh F . (1985). Multidimensional Clustering Algorithms. COMPSTAT Lectures 4. Physica-Verlag: Wuerzburg.

    Google Scholar 

  35. Ohnami S, Aoki K, Yoshida K, Hatanaka K, Suzuki K, Sasaki H et al. (2003). Expression profiles of pancreatic cancer cell lines infected with antisense K-ras-expressing adenoviral vector. Biochem Biophys Res Commun 309: 798–803.

    CAS  Article  Google Scholar 

  36. Oliva JL, Zarich N, Martinez N, Jorge R, Castrillo A, Azanedo M et al. (2004). The P34G mutation reduces the transforming activity of K-Ras and N-Ras in NIH 3T3 cells but not of H-Ras. J Biol Chem 279: 33480–33491.

    CAS  Article  Google Scholar 

  37. Perez de Castro I, Bivona TG, Philips MR, Pellicer A . (2004). Ras activation in Jurkat T cells following low-grade stimulation of the T-cell receptor is specific to N-Ras and occurs only on the Golgi apparatus. Mol Cell Biol 24: 3485–3496.

    Article  Google Scholar 

  38. Petricoin III EF, Ito S, Williams BL, Audet S, Stancato LF, Gamero A et al. (1997). Antiproliferative action of interferon-alpha requires components of T-cell-receptor signalling. Nature 390: 629–632.

    CAS  Article  Google Scholar 

  39. Pine R, Canova A, Schindler C . (1994). Tyrosine phosphorylated p91 binds to a single element in the ISGF2/1RF-1 promoter to mediate induction by IFN alpha and IFN gamma, and is likely to autoregulate the p91 gene. EMBO J 13: 158–167.

    CAS  Article  Google Scholar 

  40. Platanias LC . (2005). Mechanisms of type-I- and type-II-interferon-mediated signalling. Nat Rev Immunol 5: 375–386.

    CAS  Article  Google Scholar 

  41. Plowman SJ, Hancock JF . (2005). Ras signaling from plasma membrane and endomembrane microdomains. Biochim Biophys Acta 1746: 274–283.

    CAS  Article  Google Scholar 

  42. Plowman SJ, Williamson DJ, O’Sullivan MJ, Doig J, Ritchie AM, Harrison DJ et al. (2003). While K-ras is essential for mouse development, expression of the K-ras 4A splice variant is dispensable. Mol Cell Biol 23: 9245–9250.

    CAS  Article  Google Scholar 

  43. Potenza N, Vecchione C, Notte A, De Rienzo A, Rosica A, Bauer L et al. (2005). Replacement of K-Ras with H-Ras supports normal embryonic development despite inducing cardiovascular pathology in adult mice. EMBO Rep 6: 432–437.

    CAS  Article  Google Scholar 

  44. Reuther GW, Der CJ . (2000). The Ras branch of small GTPases: Ras family members don’t fall far from the tree. Curr Opin Cell Biol 12: 157–165.

    CAS  Article  Google Scholar 

  45. Rocks O, Peyker A, Kahms M, Verveer PJ, Koerner C, Lumbierres M et al. (2005). An acylation cycle regulates localization and activity of palmitoylated Ras isoforms. Science 307: 1746–1752.

    CAS  Article  Google Scholar 

  46. Rojas JM, Santos E . (2002). ras genes and human cancer: different implications and different roles. Curr Genom 3: 295–311.

    CAS  Article  Google Scholar 

  47. Roy S, Luetterforst R, Harding A, Apolloni A, Etheridge M, Stang E et al. (1999). Dominant-negative caveolin inhibits H-Ras function by disrupting cholesterol-rich plasma membrane domains. Nat Cell Biol 1: 98–105.

    CAS  Article  Google Scholar 

  48. Ruiz-Vela A, Opferman JT, Cheng EH, Korsmeyer SJ . (2005). Proapoptotic BAX and BAK control multiple initiator caspases. EMBO Rep 6: 379–385.

    CAS  Article  Google Scholar 

  49. Su AI, Wiltshire T, Batalov S, Lapp H, Ching KA, Block D et al. (2004). A gene atlas of the mouse and human protein-encoding transcriptomes. Proc Natl Acad Sci USA 101: 6062–6067.

    CAS  Article  Google Scholar 

  50. Sweet-Cordero A, Mukherjee S, Subramanian A, You H, Roix JJ, Ladd-Acosta C et al. (2005). An oncogenic KRAS2 expression signature identified by cross-species gene-expression analysis. Nat Genet 37: 48–55.

    CAS  Article  Google Scholar 

  51. Takahashi C, Contreras B, Iwanaga T, Takegami Y, Bakker A, Bronson RT et al. (2006). Nras loss induces metastatic conversion of Rb1-deficient neuroendocrine thyroid tumor. Nat Genet 38: 118–123.

    CAS  Article  Google Scholar 

  52. Townsend PA, Scarabelli TM, Davidson SM, Knight RA, Latchman DS, Stephanou A . (2004). STAT-1 interacts with p53 to enhance DNA damage-induced apoptosis. J Biol Chem 279: 5811–5820.

    CAS  Article  Google Scholar 

  53. Tusher VG, Tibshirani R, Chu G . (2001). Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA 98: 5116–5121.

    CAS  Article  Google Scholar 

  54. Umanoff H, Edelmann W, Pellicer A, Kucherlapati R . (1995). The murine N-ras gene is not essential for growth and development. Proc Natl Acad Sci USA 92: 1709–1713.

    CAS  Article  Google Scholar 

  55. Vasseur S, Malicet C, Calvo EL, Labrie C, Berthezene P, Dagorn JC et al. (2003). Gene expression profiling by DNA microarray analysis in mouse embryonic fibroblasts transformed by rasV12 mutated protein and the E1A oncogene. Mol Cancer 2: 19.

    Article  Google Scholar 

  56. Voice JK, Klemke RL, Le A, Jackson JH . (1999). Four human ras homologs differ in their abilities to activate Raf-1, induce transformation, and stimulate cell motility. J Biol Chem 274: 17164–17170.

    CAS  Article  Google Scholar 

  57. Walsh AB, Bar-Sagi D . (2001). Differential activation of the Rac pathway by Ha-Ras and K-Ras. J Biol Chem 276: 15609–15615.

    CAS  Article  Google Scholar 

  58. Wolfman JC, Palmby T, Der CJ, Wolfman A . (2002). Cellular N-Ras promotes cell survival by downregulation of Jun N-terminal protein kinase and p38. Mol Cell Biol 22: 1589–1606.

    CAS  Article  Google Scholar 

  59. Wulfkuhle JD, Aquino JA, Calvert VS, Fishman DA, Coukos G, Liotta LA et al. (2003). Signal pathway profiling of ovarian cancer from human tissue specimens using reverse-phase protein microarrays. Proteomics 3: 2085–2090.

    CAS  Article  Google Scholar 

  60. Yan J, Roy S, Apolloni A, Lane A, Hancock JF . (1998). Ras isoforms vary in their ability to activate Raf-1 and phosphoinositide 3-kinase. J Biol Chem 273: 24052–24056.

    CAS  Article  Google Scholar 

  61. Zuber J, Tchernitsa OI, Hinzmann B, Schmitz AC, Grips M, Hellriegel M et al. (2000). A genome-wide survey of RAS transformation targets. Nat Genet 24: 144–152.

    CAS  Article  Google Scholar 

Download references

Acknowledgements

We thank R Pine (The Public Health Research Institute, Newark, NJ) for ISRE and GAS reporter constructs and E Petricoin (FDA/NIH Proteomic Facility, Bethesda, MD, USA) for support with protein array layout generation and analysis. This work was supported by Grants SAF2003-04177 and GEN2003-20239-C06-02 from MEC and Grant PI021570 from MSC, as well as institutional support from Red Temática C03/10 de Investigación de Centros de Cáncer (RTICCC) from ISCIII, MSC, Spain. CG was supported by Ramón y Cajal Program. EC was a predoctoral fellow from MEC.

Author information

Affiliations

Authors

Corresponding author

Correspondence to E Santos.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Castellano, E., De Las Rivas, J., Guerrero, C. et al. Transcriptional networks of knockout cell lines identify functional specificities of H-Ras and N-Ras: significant involvement of N-Ras in biotic and defense responses. Oncogene 26, 917–933 (2007). https://doi.org/10.1038/sj.onc.1209845

Download citation

Keywords

  • H-ras
  • N-ras
  • microarray
  • genomics
  • transcriptome
  • cluster analysis
  • differential gene expression

Further reading

Search

Quick links