Serum-dependent transcriptional networks identify distinct functional roles for H-Ras and N-Ras during initial stages of the cell cycle
© Castellano et al.; licensee BioMed Central Ltd. 2009
Received: 2 July 2009
Accepted: 6 November 2009
Published: 6 November 2009
Using oligonucleotide microarrays, we compared transcriptional profiles corresponding to the initial cell cycle stages of mouse fibroblasts lacking the small GTPases H-Ras and/or N-Ras with those of matching, wild-type controls.
Serum-starved wild-type and knockout ras fibroblasts had very similar transcriptional profiles, indicating that H-Ras and N-Ras do not significantly control transcriptional responses to serum deprivation stress. In contrast, genomic disruption of H-ras or N-ras, individually or in combination, determined specific differential gene expression profiles in response to post-starvation stimulation with serum for 1 hour (G0/G1 transition) or 8 hours (mid-G1 progression). The absence of N-Ras caused significantly higher changes than the absence of H-Ras in the wave of transcriptional activation linked to G0/G1 transition. In contrast, the absence of H-Ras affected the profile of the transcriptional wave detected during G1 progression more strongly than did the absence of N-Ras. H-Ras was predominantly functionally associated with growth and proliferation, whereas N-Ras had a closer link to the regulation of development, the cell cycle, immunomodulation and apoptosis. Mechanistic analysis indicated that extracellular signal-regulated kinase (ERK)-dependent activation of signal transducer and activator of transcription 1 (Stat1) mediates the regulatory effect of N-Ras on defense and immunity, whereas the pro-apoptotic effects of N-Ras are mediated through ERK and p38 mitogen-activated protein kinase signaling.
Our observations confirm the notion of an absolute requirement for different peaks of Ras activity during the initial stages of the cell cycle and document the functional specificity of H-Ras and N-Ras during those processes.
The mammalian H-Ras, N-Ras and K-Ras proteins are highly related small GTPases functioning as critical components of cellular signaling pathways controlling proliferation, differentiation or survival. They act as molecular switches cycling between inactive (GDP-bound) and active (GTP-bound) states in a process modulated under physiological conditions by a variety of specific regulatory proteins, including GAPs (GTPase activating proteins) and GEFs (guanine nucleotide exchange factors) [1–3]. Hyperactivating point mutations of these proteins are frequently associated with pathological conditions, particularly the development of various forms of human cancer [4, 5]. The three main mammalian ras genes appear to be ubiquitously expressed, although specific differences have been reported for particular isoforms regarding their expression levels in different cell types and tissues or their intracellular processing and subsequent location to different subcellular compartments [1, 3].
Early studies focusing on the shared sequence homology and identical in vitro effector activation pathways suggested that the three Ras protein isoforms were functionally redundant [2, 4]. However, many other reports based on different experimental approaches support the notion that these three members of the Ras family may play specialized cellular roles [1, 3, 6]. Thus, the preferential activation of specific ras genes in particular tumor types [4, 5], the different transforming potential of transfected ras genes in different cellular contexts [7, 8], the distinct sensitivities exhibited by different Ras family members for functional interactions with their GAPs, GEFs or downstream effectors [9–15], or differences among Ras isoforms regarding their intracellular processing pathways and their differential compartmentalization to specific plasma membrane microdomains or intracellular compartments [12, 14, 16–21] provide strong evidence in favor of the notion of functional specificity. The study of Ras knockout strains provides additional in vivo evidence for functional specificity. Thus, whereas disruption of K-ras 4B is embryonic lethal [22, 23], H-ras, N-ras and K-ras4A single knockout mice and H-ras/N-ras double knockout mice are perfectly viable [22, 24–26], indicating that only K-ras is necessary and sufficient for full embryonic development and suggesting that K-Ras performs specific function(s) that cannot be carried out by either H-Ras or N-Ras. A recent study describing that the knock-in of H-ras at the K-ras locus results in viable adult mice  suggests that the mortality of K-ras knockout may derive not from intrinsic inability of the other Ras isoforms to compensate for K-Ras function but rather from their inability to be expressed in the same locations (embryonic compartments) or at the same time (developmental stage) as K-Ras. Finally, additional experimental support for the notion of functional specificity of H-, N- and K-Ras proteins derives from genomic or proteomic profiling of cell lines transformed by exogenous ras oncogenes [28–34] or devoid of specific Ras proteins . In particular, our recent characterization of the transcriptional networks of actively growing cultures of fibroblast cells harboring single or double null mutations in the H-ras and N-ras loci clearly supported the notion of different functions for H-Ras and N-Ras by documenting a significant involvement of N-Ras in immunomodulation/defense and apoptotic responses .
It is also well established that Ras proteins play capital roles in regulation of the initiation and progression of the cell cycle [1, 3, 5, 36]. A number of reports have documented the absolute requirement for Ras activity at different points between G0 and S phase, after growth factor stimulation of quiescent, serum-arrested (G0) cells. Indeed, the available experimental evidence indicates that the contribution of Ras activity is absolutely needed for both the initial entry into the cell cycle (G0/G1 transition) and for the subsequent G1 progression, in a process to which multiple Ras effector pathways can contribute [36–41]. However, the exact mechanisms regulating the participation of Ras proteins in cell cycle activation and subsequent progression are still largely unknown. It is also unknown whether the different Ras isoforms play specific or redundant functional roles in those processes.
Our previous characterization of the transcriptional profiles of unsynchronized, exponentially growing cultures of H-ras and N-ras knockout fibroblasts in the presence of serum demonstrated the functional specificity of those proteins in proliferating, actively cycling cells . In this report, we were specifically interested in ascertaining whether N-Ras and H-Ras play also specific - or redundant - functional roles during the initial stages of the cell cycle. In particular, we wished to characterize the participation, if any, of these proteins in the process of entry into the cell cycle of G0, growth arrested cells (G0/G1 transition) and the subsequent steps of progression through early G1. For this purpose, we used commercial microarrays to characterize the profiles of genomic expression of wild-type (WT) and ras knockout fibroblasts (H-ras-/-, N-ras-/-, H-ras-/-/N-ras-/-) that had been subjected to serum starvation (G0) or to subsequent incubation in the presence of serum for a short, 1-hour period (G0/G1 transition) or for 8 hours (mid-G1 progression). Our data support the notion of functional specificity for H-Ras and N-Ras by documenting the occurrence of specific transcriptional profiles associated with the absence of H-Ras and/or N-Ras during defined moments of the early stages of the cell cycle.
Analysis of serum-dependent, transcriptional profiles in wild-type and rasknockout fibroblasts
To ensure statistical significance, four independent microarray hybridizations were carried out for each of the time points studied with WT cell samples, and three independent hybridizations were performed for each of the experimental conditions tested in the three different ras knockout genotypes under study (H-ras-/-, N-ras-/-, H-ras-/-/N-ras-/-). After robust normalization of the signals in all 39 separate microarray hybridizations included in this study by means of robust multi-array average software , the Significance Analysis of Microarrays (SAM) algorithm  was applied to identify the sets of differentially expressed genes showing statistically significant changes of gene expression levels when comparing the transcriptome of starved WT fibroblasts (Figure 1, Control) with that of the rest of the samples and conditions included in this study for WT and knockout cells. Figure 1 summarizes the experimental conditions and quantitative results of the microarray hybridizations performed at the different time points analyzed for each WT and ras knockout genotype under study, and shows the numbers of differentially expressed probesets (induced or repressed with regards to the 0 h, WT control) that were identified under the stringent selection conditions (false discovery rate (FDR) = 0.09) applied in the SAM comparisons.
Transcriptional profiles of serum-starved fibroblasts
Initial comparison of the gene expression patterns obtained for fibroblasts of all different genotypes analyzed after 24 hours of serum starvation showed that the transcriptional profile of the control, WT fibroblasts was very similar to those of similarly treated H-ras-/- and N-ras-/- knockout cells, indicating that H-Ras and N-Ras exert rather minor influence over the transcriptomic profile resulting from submitting fibroblasts to the stress of serum deprivation (Figure 1). We observed that the individual H-ras-/- and N-ras-/- knockouts showed negligible numbers of overall transcriptomic changes and only the simultaneous absence of both N-Ras and H-Ras in the double knockout cells allowed identification of a short list of 15 differentially expressed gene probesets in comparison to the serum-starved, control WT fibroblasts at the FDR value applied (Figure 1; Table S1 in Additional data file 1). Consideration of the short list of gene probesets distinguishing the H-ras-/- knockout cells from their corresponding WT controls suggested a predominant involvement of genes affecting cell growth and proliferation, whereas the list of genes differentially expressed in serum-starved, N-ras-/- knockout cells indicated a higher prevalence of genes related to transcriptional processes and development or differentiation (Table S1a, b in Additional data file 1). The double knockout (H-ras-/-/N-ras-/-), starved cells allowed identification of a somewhat more extensive list of differentially expressed genes (Table S1c in Additional data file 1) that confirmed some of the functional tendencies observed in the individual ras knockouts. For example, Crabp2, a gene coding for a retinoid binding protein functionally involved in morphogenesis and organogenesis [51, 52] was highly overexpressed in the single N-ras-/- cells and was also the most highly overexpressed locus detected in the double knockout (H-ras-/-/N-ras-/-) fibroblasts (Table S1b, c in Additional data file 1).
Serum-induced transcriptional profiles in wild-type fibroblasts
Besides analyzing the effect of serum deprivation on the cellular transcriptome, we also wished to determine the effect, if any, of eliminating H-Ras and/or N-Ras on the transcriptional profile of fibroblasts cultured in the presence of fetal bovine serum (FBS) for short periods of time (1 hour or 8 hours) post-starvation. Computational, pair-wise comparisons of the transcriptional profile of control WT, serum-starved fibroblasts with those obtained for the same cells after incubation in the presence of FBS generated two separate lists of differentially expressed genes reflecting the actual transcriptional changes caused in WT, growth arrested (G0) fibroblasts by stimulation with serum for 1 hour (Table S2 in Additional data file 1) or after 8 hours of serum incubation (Table S3 in Additional data file 1).
It is noteworthy that the transcriptomic profile depicted in Table S2 in Additional data file 1 for serum-deprived, growth arrested, WT fibroblasts treated with FBS for a short 1-hour period contained only induced genes, as no repressed loci could be identified as differentially expressed under the stringent comparison conditions used. As expected, the subset of loci showing highest transcriptional activation in Table S2 in Additional data file 1 included a series of genes (Jun, Fos, Egr, Atg, Atf-, Zfp-Ier-, and so on) belonging to the previously described category of IE genes [53–55] known to be activated in starved, G0 fibroblasts shortly after exposure to serum [43, 46, 47, 56–58]. Interestingly, the differential expression of a large proportion of the most highly activated IE loci detected in WT fibroblasts (Table S2 in Additional data file 1) was also observed in the transcriptional profiles of H-ras-/-, N-ras-/- and H-ras-/-/N-ras-/- knockout fibroblasts that were similarly starved and treated with serum for 1 hour, suggesting that H-Ras and N-Ras are not participating directly in the regulation of their transcriptional activation. On the other hand, we observed that a significant number of genes listed in Table S2 in Additional data file 1 at medium-low values of transcriptional activation (as judged by R.fold or d(i) values) did not score as differentially expressed in the transcriptional profiles of corresponding ras knockout fibroblasts treated under similar conditions (see the column 'Differential expression not kept' in Table S2 in Additional data file 1), suggesting that in those cases H-Ras or N-Ras may be actively involved in regulation of their expression.
The list of loci showing differential expression after 8 hours of serum stimulation (Table S3 in Additional data file 1) was longer and clearly different from that of early-expressed genes after 1 hour of serum treatment. In contrast to Table S2, Table S3 in Additional data file 1 includes both induced (168 probesets; 158 genes) and repressed (129 probesets; 126 genes) loci (Figure 1), and showed very minor overlapping with the list of induced-only, IE genes included in Table S2 in Additional data file 1. Consistent with the previously described molecular mechanisms triggering G1/S transition as a consequence of Rb phosphorylation and subsequent induction of E2F-dependent transcription, this loci list includes a number of known E2F targets (E2f3, Myc, Ctfg, Smad, Cyr61, Psme3, Tpm2, Vegfb, and so on) [48, 59–62]. Interestingly, some of the most highly overexpressed genes in Table S3 (see the 'R.fold' column) were functionally related to inhibition of proteolytic activities (Serpine1 and Serpinb2, Timp1, and so on) or to interaction with components of the extracellular matrix (Hbegf, Ctgf). Finally, as in Table S2 in Additional data file 1, a significant number of the loci differentially expressed in WT fibroblasts after 8 hours of serum stimulation did not keep such differential expression in the transcriptome of corresponding ras knockout fibroblast counterparts subjected to the same 8-hour serum incubation (see the column 'Differential expression not kept' in Table S3 in Additional data file 1). Interestingly, in most cases such loss of transcriptional activation or repression concerned specifically the single N-ras-/- or the double H-ras-/-/N-ras-/- knockout cells, an observation suggesting very different functional contributions of N-Ras and H-Ras to the regulation of gene expression during G1 progression in fibroblasts.
Transcriptional waves induced by serum in H-ras and N-rasknockout fibroblasts
Whereas the absence of H-Ras or N-Ras caused negligible transcriptional changes relative to WT, serum-deprived fibroblasts (Figure 1, 0 h), genomic disruption of H-ras-/- and/or N-ras-/-, individually or in combination, was associated with the occurrence of significant transcriptional changes caused by short-term incubation of the knockout fibroblasts with serum (Figure 1, 1 h and 8 h). Thus, important numbers of differentially expressed genes were detected when performing stringent pair-wise comparisons (FDR = 0.09) between the microarray hybridization pattern of serum-starved, G0 arrested WT fibroblasts and those of H-ras-/-, N-ras-/- or H-ras-/-/N-ras-/- fibroblasts subjected to serum starvation and subsequent stimulation with serum for 1 hour (G0/G1 transition) or 8 hours (G1 progression) (Figure 1, 1 h and 8 h).
Quantitative analysis of the microarray hybridization data showed that, among all different fibroblast genotypes tested, the N-ras-/- fibroblasts exhibited the highest numbers of IE, differentially expressed genes after 1 hour of serum stimulation (786 altered probesets in N-ras-/- fibroblasts versus 439 probesets in H-ras-/- fibroblasts) (Figure 1, 1 h). In contrast, the H-ras-/- genotype was associated with the higher number of differentially expressed loci detected during G1 progression, after 8 hours of serum stimulation (1,078 affected probesets in H-ras-/- fibroblasts versus 399 probesets in N-ras-/- fibroblasts; Figure 1, 8 h). These data suggest very different roles for H-Ras and N-Ras in regulation of cellular transcriptional responses to serum and reinforces the notion of specific, non-overlapping molecular functions for the different Ras isoforms. Our observation of two distinct waves of transcriptional activation (after 1 hour and 8 hours of serum stimulation) that are preferentially linked, respectively, to the N-ras-/- or the H-ras-/- genotype is consistent with the previously reported absolute requirement for Ras activity during at least two separate phases of the early G0 to S interval [36–41]. This raises the interesting possibility of a preferential functional involvement of N-Ras during the early phase and of H-Ras during a later phase of the period of absolute Ras activity requirement defined by means of microinjection of neutralizing Ras antibodies and dominant negative Ras forms [63–65].
Functional signatures linked to deficiency of H-Ras or N-Ras in the transcriptional profile of serum-induced fibroblasts
The dendrogram analyzing the short-term wave of transcriptional response to serum stimulation for 1 hour allowed discrimination of two main vertical branches (Figure 3a). One of them encompassed the hybridization data corresponding to the N-ras-/- and H-ras-/-/N-ras-/- knockout cells, whereas the second one contained those of the H-ras-/- and WT fibroblasts (Figure 3a, columns). This branching distribution indicated that the transcriptional profile of H-ras-/- cells after 1 hour of serum induction is closest to that of WT fibroblasts, whereas the expression pattern of the H-ras-/-/N-ras-/- cells is intermediate and more similar to that of the N-ras-/- cells, which is located farthest away from the WT branch. This behavior is consistent with our previous suggestion (Figure 1) of a preferential contribution of N-Ras over H-Ras in generating the first transcriptional wave of immediate-early responses to serum stimulation for 1 hour. The horizontal branching of the dendrogram allowed identification of a series of gene blocks that clearly discriminated the transcriptional profiles of the different WT and ras knockout genotypes under study (Figure 3a, blocks 1-8).
Functional signatures of differentially expressed genes induced or suppressed in H- ras -/- and/or N- ras -/- fibroblasts after serum stimulation for 1 hour (G0/G1 transition)
Immunity and defense
Fas, Cxcl10, Il6, Irf1, Psmb9, Mx1, Mx2, Cxcl2, Tap1, Ifi202b
Bax, Bid, Fas, Gadd45b, Perp, Tnfrsf11b, Phlda1, Tnfaip3, Trp53
Rela, Stat1, Stat5a, Trp53
MAPK signaling cascade
Fas, Mapkapk2, Gadd45b, Dusp8, Trp53, Map3k8, Flnb
Ehd1, Mx1, Mx2, Iigp2, Rhoj
Gnb1, Vegfa, Irs2
Ccnd2, Ccng2, Cdkn2a, Ppp1cc, Spin, Tsc2, Anapc4, Sash1
Cell adhesion and cytoskeleton organization
Nras, Pik3r2, Ppp1cc
Insulin signaling pathway
Nras, Pik3r2, Ppp1cc, Tsc2, Pck2
Functional signatures of differentially expressed genes induced or suppressed in H- ras -/- and/or N- ras -/- fibroblasts after serum stimulation for 8 hours (G1 progression)
Eif2s1, Rnu3ip2, Nola2, Cpsf4, Rnpc1, Mrpl20, Ddx18, Sf3a1, Hnrpll, Lsm8
Iars, Tars, Eif2s1, Eftud2, Nola2, Rpp30, Mrpl20
Eftud2, Rnu3ip2, Nola2, Mrpl20, Hnrpll, Lsm8
Rnps1, Eftud2, Sf3a1, Lsm8
Translation initiation factor activity
Eif2s1, Eif4ebp1, AU014645
Regulation of cell cycle
Ccnd2, Junb, Kras
Extracellular matrix interaction
Col18a1, Mmp10, Mmp13, Mmp9
Ccne2, Mcm5, Rbl1, Trp53, Cdc6
Mcm5, Pold1, Rrm2, Myst2, Cdc6
Birc5, Bcap29, Perp, Tnfrsf11b, Trp53
Ankrd1, Meis1, Tcf20
Taken together, these data reinforce the notion of non-overlapping functional roles for H-Ras and N-Ras in mammalian fibroblast cells and are consistent with our previous observations on actively growing fibroblasts  that pointed to preferential functional roles of H-Ras in growth and proliferation and of N-Ras in transcriptional regulation of immune/defense responses and apoptosis.
Serum-dependent gene expression signatures linked to deficiency of H-ras and/or N-ras
The list of differentially expressed genes identified in H-ras-/- fibroblasts stimulated with serum for 1 hour (Table S4 in Additional data file 1) includes a high percentage of loci related to signal transduction pathways (Figure 4), including Wnt-, transforming growth factor beta- and Ras-dependent signaling pathways. Among others, a notable change was a significant reduction in the expression level of the p110alpha subunit of phosphoinositide-3 kinase (PI3K; Table S4 in Additional data file 1). Furthermore, confirming the conclusions from the global analyses in Figure 3 and Tables 1 and 2, the expression profile of H-ras-/- fibroblasts stimulated with serum for 1 hour showed specifically increased percentages of differentially expressed genes functionally related to cell development and cell growth and proliferation (Figure 4; Table S4 in Additional data file 1).
Differential gene expression during G1 progression in H-ras-/- fibroblasts stimulated with serum for 8 hours (Table S7 in Additional data file 1) involved a high percentage of loci related to specific functional categories such as signal transduction, transcription, RNA processing, protein biosynthesis or ubiquitin interaction (Figure 4). Noticeable with regard to signal transduction was the increased expression of a number of important G protein subunits or small GTPases (including, among others, K-Ras), as well as specific regulatory proteins with GAP or GEF activity (Table S7 in Additional data file 1). In contrast to the profile of IE gene expression in H-ras-/- cells during G0/G1 transition, the profile of H-ras-/- cells stimulated with serum for 8 hours showed a clear increase in the number of differentially expressed loci related to functional categories such as RNA metabolism and processing, protein biosynthesis and ribosome biogenesis (Figure 4). Particularly interesting in this regard was the specific detection of significantly increased expression levels of various tRNA synthetases, translation regulatory factors and ribosomal proteins (both cytoplasmic and mitochondrial; Table S7 in Additional data file 1). Interestingly, the increased expression of tRNA acyl synthetases was conserved in similarly treated, double knockout H-ras-/-/N-ras-/-cells, but not in single knockout N-ras-/- cells (Tables S8 and S9 in Additional data file 1). The concentration of specific transcriptional alterations on functional categories related to cellular growth and proliferation (that is, transcription, protein biosynthesis or primary cell metabolism) is consistent with our previous proposition of a predominant role of H-Ras in controlling the second wave of serum-induced transcriptional activation occurring in fibroblasts during G1 progression after 8 h of incubation in the presence of serum (Figure 1, Tables 1 and 2).
The list of differentially expressed genes specifically associated with the absence of N-Ras in fibroblasts stimulated with serum for 1 hour (Table S5 in Additional data file 1) showed a high proportion of loci functionally related to processes of cellular signal transduction, transcription and primary metabolism. Although similarly treated H-ras-/- fibroblasts also showed predominant alteration of these functional categories (Table S4 in Additional data file 1), the identity of the genes listed under these functional headings differed significantly between the H-ras-/- and N-ras-/- genotypes. In particular, the elevated levels of specific transcription-related genes detected in N-ras-/- fibroblasts incubated with serum for 1 hour (Table S5 in Additional data file 1; Figure 4) confirms the functional signature for transcription detected in the global, multi-class analyses depicted in Tables 1 and 2 and is consistent with the predominant regulatory role previously attributed to N-Ras during the first transcriptional wave of the response of fibroblasts to serum (Figure 1). The detection of significantly increased levels of genes concerned with immunity/defense and response to interferon in these N-ras-/- fibroblasts was also striking (Table S5 in Additional data file 1; Figure 4). Interestingly, the increased expression of this functional category of genes was restricted to, and highly specific for, the N-ras-/- genotype and was of greater quantitative significance during the early transcriptional wave of response to 1 hour of stimulation with serum (G0/G1) than during G1 progression after 8 hours of serum stimulation (Figure 4). Consistent with these observations, a preferential functional involvement of N-Ras with immunity and defense responses was also previously described in serum-supplemented, unsynchronized, actively growing cultures of N-ras-/- cells . Regarding signal transduction, Table S5 in Additional data file 1 includes significant numbers of over-expressed kinase kinases as well as repressed phosphatases, G protein subunits and Ras-related small GTPases. It was also remarkable to identify Pik3ca (the p110 alpha polypeptide of PI3K) and Pik3r2 (its regulatory p85 subunit) among the most highly repressed loci in the list (Table S5 in Additional data file 1). The simultaneous differential expression of genes related to cell migration and adhesion, together with the repression of specific members of the Rho and Rac families, may suggest functional effects over cell motility under these particular experimental conditions.
The transcriptional profile of N-ras-/- cells stimulated with serum for 8 hours (Table S8 in Additional data file 1) showed specifically high representation of functional categories such as primary cell metabolism, signal transduction, cell development and differentiation and cell adhesion (Figure 4). In particular, the categories of primary cell metabolism and cell development and differentiation showed the highest quantitative increases in comparison to the same cells stimulated with serum for 1 hour only (Figure 4). The list of differentially expressed genes related to signal transduction is shorter for N-ras-/- cells stimulated with serum for 8 hours (Table S8 in Additional data file 1) than in the same cells treated with serum for 1 hour (Table S5 in Additional data file 1). Penk, coding for proenkephalin1 [67, 68], was the most highly over-expressed probeset under this functional category. Interestingly, this locus was also highly over-expressed in the same N-ras-/- fibroblasts subjected to starvation alone (Table S1 b in Additional data file 1) or to starvation and subsequent short-term, 1-hour serum stimulation (Table S5 in Additional data file 1). Compared to its transcriptional profile during G0/G1 transition, the N-ras-/- cells stimulated with serum for 8 hours shared similar repression of Pi3Kr2 and over-expression of a smaller number of different kinases. Over-expression of GAPs and repression of GEFs, as well as induction or repression of specific ras-related loci, was also observed in this case (Table S8 in Additional data file 1). Regarding cell development and differentiation, Mpg (matrix G1a protein) and Crabp2 (retinoic acid binding protein) showed the highest levels of over-expression under these conditions of serum stimulation. As with Penk, Crabp2 was already highly over-expressed in the same cells subjected to starvation alone (Table S1b in Additional data file 1). Finally, the group of differentially expressed genes listed under cell adhesion and migration showed great increases in the level of expression of specific matrix metallopeptidases or gap junction membrane channel proteins, suggesting specific functional effects on cell-extracellular matrix or cell-cell interactions in fibroblasts of this particular genotype (Table S8 in Additional data file 1).
Differential gene expression in double knockout H-ras-/-/N-ras-/- fibroblasts stimulated with serum for 1 hour (Table S6 in Additional data file 1) involved a significant percentage of genes related to signaling, metabolism and transcription. There was a specific quantitative increase in the functional categories of signal transduction and cell cycle/DNA replication when compared to the other knockout genotypes analyzed (Figure 4). In these double H-ras-/-/N-ras-/- knockout cells, the percentage of differentially expressed genes functionally assigned to signal transduction was higher during G0/G1 transition than during G1 progression (Figure 4). At both stages of the cell cycle we observed increased expression of a number of kinases, small GTPases and other G proteins as well as repression of PI3K subunits (Pik3r2, Pik3ca) (Tables S6 and S9 in Additional data file 1), a pattern consistent with that previously described in the single knockout H-ras-/- or N-ras-/- cells (Tables S4 and S5 in Additional data file 1)
The specific transcriptional profile of fibroblasts lacking both H-Ras and N-Ras during G1 progression (8 hours with serum; Table S9 in Additional data file 1) also showed significant involvement of signaling, transcription or cell metabolism. A specific, visible increase in the categories of cell cycle/DNA replication, RNA processing and ubiquitin cycle was also observed in this case (Figure 4).
In general, the percentage profile of functional categories associated with the absence of both H-Ras and N-Ras in fibroblasts paralleled for the most part that of the same functional categories in one or both of the individual H-ras-/- or N-ras-/-knockout genotypes. For example, the H-ras-/-/N-ras-/- fibroblasts behaved like H-ras-/- cells with regard to development and differentiation or like N-ras-/- cells with regard to growth and proliferation after 1 hour of serum stimulation. Likewise, a similar percentage distribution was detected for functional categories such as RNA metabolism or ubiquitin cycle between H-ras-/-/N-ras-/- and H-ras-/-fibroblasts stimulated with serum for 8 hours (Figure 4). A contrasting exception to that behavior was seen with the category of cell cycle/DNA replication, which clearly showed an additive behavior in comparison to the individual H-ras-/- and N-ras-/- knockout cells (Figure 4).
Functional verification of microarray-based expression data
Various alternative experimental approaches were used to validate the transcriptional data generated with microarrays. Quantitative real time PCR of a randomly selected collection of the differentially expressed genes listed in Tables S4 to S9 in Additional data file 1 was first carried out with microfluidic cards using the signal of the18S ribosomal subunit as control. Confirmation by this technique of the transcriptional trends previously detected with microarrays is indicated by the asterisks in the R.fold column of Tables S4 to S9. In general, a good qualitative agreement was observed between the microarray-derived data and the quantitative real time PCR results, although some quantitative differences were sometimes observed. Additional validation of the microarray-based transcriptional data was obtained in other cases by means of western immunoblots of cellular extracts of the same ras knockout fibroblast lines analyzed with microarrays after serum stimulation. This approach also confirmed the over-expression or the repression of the protein products of a series of differentially expressed genes, as indicated by the hash signs in the R.fold columns of the pertinent tables.
We also explored the possibility of functional links between the above described alterations of gene expression and potential defects in signal transduction. Analysis with protein microarrays of the status of a number of known components of Ras effector signaling pathways showed in N-ras-/- knockout cells a significant decrease in extracellular signal-regulated kinase (ERK) phosphorylation (T202/Y204 residues) occurring after both starvation or short-term serum stimulation (1 hour), suggesting a specific deficiency in ERK-related signaling under those conditions (Figure 5c). Regarding the H-ras-/- fibroblasts, our data suggested a specific deregulation in Ras-PI3K pathways as we consistently detected a significant increase of phosphorylated AKT (S473 residue) in these cells under both starvation and/or serum stimulation, as well as increased PTEN levels after stimulation with serum for 8 hours (Figure 5c).
N-Ras regulation of Stat1 expression and activity through the Ras-ERK signaling pathway
Enhanced apoptosis in N-ras-/- and H-ras-/-N-ras-/-fibroblasts involves intrinsic and extrinsic pathway components
N-Ras is a direct regulator of Bax and Perp expression
Various experimental approaches, including studies of over-expression, subcellular location/processing, genomic disruption and genomic/proteomic profiling support the notion that the mammalian H-Ras, N-Ras and K-Ras isoforms play non-overlapping, differentiated functional roles [1, 3, 6]. For example, our recent characterization of the transcriptomic profile of actively growing fibroblasts lacking H-Ras and/or N-Ras provided significant evidence for the functional involvement of N-Ras in cellular responses related to immunomodulation/host defense and apoptosis . Other reports indicate also that the mammalian Ras proteins play essential functional roles in regulation of the cell cycle [1, 3, 5, 36]. This is based on the observation that microinjection of non-specific, neutralizing Ras antibodies has demonstrated an absolute requirement for Ras activity at several points during serum stimulation of quiescent cells [36–41]. However, little is known about the exact mechanisms mediating the participation of Ras proteins in cell cycle progression or about the possibility that different Ras isoforms play differential functional contributions in this process.
The present study, focused on the joint analysis of the genomic expression profiles of WT and ras knockout (H-ras-/-, N-ras-/-, H-ras-/-/N-ras-/-) fibroblasts subjected to serum starvation or to subsequent stimulation with serum for short periods of time, provides a valid experimental system to test whether N-Ras and H-Ras play specific -or redundant - functional roles during the initial stages of the cell cycle, and to analyze potential mechanisms involved. Thus, microarray-based analysis of the transcriptomic profiles of the serum starved, G0-arrested fibroblasts enables the participation of the Ras isoforms in cellular responses to the stress of serum deprivation to be gauged. On the other hand, the study of the transcriptomic profiles of the same set of serum-arrested fibroblast lines after stimulation with serum for 1 hour or 8 hours was instrumental to discern different functional contributions of N-Ras or H-Ras during G0/G1 transition (1 hour) or mid-G1 progression (8 hours).
The meaningful, joint analysis of the complete set of different transcriptional profiles generated in this study involved in most instances the comparison of the profiles of G0-arrested WT cells with those of the other samples and conditions studied here by means of microarray hybridization. Interestingly, the comparison of the gene expression patterns of G0-arrested fibroblasts (after 24 hours of serum starvation) of all different genotypes tested showed negligible differences among the transcriptional profiles of the WT controls and those of the H-ras-/- or N-ras-/- knockout cells (Tables 1 and 2), indicating that H-Ras and N-Ras do not play a highly significant functional role in generating the transcriptional response of cultured fibroblasts to the stress of serum deprivation.
The hybridization data generated here also allowed us to ascertain whether H-Ras and N-Ras had any specific effect on the transcriptional responses of the starved fibroblasts to serum stimulation. In particular, the microarray hybridizations corresponding to fibroblasts incubated with serum for 1 hour were aimed at targeting the specific gene population transcribed immediately after exit of G0 and re-entry into G1 of the cell cycle (G0/G1 transition) [43, 46, 47, 56–58], whereas those corresponding to cells stimulated with serum for 8 hours were geared to characterize the profile of induced/repressed genes occurring in fibroblasts progressing through the early-mid stages of G1 phase in the cell cycle [48, 59–62]. Accordingly, the list of differentially expressed genes resulting from comparing the profile of G0-arrested WT cells with that of the same WT cells after short-term stimulation (1 hour) with serum contained only induced genes that corresponded, for the most part, with the expected population of so-called IE genes (jun, fos, and so on) known to be transcribed in starved G0 fibroblasts shortly after exposure to serum in culture [43, 46, 47, 56–58]. Interestingly, the profiles of H-ras-/-, N-ras-/- and H-ras-/-/N-ras-/- knockout fibroblasts shared high differential expression of many of the IE loci detected in WT cells, suggesting that, in those cases, H-Ras and N-Ras do not have a direct functional contribution to the transcriptional activation of IE loci and that the regulation of these early serum responses is probably mediated through other Ras-independent signaling pathways. On the other hand, a significant number of differentially expressed, primary response genes were also identified in the WT cells that did not score as differentially expressed in the transcriptional profiles of corresponding ras knockout fibroblasts treated under similar conditions, suggesting that in those cases H-Ras or N-Ras may be actively involved in regulation of their expression. The transcriptional profile of WT fibroblasts stimulated with serum for 8 hours was clearly different from that detected during G0/G1 transition (1 hour) and includes a long list of induced and repressed genes encompassing E2F targets that would be expected as a consequence of the process of G1 to S progression, after Rb phosphorylation and subsequent E2F transcriptional activation [48, 59–62]. Interestingly, the transcriptional activation of many differentially expressed loci detected in the WT cells was lost in the ras knockout fibroblasts subjected to the same treatment with serum. Such loss of transcriptional activation was particularly noticeable in the case of the N-ras-/- and H-ras-/-/N-ras-/- knockout cells, suggesting a major functional participation of Ras proteins, particularly N-Ras, in the regulation of transcriptional programs during early G1 progression.
Whereas the absence of H-Ras or N-Ras did not seem to modify the cellular responses to serum deprivation stress, the genomic disruption of H-ras-/- and/or N-ras-/-, individually or in combination, led to very different transcriptional responses to serum stimulation in comparison to the G0-arrested, WT fibroblasts. Our data clearly show that the absence of N-Ras causes the highest quantitative changes in the first wave of transcriptional activation occurring during G0/G1 transition (1-hour serum stimulation), whereas the absence of H-Ras was associated with the largest size of the second wave of transcriptional activation corresponding to mid-G1 progression (8-hour serum stimulation). The preferential association of N-Ras and H-Ras with each of these two distinct transcriptional waves is consistent with previous reports documenting the absolute requirement for Ras activity during different moments of the early G0 to S interval [36–41], and raises the interesting possibility of a preferential functional involvement of N-Ras with the immediate-early cellular responses to serum stimulation and of H-Ras with the cellular responses related to growth and proliferation during mid-G1 progression.
The analysis of functional annotations corresponding to the differentially expressed genes identified in the multi-class comparisons depicted in the Figure 3 dendrograms and the pair-wise comparisons described in Tables S4 to S9 in Additional data file 1 was instrumental for the assignment of specific functional signatures to H-Ras and N-Ras during the two specific stages of the early cell cycle (G0/G1 transition and mid-G1) that were studied here. Thus, consistent with our previous conclusion attributing a preferential functional role to N-Ras in control of the early (G0/G1 transition) transcriptional wave, and to H-Ras in control of the second (mid-G1) transcriptional wave, the branching of the respective dendrograms clearly shows that the transcriptional pattern of N-ras-/- cells was the most distant from that of the WT control during the early G0/G1 transition and, in contrast, that of H-ras-/- fibroblasts clustered farthest away from its WT control in the set of samples corresponding to stimulation with serum for 8 hours, during mid-G1 progression. Computational evaluation (Genecodis) of the functional annotations for the components of the clusters in the dendrograms provided statistically significant evidence linking the absence of N-Ras during G0/G1 transition to induction of loci related to four main categories of cellular functions, including immune defense responses, apoptosis, transcription and MAPK signaling, and to repression of loci functionally related to cell cycle control, cell adhesion and insulin signaling. The same computational analyses also demonstrated the occurrence of a statistically significant link between the absence of H-Ras and induction of genes related to RNA binding/metabolism/processing and ribosomal protein biosynthesis during the second transcriptional wave analyzed in this study (8-hour serum stimulation; mid-G1 progression). These observations during early stages of the cell cycle are clearly consistent with previous observations from our laboratory with actively growing fibroblasts  that pointed to preferential functional roles of H-Ras in growth and proliferation and of N-Ras in transcriptional regulation of apoptosis and immune/defense responses. Our conclusions are further supported by recent reports [75–77] on the contribution of Stat proteins and interferon signaling to oncogenic transformation and human tumor development. All these observations thus reinforce the notion of non-overlapping functional roles for H-Ras and N-Ras in mammalian fibroblast cells.
The global functional analyses were further complemented and reinforced by the study of the functional annotations of the individual genes listed in the pair-wise comparisons summarized in Tables S4 to S9 in Additional data file 1. The identification of individual genes whose transcription was most specifically linked to the absence of either H-Ras or N-Ras was facilitated by excluding from consideration all loci showing similar levels of differential expression (d-value or R.fold parameters in pair-wise comparisons to G0-arrested, WT cells) for both the WT and the ras knockout (H-ras-/-, N-ras-/- or H-ras-/-/N-ras-/-) cells subjected to stimulation with serum for the same time (1 hour or 8 hours). Confirming the previous global analysis, the list of differentially expressed genes in H-ras-/- fibroblasts subjected to serum stimulation included many different loci that were functionally related to development, growth and proliferation. Particularly striking in this regard was the elevated number of genes coding for tRNA synthetases and ribosomal proteins in both the single H-ras-/- and double H-ras-/-/N-ras-/- knockout cells, but not in N-ras-/- cells, suggesting a specific, direct link between H-Ras and these types of cellular functions related to growth processes. The transcriptional profile of N-Ras-deficient cells displayed many individual genes falling under the functional categories of defense and apoptosis (as previously noted), as well as cell adhesion, motility and signal transduction processes. Regarding this latter category, it was remarkable to observe in serum-stimulated N-ras-/- cells a significant reduction in expression level of components of PI3K signaling pathways, in particular the p85 and p110 subunits of this enzyme, suggesting a significant contribution of N-Ras to cellular signaling through this pathway. All in all, these observations are consistent with the suggestion of a significant functional contribution of N-Ras to the first wave of transcriptional activation associated with G0/G1 re-entry into the cell cycle. Finally, the profile of functional categories affected in the double H-ras-/-/N-ras-/- knockouts reflected, in general, the individual profiles exhibited by the individual H-ras-/-or N-ras-/- genotypes, with a notable exception in the category of cell cycle/DNA replication, where the behavior of the double knockout fibroblasts was additive in relation to the individual knockout genotypes, suggesting that H-Ras and N-Ras complement each other functionally with regards to cellular functions affecting cell cycle progression. In any event, the validation of any proposed functional link resulting from the analysis of transcriptional profiles requires further direct confirmation by means of specific, in vivo functional assays.
Various experimental approaches, including reverse phase protein arrays and direct functional assays of knockout fibroblasts of the specific genotypes under study provided direct support for some of the functional roles attributed to N-Ras or H-Ras on the basis of the transcriptional profiles of pertinent knockout cells, and also offered specific hints on the possible mechanisms involved. For example, with regards to cellular defense processes, our results demonstrated the specific increase of Stat1 expression and phosphorylation in N-Ras-deficient cells and provided direct evidence for the participation of Ras-ERK signaling pathways to mediate the transcriptional regulation of Stat1 by N-Ras. Our data also documented the enhanced apoptotic responses associated with the absence of N-Ras in fibroblasts and provided evidence for the participation of both intrinsic and extrinsic pathways in a process involving direct transcriptional and post-transcriptional regulation by N-Ras of major components, such as Bax and Perp, through ERK- and p38-mediated pathways.
We have shown that the transcriptional profiles of G0-arrested, serum-starved WT and ras knockout fibroblasts (H-ras-/-, N-ras-/-, H-ras-/-/N-ras-/-) are very similar, indicating that these Ras proteins do not play highly important roles in regulation of transcriptional responses to the stress of serum deprivation. In sharp contrast, the transcriptional profiles of knockout fibroblasts lacking H-Ras and/or N-Ras are very different from those of their WT controls after serum stimulation for 1 hour (G0/G1 transition) or 8 hours (mid G1 progression), indicating that H-Ras and N-Ras exert distinct, specific cellular functions during the initial stages of the cell cycle. Whereas all three different ras knockout strains exhibited important transcriptional alterations during both stages of the cell cycle, the absence of N-Ras was quantitatively more disruptive for the first transcriptional wave linked to G0/G1 transition, and the absence of H-Ras affected more potently the transcriptional wave linked to G1 progression. Furthermore, the transcriptional changes of H-Ras-deficient cells showed preferential involvement of loci functionally related to growth and proliferation whereas those of N-Ras-deficient cells were more frequently concerned with development, cell cycle regulation, immunomodulation and apoptosis. Functional analysis indicates that N-Ras contributions to cellular immunity/defense responses is mediated, at least in part, through ERK-dependent regulation of Stat1 expression and activity, whereas its participation in apoptotic responses involves transcriptional regulation of various genes (Bax and Perp) via ERK and p38 signaling pathways.
Our data documenting the occurrence of specific transcriptional profiles associated with the absence of H-Ras and/or N-Ras during early cell cycle stages are consistent with previous reports showing absolute requirements for different peaks of Ras activity during the initial stages of the cell cycle and confirm the notion of functional specificity for the H-Ras and N-Ras isoform proteins.
Materials and methods
Cell lines from the appropriate ras genotype were harvested on Dulbecco's modified Eagle's medium (DMEM; Gibco Paisley, UK) supplemented with FBS (10%; Hyclone, Logan, Utah, USA), glutamine (2 mM), penicillin (100 U/ml) and streptomycin (100 mg/ml). Cultures were grown in a humidified CO2 (5%) atmosphere at 37°C and when subconfluent cells were starved for 24 hours. After starvation cells were either used for RNA/protein isolation, or induced for 1 hour or 8 hours with 20% FBS and then RNA/protein isolation was carried out.
When using the pharmacological inhibitors PD098059 (37 μM), SB203580 (10 μM), LY294002 (20 μM), Genistein (100 μM), and PD153035 (10 μM), WT fibroblasts were cultured as usual and when 70 to 80% confluence was reached they were treated for 24 to 48 hours in the presence of the inhibitor and then collected for protein extraction. All the inhibitors were purchased from Calbiochem® (Darmstadt, Germany).
RNA isolation, cDNA synthesis and microarray hybridization
For each cell line and time point under study RNA was purified from two 10-cm culture dishes per cell line using a commercial kit (RNeasy, Qiagen, Hilden, Germany). Concentration was measured at 260 nm (Ultrospec 2000, Pharmacia Biotech Buckinghamshire, UK) and purity and quality was determined using RNA 6000 Nanochips (Agilent Technologies, Santa Clara, CA, USA). RNA was then used to synthesize cRNA probes for hybridization to Affymetrix MGU74Av2 GeneChip high-density oligonucleotide microarrays. Microarray hybridization was carried out as described in the Gene Expression Analysis Technical Manual provided by Affymetrix .
Microarray hybridization data analysis: normalization, differential gene expression and clustering
Pre-confluent cultures of at least two separate cell lines belonging to each of the ras-related genotype(s) under study (WT, H-ras-/- and N-ras-/- and H-ras-/-/N-ras-/-) were harvested and their RNA extracted for subsequent analysis using Affymetrix high density oligonucleotide microarrays MGU74Av2. At least three independent microarray hybridizations were performed with RNA corresponding to each of the null mutant ras genotypes in the experimental conditions under study. Thus, this study encompassed a total of 3 different data sets (starved cells, cells stimulated with serum for 1 hour and cells stimulated with serum for 8 hours), each consisting of 13 separate chip microarray hybridizations (4 for controls and 3 for each of the three null mutant genotypes). All array hybridization data are available at the NCBI, Gene Expression Omnibus database [GEO:GSE14829] .
Data analysis was carried out using the robust multi-array average and SAM algorithms as previously described . Changes in probeset expression level in knockout cell lines compared to their WT counterparts were identified as significant using a FDR cutoff value of 0.09. Following identification of the differentially expressed probesets, the corresponding matrix of expression values for all microarray hybridizations performed were analyzed using the hclust clustering algorithm implemented in R . This algorithm performs hierarchical cluster analysis with complete linkage to find similarity between probesets based on their expression values in the different chip microarrays analyzed. The algorithm classifies the probesets in correlated groups presenting similar expression profiles or expression signatures. The statistical significance of functional Gene Ontology annotations was estimated by means of P-values of confidence calculated by running Fisher's exact test to compare the number of genes assigned to the various functional categories within each cluster of the dendrogram.
Functional analysis of the significant genes obtained for each induced state was done using a functional annotation tool called GeneCodis (Gene Annotation Co-occurrence Discovery) [66, 81]. This tool finds combinations of co-occurrent annotations that are significantly associated with a list of genes under study with respect to a reference list. The significance of the annotations is calculated using a hypergeometric statistical test with FDR P-value correction and using as reference the mouse genome. The annotations were done at the same time to the full Gene Ontology) database  and to the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways database . After the analyses were done with GeneCodis, the redundancy on the list of genes that are assigned to each functional class was depurated by manual curation in order to identify distinct groups of genes that include similar or related biological functions and that can be enclosed in more general cellular processes as presented in Tables 1 and 2.
RNA from mouse embryo fibroblasts subjected to the different experimental conditions under study was used for quantitative PCR validation on low density microarrays, microfluidic cards (Applied Biosystems, Foster City, CA, USA) using the 18 s ribosomal subunit as an internal control. RNA (10 μg; 1 μl final reaction volume) were reverse transcribed using the High Capacity cDNA Archive Kit (Applied Biosystems) as recommended by the supplier. The previously synthesized cDNA (5 μl) was then mixed with 50 μl of the Taqman® Universal PCR Master Mix (Applied Biosystems) and 50 μl of RNAses free water. Samples were loaded into the microfluidic cards containing the lyophilized oligos in each well and then centrifuged at 1,200 rpm for 2 minutes. Cards were sealed using a Low Density Array Sealer (Applied Biosystems) and the PCR reaction was carried out in an ABI PRISM® 7900HT termocycler (Applied Biosystems). Results were analyzed using the software Sequence Detection Systems (SDS) v2.1 (Applied Biosystems).
Western blot analysis of cellular extracts
Protein lysates were obtained and quantified as previously described  Lysates (30 to 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 (10 mM Tris-HCl (pH 8.0), 150 mM NaCl, 0.05% Tween 20) plus 1% bovine serum albumin were incubated, as appropriate, with dilutions of 0.2 mg/ml of commercial antibodies from Santa Cruz Biotechnologies (Santa Cruz, CA, USA) and horseradish peroxidase-conjugated (Amersham Bioscience, Buckinghamshire, UK) were used as secondary antibodies. Immunoblots were developed using the commercial Enhanced Chemiluminescence (ECL) and ECL plus kits (Amershan Pharmacia Biotech, Piscataway, NJ, USA) following the supplier's recommendations.
Reverse-phase protein lysate array layout and antibody staining
Reverse phase protein microarrays were done as previously described . Origin and dilution of the antibodies used is shown in Table S10 in Additional data file 1. Development of antibody-stained arrays and quantification of the signal data obtained after scanning the arrays were carried out as described [84, 85].
Luciferase reporter assays
Transcriptional activity of control, N-ras-/- and the double H-ras-/-/N-ras-/- cells was assayed using luciferase reporter constructs 8 ISRE-tkLuc (kindly provided by Dr R Pine, The Public Health Research Institute, Newark, NJ, USA)), Bax-pGL3 and PERP-pGL3 (kindly provided by Dr P Lazo, Centro de Investigacion del Cancer, Salamanca, Spain). Cells seeded in six-well plates (5 × 105 cells/well) and cultured for 12 hours were transfected with reporter plasmids (5.0 μg) using JetPEI (Polyplus transfection, Illkirch, France). phRL-tk plasmid (50 ng; Promega, Madison, WI, USA) was co-transfected as an internal control. After further culture for 24 to 36 hours in DMEM with 10% FBS serum, cell extracts were assayed for luciferase activity. Where indicated, cotransfections were done by adding 5.0 μg of a construct containing N-ras (N-ras-pCEFL) or H-ras (H-ras-pCEFL) genes. Luciferase assays were performed using a dual luciferase reporter kit (Promega). Luminescence was determined with a MiniLumat LB9506 luminometer (Berthold, Bad Wildbad, Germany).
Caspase 8 and caspase 9 activity assays
We seeded 5 × 105 cells in six-well plates and once attached they were starved for 24 hours and/or serum stimulated for 1 hour or 8 hours as previously described. After washing twice with cold phosphate-buffered saline cells were lysated with Reporter lysis buffer 1× (Promega), centrifuged for 5 minutes at 12,000 rpm and 4°C and supernatant collected into a new tube. Caspase 8 and 9 activity was measured by adding to the lysates the corresponding reagent (Caspase-Glo® 8 or Caspase-Glo® 9, Promega) in a 1:1 ratio. After 1 hour incubation at room temperature caspase 8 and caspase 9 activity was determined using a MiniLumat LB506 luminometer (Berthold).
Additional data files
The following additional data are available with the online version of this paper: Tables S1 to S10 listing the differential expression detected in WT and knockout fibroblasts of the indicated genotypes that were cultured under conditions of serum starvation or stimulation, as specified in each case (Additional data file 1).
Dulbecco's modified Eagle's medium
extracellular signal-regulated kinase
fetal bovine serum
false discovery rate
GTPase activating protein
guanosine nucleotide exchange factor
interferon-stimulated response element
Significance Analysis of Microarrays
signal transducer and activator of transcription
We thank Dr R Pine (The Public Health Research Institute, Newark, NJ) for ISRE reporter constructs, Dr P Crespo (Instituto de Investigaciones Biomédicas, Santander, Spain) for H-Ras and N-Ras reporter constructs, Dr P Lazo (Centro de Investigación del Cancer, Salamanca, Spain) for PERP reporter construct and E Petricoin (George Mason University, VA) for support with protein array layout generation and analysis. This work was supported in part by grants from Junta Castilla y León (SA044A08 and GR93) and from Instituto de Salud Carlos III (FIS PI021570), as well as institutional support from RTICC (RD06/0020/000), and Acción Transversal en Cáncer 2008 from ISCIII, MSC, Spain. CG was supported by Ramon y Cajal Programa and EC was supported by MEC and FICUS (Fundación de Investigación del Cáncer Universidad de Salamanca) fellowships.
- Abankwa D, Gorfe AA, Hancock JF: Mechanisms of Ras membrane organization and signalling: Ras on a rocker. Cell Cycle. 2008, 7: 2667-2673.PubMedPubMed CentralView ArticleGoogle Scholar
- Colicelli J: Human RAS superfamily proteins and related GTPases. Sci STKE. 2004, 2004: RE13-10.1126/stke.2502004re13.PubMedPubMed CentralGoogle Scholar
- Karnoub AE, Weinberg RA: Ras oncogenes: split personalities. Nat Rev Mol Cell Biol. 2008, 9: 517-531. 10.1038/nrm2438.PubMedPubMed CentralView ArticleGoogle Scholar
- Rojas JM, Santos E: ras genes and human cancer: different implications and different roles. Curr Genom. 2002, 3: 295-311. 10.2174/1389202023350381.View ArticleGoogle Scholar
- Schubbert S, Shannon K, Bollag G: Hyperactive Ras in developmental disorders and cancer. Nat Rev Cancer. 2007, 7: 295-308. 10.1038/nrc2109.PubMedView ArticleGoogle Scholar
- Omerovic J, Hammond DE, Clague MJ, Prior IA: Ras isoform abundance and signalling in human cancer cell lines. Oncogene. 2008, 27: 2754-2762. 10.1038/sj.onc.1210925.PubMedPubMed CentralView ArticleGoogle Scholar
- Maher J, Baker DA, Manning M, Dibb NJ, Roberts IA: Evidence for cell-specific differences in transformation by N-, H- and K-ras. Oncogene. 1995, 11: 1639-1647.PubMedGoogle Scholar
- Oliva JL, Zarich N, Martinez N, Jorge R, Castrillo A, Azanedo M, Garcia-Vargas S, Gutierrez-Eisman S, Juarranz A, Bosca L, Gutkind JS, Rojas JM: 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. 2004, 279: 33480-33491. 10.1074/jbc.M404058200.PubMedView ArticleGoogle Scholar
- Bollag G, McCormick F: Differential regulation of rasGAP and neurofibromatosis gene product activities. Nature. 1991, 351: 576-579. 10.1038/351576a0.PubMedView ArticleGoogle Scholar
- Clyde-Smith J, Silins G, Gartside M, Grimmond S, Etheridge M, Apolloni A, Hayward N, Hancock JF: Characterization of RasGRP2, a plasma membrane-targeted, dual specificity Ras/Rap exchange factor. J Biol Chem. 2000, 275: 32260-32267. 10.1074/jbc.M006087200.PubMedView ArticleGoogle Scholar
- Jones MK, Jackson JH: Ras-GRF activates Ha-Ras, but not N-Ras or K-Ras 4B, protein in vivo . J Biol Chem. 1998, 273: 1782-1787. 10.1074/jbc.273.3.1782.PubMedView ArticleGoogle Scholar
- Hancock JF: Ras proteins: different signals from different locations. Nat Rev Mol Cell Biol. 2003, 4: 373-384. 10.1038/nrm1105.PubMedView ArticleGoogle Scholar
- Plowman SJ, Hancock JF: Ras signaling from plasma membrane and endomembrane microdomains. Biochim Biophys Acta. 2005, 1746: 274-283. 10.1016/j.bbamcr.2005.06.004.PubMedView ArticleGoogle Scholar
- Plowman SJ, Muncke C, Parton RG, Hancock JF: H-ras, K-ras, and inner plasma membrane raft proteins operate in nanoclusters with differential dependence on the actin cytoskeleton. Proc Natl Acad Sci USA. 2005, 102: 15500-15505. 10.1073/pnas.0504114102.PubMedPubMed CentralView ArticleGoogle Scholar
- Walsh AB, Bar-Sagi D: Differential activation of the Rac pathway by Ha-Ras and K-Ras. J Biol Chem. 2001, 276: 15609-15615. 10.1074/jbc.M010573200.PubMedView ArticleGoogle Scholar
- Ehrhardt A, David MD, Ehrhardt GR, Schrader JW: 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. 2004, 24: 6311-6323. 10.1128/MCB.24.14.6311-6323.2004.PubMedPubMed CentralView ArticleGoogle Scholar
- Hamilton M, Wolfman A: Oncogenic Ha-Ras-dependent mitogen-activated protein kinase activity requires signaling through the epidermal growth factor receptor. J Biol Chem. 1998, 273: 28155-28162. 10.1074/jbc.273.43.28155.PubMedView ArticleGoogle Scholar
- Matallanas D, Sanz-Moreno V, Arozarena I, Calvo F, Agudo-Ibanez L, Santos E, Berciano MT, Crespo P: 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. 2006, 26: 100-116. 10.1128/MCB.26.1.100-116.2006.PubMedPubMed CentralView ArticleGoogle Scholar
- Prior IA, Muncke C, Parton RG, Hancock JF: Direct visualization of Ras proteins in spatially distinct cell surface microdomains. J Cell Biol. 2003, 160: 165-170. 10.1083/jcb.200209091.PubMedPubMed CentralView ArticleGoogle Scholar
- Rocks O, Peyker A, Bastiaens PI: Spatio-temporal segregation of Ras signals: one ship, three anchors, many harbors. Curr Opin Cell Biol. 2006, 18: 351-357. 10.1016/j.ceb.2006.06.007.PubMedView ArticleGoogle Scholar
- Voice JK, Klemke RL, Le A, Jackson JH: Four human ras homologs differ in their abilities to activate Raf-1, induce transformation, and stimulate cell motility. J Biol Chem. 1999, 274: 17164-17170. 10.1074/jbc.274.24.17164.PubMedView ArticleGoogle Scholar
- Johnson L, Greenbaum D, Cichowski K, Mercer K, Murphy E, Schmitt E, Bronson RT, Umanoff H, Edelmann W, Kucherlapati R, Jacks T: K-ras is an essential gene in the mouse with partial functional overlap with N-ras. Genes Dev. 1997, 11: 2468-2481. 10.1101/gad.11.19.2468.PubMedPubMed CentralView ArticleGoogle Scholar
- Koera K, Nakamura K, Nakao K, Miyoshi J, Toyoshima K, Hatta T, Otani H, Aiba A, Katsuki M: K-ras is essential for the development of the mouse embryo. Oncogene. 1997, 15: 1151-1159. 10.1038/sj.onc.1201284.PubMedView ArticleGoogle Scholar
- Esteban LM, Vicario-Abejon C, Fernandez-Salguero P, Fernandez-Medarde A, Swaminathan N, Yienger K, Lopez E, Malumbres M, McKay R, Ward JM, Pellicer A, Santos E: 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. 2001, 21: 1444-1452. 10.1128/MCB.21.5.1444-1452.2001.PubMedPubMed CentralView ArticleGoogle Scholar
- Plowman SJ, Williamson DJ, O'Sullivan MJ, Doig J, Ritchie AM, Harrison DJ, Melton DW, Arends MJ, Hooper ML, Patek CE: While K-ras is essential for mouse development, expression of the K-ras 4A splice variant is dispensable. Mol Cell Biol. 2003, 23: 9245-9250. 10.1128/MCB.23.24.9245-9250.2003.PubMedPubMed CentralView ArticleGoogle Scholar
- Umanoff H, Edelmann W, Pellicer A, Kucherlapati R: The murine N-ras gene is not essential for growth and development. Proc Natl Acad Sci USA. 1995, 92: 1709-1713. 10.1073/pnas.92.5.1709.PubMedPubMed CentralView ArticleGoogle Scholar
- Potenza N, Vecchione C, Notte A, De Rienzo A, Rosica A, Bauer L, Affuso A, De Felice M, Russo T, Poulet R, Cifelli G, De Vita G, Lembo G, Di Lauro R: Replacement of K-Ras with H-Ras supports normal embryonic development despite inducing cardiovascular pathology in adult mice. EMBO Rep. 2005, 6: 432-437. 10.1038/sj.embor.7400397.PubMedPubMed CentralView ArticleGoogle Scholar
- Brem R, Certa U, Neeb M, Nair AP, Moroni C: Global analysis of differential gene expression after transformation with the v-H-ras oncogene in a murine tumor model. Oncogene. 2001, 20: 2854-2858. 10.1038/sj.onc.1204403.PubMedView ArticleGoogle Scholar
- Croonquist PA, Linden MA, Zhao F, Van Ness BG: 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. 2003, 102: 2581-2592. 10.1182/blood-2003-04-1227.PubMedView ArticleGoogle Scholar
- Kim S, Lee YZ, Kim YS, Bahk YY: A proteomic approach for protein-profiling the oncogenic ras induced transformation (H-, K-, and N-Ras) in NIH/3T3 mouse embryonic fibroblasts. Proteomics. 2008, 8: 3082-3093. 10.1002/pmic.200800106.PubMedView ArticleGoogle Scholar
- Ohnami S, Aoki K, Yoshida K, Hatanaka K, Suzuki K, Sasaki H, Yoshida T: Expression profiles of pancreatic cancer cell lines infected with antisense K-ras-expressing adenoviral vector. Biochem Biophys Res Commun. 2003, 309: 798-803. 10.1016/j.bbrc.2003.08.073.PubMedView ArticleGoogle Scholar
- Sweet-Cordero A, Mukherjee S, Subramanian A, You H, Roix JJ, Ladd-Acosta C, Mesirov J, Golub TR, Jacks T: An oncogenic KRAS2 expression signature identified by cross-species gene-expression analysis. Nat Genet. 2005, 37: 48-55.PubMedGoogle Scholar
- Vasseur S, Malicet C, Calvo EL, Labrie C, Berthezene P, Dagorn JC, Iovanna JL: Gene expression profiling by DNA microarray analysis in mouse embryonic fibroblasts transformed by rasV12 mutated protein and the E1A oncogene. Mol Cancer. 2003, 2: 19-10.1186/1476-4598-2-19.PubMedPubMed CentralView ArticleGoogle Scholar
- Zuber J, Tchernitsa OI, Hinzmann B, Schmitz AC, Grips M, Hellriegel M, Sers C, Rosenthal A, Schafer R: A genome-wide survey of RAS transformation targets. Nat Genet. 2000, 24: 144-152. 10.1038/72799.PubMedView ArticleGoogle Scholar
- Castellano E, De Las Rivas J, Guerrero C, Santos E: 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. 2007, 26: 917-933. 10.1038/sj.onc.1209845.PubMedView ArticleGoogle Scholar
- Pruitt K, Der CJ: Ras and Rho regulation of the cell cycle and oncogenesis. Cancer Lett. 2001, 171: 1-10. 10.1016/S0304-3835(01)00528-6.PubMedView ArticleGoogle Scholar
- Gille H, Downward J: Multiple ras effector pathways contribute to G(1) cell cycle progression. J Biol Chem. 1999, 274: 22033-22040. 10.1074/jbc.274.31.22033.PubMedView ArticleGoogle Scholar
- Marshall C: How do small GTPase signal transduction pathways regulate cell cycle entry?. Curr Opin Cell Biol. 1999, 11: 732-736. 10.1016/S0955-0674(99)00044-7.PubMedView ArticleGoogle Scholar
- Stacey D, Kazlauskas A: Regulation of Ras signaling by the cell cycle. Curr Opin Genet Dev. 2002, 12: 44-46. 10.1016/S0959-437X(01)00262-3.PubMedView ArticleGoogle Scholar
- Takuwa N, Takuwa Y: Regulation of cell cycle molecules by the Ras effector system. Mol Cell Endocrinol. 2001, 177: 25-33. 10.1016/S0303-7207(01)00439-7.PubMedView ArticleGoogle Scholar
- Taylor SJ, Shalloway D: Cell cycle-dependent activation of Ras. Curr Biol. 1996, 6: 1621-1627. 10.1016/S0960-9822(02)70785-9.PubMedView ArticleGoogle Scholar
- Bar-Sagi D, Feramisco JR: Microinjection of the ras oncogene protein into PC12 cells induces morphological differentiation. Cell. 1985, 42: 841-848. 10.1016/0092-8674(85)90280-6.PubMedView ArticleGoogle Scholar
- Greenberg ME, Hermanowski AL, Ziff EB: Effect of protein synthesis inhibitors on growth factor activation of c-fos, c-myc, and actin gene transcription. Mol Cell Biol. 1986, 6: 1050-1057.PubMedPubMed CentralView ArticleGoogle Scholar
- Jones SM, Kazlauskas A: Growth-factor-dependent mitogenesis requires two distinct phases of signalling. Nat Cell Biol. 2001, 3: 165-172. 10.1038/35055073.PubMedView ArticleGoogle Scholar
- Pardee AB: G1 events and regulation of cell proliferation. Science. 1989, 246: 603-608. 10.1126/science.2683075.PubMedView ArticleGoogle Scholar
- Riabowol KT, Vosatka RJ, Ziff EB, Lamb NJ, Feramisco JR: Microinjection of fos-specific antibodies blocks DNA synthesis in fibroblast cells. Mol Cell Biol. 1988, 8: 1670-1676.PubMedPubMed CentralView ArticleGoogle Scholar
- Siegfried Z, Ziff EB: Transcription activation by serum, PDGF, and TPA through the c-fos DSE: cell type specific requirements for induction. Oncogene. 1989, 4: 3-11.PubMedGoogle Scholar
- Sun A, Bagella L, Tutton S, Romano G, Giordano A: From G0 to S phase: a view of the roles played by the retinoblastoma (Rb) family members in the Rb-E2F pathway. J Cell Biochem. 2007, 102: 1400-1404. 10.1002/jcb.21609.PubMedView ArticleGoogle Scholar
- Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U, Speed TP: Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics. 2003, 4: 249-264. 10.1093/biostatistics/4.2.249.PubMedView ArticleGoogle Scholar
- Tusher VG, Tibshirani R, Chu G: Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA. 2001, 98: 5116-5121. 10.1073/pnas.091062498.PubMedPubMed CentralView ArticleGoogle Scholar
- Collins CA, Watt FM: Dynamic regulation of retinoic acid-binding proteins in developing, adult and neoplastic skin reveals roles for beta-catenin and Notch signalling. Dev Biol. 2008, 324: 55-67. 10.1016/j.ydbio.2008.08.034.PubMedView ArticleGoogle Scholar
- Lane MA, Xu J, Wilen EW, Sylvester R, Derguini F, Gudas LJ: LIF removal increases CRABPI and CRABPII transcripts in embryonic stem cells cultured in retinol or 4-oxoretinol. Mol Cell Endocrinol. 2008, 280: 63-74. 10.1016/j.mce.2007.09.010.PubMedPubMed CentralView ArticleGoogle Scholar
- Fambrough D, McClure K, Kazlauskas A, Lander ES: Diverse signaling pathways activated by growth factor receptors induce broadly overlapping, rather than independent, sets of genes. Cell. 1999, 97: 727-741. 10.1016/S0092-8674(00)80785-0.PubMedView ArticleGoogle Scholar
- Herschman HR: Primary response genes induced by growth factors and tumor promoters. Annu Rev Biochem. 1991, 60: 281-319. 10.1146/annurev.bi.60.070191.001433.PubMedView ArticleGoogle Scholar
- Rollins BJ, Stiles CD: Serum-inducible genes. Adv Cancer Res. 1989, 53: 1-32. full_text.PubMedView ArticleGoogle Scholar
- Hopewell R, Li L, MacGregor D, Nerlov C, Ziff EB: Regulation of cell proliferation and differentiation by Myc. J Cell Sci Suppl. 1995, 19: 85-89.PubMedView ArticleGoogle Scholar
- Leonard DG, Ziff EB, Greene LA: Identification and characterization of mRNAs regulated by nerve growth factor in PC12 cells. Mol Cell Biol. 1987, 7: 3156-3167.PubMedPubMed CentralView ArticleGoogle Scholar
- Siegfried Z, Ziff EB: Altered transcriptional activity of c-fos promoter plasmids in v-raf-transformed NIH 3T3 cells. Mol Cell Biol. 1990, 10: 6073-6078.PubMedPubMed CentralView ArticleGoogle Scholar
- Blais A, Dynlacht BD: Hitting their targets: an emerging picture of E2F and cell cycle control. Curr Opin Genet Dev. 2004, 14: 527-532. 10.1016/j.gde.2004.07.003.PubMedView ArticleGoogle Scholar
- Ma Y, Croxton R, Moorer RL, Cress WD: Identification of novel E2F1-regulated genes by microarray. Arch Biochem Biophys. 2002, 399: 212-224. 10.1006/abbi.2002.2761.PubMedView ArticleGoogle Scholar
- Polager S, Kalma Y, Berkovich E, Ginsberg D: E2Fs up-regulate expression of genes involved in DNA replication, DNA repair and mitosis. Oncogene. 2002, 21: 437-446. 10.1038/sj.onc.1205102.PubMedView ArticleGoogle Scholar
- Wells J, Graveel CR, Bartley SM, Madore SJ, Farnham PJ: The identification of E2F1-specific target genes. Proc Natl Acad Sci USA. 2002, 99: 3890-3895. 10.1073/pnas.062047499.PubMedPubMed CentralView ArticleGoogle Scholar
- Dobrowolski S, Harter M, Stacey DW: Cellular ras activity is required for passage through multiple points of the G0/G1 phase in BALB/c 3T3 cells. Mol Cell Biol. 1994, 14: 5441-5449.PubMedPubMed CentralView ArticleGoogle Scholar
- Feig LA, Cooper GM: Inhibition of NIH 3T3 cell proliferation by a mutant ras protein with preferential affinity for GDP. Mol Cell Biol. 1988, 8: 3235-3243.PubMedPubMed CentralView ArticleGoogle Scholar
- Mulcahy LS, Smith MR, Stacey DW: Requirement for ras proto-oncogene function during serum-stimulated growth of NIH 3T3 cells. Nature. 1985, 313: 241-243. 10.1038/313241a0.PubMedView ArticleGoogle Scholar
- Carmona-Saez P, Chagoyen M, Tirado F, Carazo JM, Pascual-Montano A: GENECODIS: a web-based tool for finding significant concurrent annotations in gene lists. Genome Biol. 2007, 8: R3-10.1186/gb-2007-8-1-r3.PubMedPubMed CentralView ArticleGoogle Scholar
- Boggs DF, Miller JH: Absence of an hypoxic depression of metabolism in preproenkephalin knockout mice. Respir Physiol Neurobiol. 2006, 152: 92-99. 10.1016/j.resp.2005.06.004.PubMedView ArticleGoogle Scholar
- Nitsche JF, Schuller AG, King MA, Zengh M, Pasternak GW, Pintar JE: Genetic dissociation of opiate tolerance and physical dependence in delta-opioid receptor-1 and preproenkephalin knock-out mice. J Neurosci. 2002, 22: 10906-10913.PubMedGoogle Scholar
- Dimberg A, Karlberg I, Nilsson K, Oberg F: Ser727/Tyr701-phosphorylated Stat1 is required for the regulation of c-Myc, cyclins, and p27Kip1 associated with ATRA-induced G0/G1 arrest of U-937 cells. Blood. 2003, 102: 254-261. 10.1182/blood-2002-10-3149.PubMedView ArticleGoogle Scholar
- Pilz A, Ramsauer K, Heidari H, Leitges M, Kovarik P, Decker T: Phosphorylation of the Stat1 transactivating domain is required for the response to type I interferons. EMBO Rep. 2003, 4: 368-373. 10.1038/sj.embor.embor802.PubMedPubMed CentralView ArticleGoogle Scholar
- Shuai K, Schindler C, Prezioso VR, Darnell JE: Activation of transcription by IFN-gamma: tyrosine phosphorylation of a 91-kD DNA binding protein. Science. 1992, 258: 1808-1812. 10.1126/science.1281555.PubMedView ArticleGoogle Scholar
- Hengartner MO: The biochemistry of apoptosis. Nature. 2000, 407: 770-776. 10.1038/35037710.PubMedView ArticleGoogle Scholar
- Wajant H: Death receptors. Essays Biochem. 2003, 39: 53-71.PubMedView ArticleGoogle Scholar
- Parone PA, James D, Martinou JC: Mitochondria: regulating the inevitable. Biochimie. 2002, 84: 105-111. 10.1016/S0300-9084(02)01380-9.PubMedView ArticleGoogle Scholar
- Critchley-Thorne RJ, Simons DL, Yan N, Miyahira AK, Dirbas FM, Johnson DL, Swetter SM, Carlson RW, Fisher GA, Koong A, Holmes S, Lee PP: Impaired interferon signaling is a common immune defect in human cancer. Proc Natl Acad Sci USA. 2009, 106: 9010-9015. 10.1073/pnas.0901329106.PubMedPubMed CentralView ArticleGoogle Scholar
- Wang S, Koromilas AE: Stat1 is an inhibitor of Ras-MAPK signaling and Rho small GTPase expression with implications in the transcriptional signature of Ras transformed cells. Cell Cycle. 2009, 8: 2070-2079.PubMedView ArticleGoogle Scholar
- Gough DJ, Corlett A, Schlessinger K, Wegrzyn J, Larner AC, Levy DE: Mitochondrial STAT3 supports Ras-dependent oncogenic transformation. Science. 2009, 324: 1713-1716. 10.1126/science.1171721.PubMedPubMed CentralView ArticleGoogle Scholar
- Affymetrix. [http://affymetrix.com/index.affx]
- NCBI Gene Expresión Omnibus (GEO) Accesion Display. [http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE14829]
- Murtagh F: Multidimensional clustering algorithms. COMPSTAT Lectures 4. 1985, Wuerzburg: Physica-Verlag, 131-Google Scholar
- Genecodis 2.0, UCM access, Madrid. [http://genecodis.dacya.ucm.es/]
- The Gene Ontology Database. [http://www.geneontology.org/]
- KEGG Pathway Database. [http://www.genome.ad.jp/kegg/pathway.html]
- Herrmann PC, Gillespie JW, Charboneau L, Bichsel VE, Paweletz CP, Calvert VS, Kohn EC, Emmert-Buck MR, Liotta LA, Petricoin EF: Mitochondrial proteome: altered cytochrome c oxidase subunit levels in prostate cancer. Proteomics. 2003, 3: 1801-1810. 10.1002/pmic.200300461.PubMedView ArticleGoogle Scholar
- Wulfkuhle JD, Aquino JA, Calvert VS, Fishman DA, Coukos G, Liotta LA, Petricoin EF: Signal pathway profiling of ovarian cancer from human tissue specimens using reverse-phase protein microarrays. Proteomics. 2003, 3: 2085-2090. 10.1002/pmic.200300591.PubMedView ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.