Spliceosomal protein E regulates neoplastic cell growth by modulating expression of cyclin E/CDK2 and G2/M checkpoint proteins.

Small nuclear ribonucleoproteins are essential splicing factors. We previously identified the spliceosomal protein E (SmE) as a downstream effector of E2F1 in p53-deficient human carcinoma cells. Here, we investigated the biological relevance of SmE in determining the fate of cancer and non-tumourigenic cells. Adenovirus-mediated expression of SmE selectively reduces growth of cancerous cells due to decreased cell proliferation but not apoptosis. A similar growth inhibitory effect for SmD1 suggests that this is a general function of Sm-family members. Deletion of Sm-motifs reveals the importance of the Sm-1 domain for growth suppression. Consistently, SmE overexpression leads to inhibition of DNA synthesis and G2 arrest as shown by BrdU-incorporation and MPM2-staining. Real-time RT-PCR and immunoblotting showed that growth arrest by SmE directly correlates with the reduction of cyclin E, CDK2, CDC25C and CDC2 expression, and up-regulation of p27Kip. Importantly, SmE activity was not associated with enhanced expression of other spliceosome components such as U1 SnRNP70, suggesting that the growth inhibitory effect of SmE is distinct from its pre-mRNA splicing function. Furthermore, specific inactivation of SmE by shRNA significantly increased the percentage of cells in S phase, whereas the amount of G2/M arrested cells was reduced. Our data provide evidence that Sm proteins function as suppressors of tumour cell growth and may have major implications as cancer therapeutics.

Introduction

E2F transcription factors are key components in the cell cycle regulatory machinery. They control cell proliferation by regulating the timely expression of many genes required for cell cycle progression, particularly those involved in transition from G1 into S phase. Several regulators of the cell cycle, such as CDC2, cyclin E, cyclin A and p27Kip contain E2F binding sites in their promoters [1, 2]. Novel findings indicated that E2F integrates cell cycle progression with DNA repair, replication, and G2/M checkpoints through co-ordinate regulation of genes essential for both DNA synthesis as well as cell surveillance [3]. It has been shown that E2F regulates expression of several genes required for DNA repair, chromatin assembly, condensation and segregation, as well as multiple checkpoints that ensure genomic integrity. For example, E2F mediates sustained G2 arrest through repression of two mitotic genes, stathmin and Aurora- and Ip11- like midbody-associated protein 1 (AIM-1), in response to genotoxic stress [4].

E2F's activity is controlled in many ways including interaction with Rb. In quiescent cells, E2F is inhibited through association with hypophosphorylated Rb and its pocket protein family members p107 and p130. During cell cycle progression, D-type cyclin associated kinases initiate phosphorylation of Rb family members, which results in the release of E2F and transactivation of E2F-regulated genes [5, 6]. However, the role of E2F in determining cell fate is not restricted to its effect on cell cycle progression, since recent findings demonstrated that E2F affects additional processes including differentiation, development and responses to DNA damage [7]. At least one member of the E2F family, E2F1, can also efficiently induce cells to undergo apoptosis. Previously, the E2F1-marked box domain was reported as the unique pro-apoptotic activity that distinguishes E2F1 from other E2Fs. E2F1-induced apoptosis occurs via both p53-dependent and p53-independent pathways [8]. In response to DNA damage E2F1 is phosphorylated by ataxia-telangiectasia mutated (ATM) and Chk2 and thereby stabilized [9-11].

Interestingly, we have identified the small nuclear ribonucleoprotein E (snRNP E), also known as spliceosomal protein E (SmE), as a novel target of E2F1 by a genetic so called technical knockout (TKO) approach [12]. The 11 kD basic SmE belongs to a large family of polypeptides containing Sm and Sm-like (Lsm) proteins, which are conserved in eukaryotes and in archaebacteria [13, 14]. Sequence comparison of the so far known seven Sm family members (B/B′, D1, D2, D3, E, F and G) from a range of species revealed a highly conserved Sm core protein motif [15]. This motif is composed of two blocks of amino acids, the Sm-1 and Sm-2 motif, responsible for the assembly of U snRNAs (U1, U2, U4/U6 and U5) in an ordered manner to form the Sm core of the splicesomal snRNPs [14], thereby involved in RNA processing and mRNA degradation [16]. Generally, mRNA processing factors are traditionally thought to function only in the control of global gene expression and are involved in essential pre-mRNA splicing. However, an increasing number of evidence demonstrated that Sm and Sm-like proteins also contribute to other physiological activities independent of its concanical RNA processing tasks [17-19]. In fact, it has been shown that Sm proteins control germ granule localization during early embryogenesis of C. elegans, and SmE and SmG are required to maintain transcriptional quiescence in embryonic germ cell precursors through inhibition of RNA poly-merase II activity [18]. In addition, Sm proteins were also detected as components of nuage and mitochondrial cement in Xenopus oocytes [19]. Furthermore, the Sm-like Hfq protein in E.coli. binds numerous small RNA to regulate the stability or translation of specific mRNA [20], suggesting that Sm proteins are implicated in multiple biological functions. In addition, it is known that human Sm proteins cross-react with antisera from patients suffering from the autoimmune disorder systemic lupus erythematosus (SLE), which is characterized by the generation of autoantibodies specific for a variety of self-antigens, many of which are nuclear in origin, including DNA, histones and Sm proteins [21]. In SLE patients, as much as 20% of the entire Ig repertoire may bind Sm protein These anti-Sm autoantibodies individually target a number of components of the splicesome, namely SmB/B′, D1, D2, D3, E, F and G. However, the relevance of Sm proteins and the presence of the autoantibodies in the aetiology and development of autoimmune disorders are still unclear.

Here, we investigated the role of SmE in determining the fate of p53-deficient and p53-proficient human tumour versus non-tumourigenic cells. Apart from its pre-mRNA splicing function, enhanced expression of SmE-reduced cell growth in neoplas-tic cells and lead to the inhibition of DNA synthesis and G2 arrest, which correlates with the regulation of cell cycle checkpoint proteins, such as cyclin E, CDC2, CDK2, CDC25C and p27Kip. Our results provide first evidence that Sm proteins play an essential role in the modulation of cell cycle progression in tumour cells without affecting the growth of normal cells.

Materials and methods

Cell culture

Human H1299 lung cancer cells, human colon tumour (HCT) 116 (p53+/+) and HCT 116 (p53−/− colon cancer cells (kindly provided by B. Vogelstein), the human VH6 foreskin and WI-38 lung fibroblast cell line (Promochem, Wesel, Germany) were maintained in Dulbecco's modified eagle medium (DMEM; Invitrogen, Karlsruhe, Germany) supplemented with 10% foetal calf serum (FCS; Biochrom, Berlin, Germany). Medium contained 2 mM L-glutamine, penicillin at 100 μg/ml and streptomycin at 100 U/ml. The non-immortalized VIT1 human primary pancreatic mesenchymal cell line (Chemicon, Temecula, CA, USA) was grown in Pancreatic Cell Culture Medium (SCR016) supplemented with SCR015 (Chemicon, Temecula, CA, USA). Viruses were grown in 293 cells maintained in DMEM with 10% FCS.

Adenoviral vector construction and RNAi

Ad vectors were generated using the AdEasy System. SmE and SmD1 full-length cDNA was amplified by RT-PCR using the following primer sets: SmE 5′-ATGGCGTACCGTGGCCAGGGT-3′ and 5′-CTAGTTGGAGA-CACTTTGTAG-3′, SmD1 5′-ATGAAGCTCGTGAGATTTTTG-3′ and 5′-TTATCGCCTAGGACCCCCTCT-3′. SmEΔ1 mutant cDNA carrying a deletion of the conserved 19 amino acids in the Sm-1 motif (SmE(1) and the SmEΔ2 mutant deleted of 14 amino acids in the Sm-2 motif were generated by overlapping PCRs using two separate primer pairs: set I 5′-ATG-GCGTACCGTGGCCAGGGT-3′ and 5′-TGTTTTAGAATGAATTTCTATCCG-CATATTCACTT-3′, and set II 5′-AATATGCGGATAGAAATTCATTCTAAAA-CAAAGTC-3′ and 5′-CTAGTTGGAGACACTTTGTAG-3′. Amplification of the SmEΔ2 mutant deleted of 14 amino acids in the Sm-2 motif was performed using primers 5′-ATGGCGTACCGTGGCCAGGGT-3′ and 5′-CTATTGTTTTCTTGACTTTGTTTT-3′. Products were ligated into the pcDNA 3.1 V5-His TOPO vector (Invitrogen, Karlsruhe, Germany) and subcloned in the pShuttle plasmid under control of the cytomegalovirus (CMV) promoter using the Hind III and EcoR V restriction sites. Virus was generated by homologous recombination following cotransformation with pAdEasy1 in E. coli BJ5183. Ad vectors expressing short hairpin RNAs (shRNAs) against human SmE (Ad-shSmE) was generated with the GeneSuppressor System (Biocarta, Hamburg, Germany) according to the supplier's protocol. Specific oligonucleotides were designed using online (http://www.imgenex.com) software: The target sequence for SmE is ATAGATCGCGGATTCAGGTGT (bp 80–100). Synthesized oligos (Invitogen, Karlsruhe, Germany) were ligated into pSuppressorAdeno shuttle vector and cotransfected with GeneSupressor backbone plasmid in 293 cells. Ad-shgreen florescence protein (GFP) has been described elsewhere [23]. All viruses were propagated and purified as described [24], and titrated using the Adeno-X Rapid Titer Kit (BD Biosciences Clontech, Heidelberg, Germany).

Growth rate determination, colony formation and XTT assay

Cells were seeded at a density of 5 × 104 cells/dish for 24 hrs before viral infection. Cells were infected with virus at a multiplicity of infection (MOI) of 50 allowing 100% transduction. Triplicate dishes of each treatment were counted at daily intervals over 5 days. Cell viability was determined by trypan blue exclusion. For colony formation assay, infected cells were cultured for 2 weeks. Colonies were stained with 0.025% crystal violet in 20% methanol and counted under light microscopy. For XTT assay, cells seeded on 96-well plates were incubated with TACS™ XTT labelling mixture for 4 hrs (Biozol Diagnostica, Eching, Germany). Conversion of XTT to formazan was quantitated by measuring the absorbance at 450 nm. For blocking glycogen synthase kinase 3β (GSK3β) activity, tumour cells were treated with the GSK3β specific inhibitor 4-Benzyl-2-methyl-1,2,4-thiadiazolidine-3,5-dione (TDZD-8; 10 μM) (Calbiochem, San Diego, CA USA).

For treatment with chemotherapeutic agents, 7 × 106 cells were treated by continuous exposure to 0.2 μg/ml daunorubicin or etoposide at indicated concentrations for 2 days. Drug treatment of transfected cultures was started 24 hrs following transfection.

Flow cytometry, caspase-3 activity assay

Cells were harvested at indicated time points after infection, fixed in 70% ethanol, and stained for DNA content with propidium iodide (PI). Flow cytometric analysis was carried out in a Fluorescence-activated-cell-sorter (FACS) Calibur (BD Bioscience, Heidelberg, Germany) using CellQuest software. Caspase-3 activity was assayed by measuring the cleavage of the chromophore p-nitroanilide (p-NA) from a p-NA-labelled substrate according to the manufacturer's instructions (BD Biosciences, Heidelberg, Germany). Absorbance was measured at 405 nm in a spectrophotometer.

Immunoblotting

For Western blot analysis, cells were lysed in RIPA buffer (50 mM Tris/HCl (pH 7.2), 150 mM NaCl, 1% Triton X-100 (V/V), 1% Na-Deoxycholate (V/V), 0.1% SDS (W/V) and total protein concentration was quantified by a modified Bradford assay (Bio-Rad, München, Germany). Equal amounts of protein were separated by SDS-PAGE and transferred to nitrocellulose membranes (Amersham Biosciences, Freiburg, Germany). Membranes were probed with antibodies against CDK2, Cyclin E, Cyclin B1 [H20], p27Kip[C-19], CDC25C [C-20], CDC25A [144], U1 SnRNP 70 (C-18), Sm D1(C-15) and actin from Santa Cruz Biotechnology (Heidelberg, Germany), or anti-GFP monoclonal antibody 3E6 (Invitrogen, Karlsruhe, Germany). Expression of CDC2 was analysed using the PhosphoPlus™ CDC2(tyr15) Antibody Kit (New England Biolabs, MA, USA). Primary antibodies were detected with appropriate secondary antibody-horseradish peroxidase conjugates (Amersham Biosciences, Freiburg, Germany). Membranes were developed using the ECL system (Amersham Biosciences, Freiburg, Germany).

Real-time RT-PCR

Total RNA was extracted with the RNeasy Mini Kit (Qiagen, Hilden, Germany) and reverse transcribed using Omniscript RT (Qiagen, Hilden, Germany) and Oligo-dT primer. Real-time PCR was performed using TaqMan Universal PCR Master Mix (Applied Biosystems, Darmstadt, Germany) in conjunction with ABI PRISM 7700 HT Sequence Detection Systems as previously described [12]. Assay on demand kits for CDC2 (Hs00364293_m1), CDK2 (Hs00608082_m1), and CDC25C (Hs00156411_m1) were purchased from Applied Biosystems (Darmstadt, Germany). Gene expression profile was achieved using the Comparative CT method of relative quantification.

5-Bromo-2′-deoxyuridine (BrdU) incorporation and MPM2 staining

Cells were seeded at a density of 5 × 105. Cells in G1, M and G2/M phase were analysed using the Cell Cycle Analysis with BrdU incorporation Kit according to the instructions (BD Biosciences, Heidelberg, Germany). Cells were labelled with 10 mM BrdU for 1 hr and fixed in 70% ethanol, followed by DNase I digestion for 1 hr. For detection of M phase cells, cells were incubated with Fluorescein isothiocyanate (FITC)-conjugated mouse monoclonal anti-mPM2 antibody (Upstate, Charlottesville, VA, USA) at a final concentration of 1 μg/ml, FITC-labelled goat antimouse antibody (DakoCytomation AG, Hamburg, Germany) and stained for DNA content with PI.

Statistical analysis

Statistical significance was calculated by paired Student's t-test. All statistical tests employed in this study were two-sided.

Results

Overexpression of SmE leads to attenuated cell proliferation but not apoptosis

We have recently identified the small nuclear ribonucleoprotein E as an effector of E2F1 in p53-deficient cancer cells using the technical TKO method [12]. This is a genetic tool based on the assumption that specific inactivation of growth inhibitory genes conveys growth advantage in a specific restrictive environment, followed by selection of phenotypic changes caused by its inactivation [25]. It allows identification of both cell death genes and mediators of growth inhibitory signals. To clarify the possible contribution of Sm proteins to cell fate, we generated Ad vectors that express SmE and another Sm family member, SmD1, both containing the highly conserved Sm-1 and Sm-2 motifs (Fig.1A), and two truncated mutants of the SmE protein, SmEΔ1 and SmEΔ2, with a deletion of Sm-1 and Sm-2, respectively. The Sm domains were shown to interact with other Sm proteins to form the spliceo-somal complex during pre-mRNA processing [15, 26]. The effect of Sm proteins on cell growth was determined after infection of H1299 cells with Ad-SmD1, Ad-SmE and its deletion mutants by counting the cell number over a period of 5 days. As shown in Figure 1B, growth of Ad-SmE infected cells was significantly suppressed after 5 days with a more than 50% decrease in cell number compared to the Ad-GFP infected control. A similar result was also observed with SmD1, suggesting that the growth inhibitory effect might be a general function of this protein family. In contrast, inhibition of cell growth was substantially compromised in SmEΔ1 expressing cells, whereas Ad-SmEΔ2-infected cells exhibited a similar suppression effect as shown for the full-length SmE protein. The protein expression levels of SmD1 and different SmE constructs in H1299 cells at five days after infection are as indicated (Fig.1B, upper panel). In addition, we analysed the long-term effect of Sm protein expression on cell growth by colony formation assay. After virus infection, H1299 cells were cultured on 6-well plates for 2 weeks. Consistent with the cell growth rates shown in Figure 1B, cells expressing SmD1 (40.4±9.6), SmE (29.5±5.7) and its mutant SmEΔ2 (42±4.6) gave rise to a significantly lower number of colonies than cells infected with Ad-SmEΔ1 (86.5±9.5) or the control virus Ad-GFP (98.3±10.2, Fig.1C). These data indicate that the SmE(1 is essential for SmE-induced cell growth inhibition.

Figure 1

Growth inhibitory effect of SmD1, SmE and its truncated mutants in H1299 cells. (A) Structural homology of human Sm protein family members SmE (RUXE) and SmD1 (SMD1) with highly conserved Sm-1 and Sm-2 motifs indicated. (B) H1299 cells were seeded at a low density of 5 × 105 cells/35-mm dish 24 hrs prior to viral infection with Ad-SmD1 (square), Ad-SmE (triangle), Ad-SmEΔ1 (grey rhombus) and Ad-SmEΔ2 (circle) carrying a deletion of the Sm-1 or Sm-2 motif, respectively, or control vector Ad-GFP (rhombus). At indicated time points after infection, cells were counted and viability determined by trypan blue exclusion Each graph represents the mean ± SD of three independent experiments. The protein expression levels of SmD1 and different SmE constructs from infected cells are shown before and 5 days after viral infection. (upper panel). The expression of GFP served as control. (C) H1299 cells seeded at equal cell numbers in 6-well plates were infected as in (B). After 2 weeks, cells were stained with 0.025% crystal violet and the number of colonies counted.

Several E2F1 effector proteins are known to be potent mediators of cell death [8]. Therefore, we examined the effect of SmE on cell viability of different human cancer cell lines and non-tumourigenic fibroblasts using the tumour suppressor p53 as a positive control. Compared to Ad-GFP virus infected cells, enforced expression of SmE resulted in a significant loss of cell viability (by approximately 30–40%) at 72 hrs after treatment in all tested tumour cell lines independent of their endogenous p53 and/or Rb status (Fig.2A). Importantly, the activity of SmE was restricted to transformed cells and, similar to p53, not observed in karyotypically normal cell lines, such as human VH6 foreskin, WI-38 lung and VIT1 primary pancreatic fibroblasts. Equal SmE protein levels in all cell lines are shown in Figure 2B. A National Center for Biotechnology Information (NCBI) protein database search revealed that other proteins with cell growth regulatory functions, such as the glycogen synthase kinase 3 (GSK3) contain a Sm-1 like motif. To exclude that SmE acts by competing Sm-1 motifs of these proteins rather than by an intrinsic inhibitory activity, the cell growth inhibitory effect of SmE was also measured in Ad-SmE-infected tumour cell lines in the presence of GSK3β specific inhibitor TDZD-8. No difference in cell viability was observed compared to cells overexpressing SmE in the absence of GSK3β inhibitor, suggesting that growth inhibition is a direct activity of the SmE protein (data not shown).

Figure 2

Ectopically expressed SmE leads to loss of cell viability in human cancer cells without affecting normal fibroblasts. (A) H1299 (p53−/−), Saos-2 (p53−/−), HCT 116 (p53−/−), HCT 116 (p53+/+) as well as human VH6, WI-38 and VIT1 fibroblasts were infected with Ad-GFP (black column), Ad-SmE (grey column) or Ad-p53 (white column). Cell viability was measured 72 hrs after infection by XTT assay. The number of viable cells is the mean ± SD of triplicate experiments. Bars show the relative viability after normalization with control vector infected cells. Significant differences between treated and control cells are labelled with asterisks (P < 0.05; paired,two-sided t-test). (B) Western blot analysis of SmE, p53 and GFP protein expression from each cell line at 72 hrs after Ad vector infection.

To investigate whether the observed loss of viability in neo-plastic cells upon overexpression of SmE is due to apoptosis, the sub-G1 DNA content of Ad-SmE infected H1299 cells was analysed by flow cytometry. Quantification of the sub-G1 population revealed no increase in the amount of apoptotic cells even at 96 hrs after infection (Fig.3A). Concomitantly, no caspase-3 activation was evident in cells expressing SmE (Fig.3B), whereas caspase-3 activity strongly increased after infection with Ad-E2F1. These results demonstrate that SmE has no apoptosis inducing properties.

Figure 3

SmE-induced cytotoxicity is not caused by apoptosis induction. (A) H1299 cells infected with Ad-SmE were analysed by flow cytometry at daily intervals over 4 days. The percentage of cells with a sub-G1 DNA content (M1) is as indicated. GFP infected cells were used as control (left panel). Western blot analysis of SmE and GFP protein expression levels in whole cell extracts from Ad-SmE or Ad-GFP infected cells. Extracts were probed with an anti-actin antibody as a loading control (right panel). (B) Caspase-3 activity of Ad-SmE infected cells was measured after 96 hrs in the absence (black column) or presence by the caspase inhibitor zVAD-fmk (50 μM) (grey column). Cells infected with Ad-E2F1 and Ad-GFP were used as positive and negative controls, respectively. The average caspase-3 activity obtained from triplicate experiments is shown. Error bars indicate standard deviations.

Enforced expression of SmE inhibits DNA synthesis and arrests cells in G2 phase

We further investigated whether SmE contributes to the observed loss of cell viability by interfering with cell proliferation. To test this hypothesis, we examined the fraction of actively proliferating cells by determining the percent of H1299 cells in S phase after Ad-SmE or Ad-GFP infection. Expression of SmE was associated with a marked decrease of the population of cells in S phase to approximately 50% after 72 hrs, compared with that of H1299 cells infected with Ad-GFP (Fig.4A). Instead, the majority of SmE expressing cells accumulated in G2/M phase (27%± 3.4 versus 16%± 2.6 in control vector treated cells at 72 hr after infection). In contrast, only a slight increase of the G1 cell population was observed following expression of SmE (Fig.4A, bottom panel).

Figure 4

SmE protein inhibits DNA synthesis and arrests cells in G2 phase. (A) Cells were seeded at 5 × 105/dish prior to Ad-SmE (grey bar) or Ad-shGFP (negative control, black bar) treatment. At indicated time points after infection, cells were labelled with BrdU and stained for DNA content with 7-amino-actinomycinD (7-AAD). BrdU- and 7-AAD labelled cells were measured by FACS analysis. The percentage of cells resembling actively proliferating cells in S, G1 and G2/M phase is shown. Each graph represents the mean ± SD of three independent experiments. (B) M phase cells were visualized by a FITC-conjugated anti-MPM2 antibody and stained for DNA content with propidium iodide (PI). Cells treated with 10 mM nocodazole for 16 hrs were used as a positive control [POS]. pAd-siGFP infected cells served as negative control [NEG]. The percentage of cells with M phase DNA content is as indicated (M).

To further distinguish G2 and M phase cells, we performed MPM2 staining using a FITC-conjugated antibody against phospho-proteins that are active during mitosis (Fig.4B). After 16 hrs only 1–3% of cells expressing SmE showed an M phase DNA content in FACS analysis, similar to control vector infected cells. In contrast, 38% of cells stained positive after treatment with nocodazole, illustrating that SmE promotes G2 arrest, but not M phase entry.

SmE modulates expression of cell cycle checkpoint associated genes

Cell cycle progression is driven by a co-ordinated regulation of the activating of cyclin-dependent kinases (CDKs), CDK inhibitors (CDKIs) and their positive regulatory cyclins. To examine the mechanism underlying SmE-mediated cell growth arrest, we analysed the expression levels of several key regulators of the cell cycle machinery (Fig.5). It is well known that cyclin E, which is expressed in mid or late G1 phase complexes with CDK2, and the resulting kinase activity is required for S phase entry and initiation of DNA replication. Consistent with the observed inhibition of DNA synthesis by SmE, expression of cyclin E and CDK2 was profoundly reduced in H1299 cells overexpressing SmE (Fig.5A, upper panel). This decrease of cyclin E/CDK2 was accompanied by a moderate up-regulation of the CDK inhibitor p27KIP1, which especially interacts with CDK2, thereby inhibiting its catalytic activity [26]. In contrast, we did not find alterations of CDC25A phosphatase expression levels that specifically induces G1 arrest, which is in accordance with our FACS data.

Figure 5

Expression changes of cyclin E/CDK2, G2/M checkpoint proteins and spliceosome component U1 small nuclear ribonucleoprotein (SnRNP) by spliceosomal protein E (SmE). (A) At different time points after infection of H1299 cells with Ad-SmE or Ad-GFP (control), whole-cell lysates were prepared and subjected to immunoblot analysis with the indicated antibodies. Actin was used as a control for equal loading. The ratio of pCDC2/CDC2 was calculated by measuring the relative intensity of each protein band through bio-imaging analysis using the TINA program version 2.09. (B) Quantitative RT-PCR analysis was performed on total RNA prepared from Ad-SmE (grey bar) or Ad-GFP (black bar) treated H1299 cells. CDC2, CDK2 and CDC25C are shown at 24, 48 and 72 hrs after infection. Bars show the copy number x100.000 after normalization for GAPDH as a housekeeping gene. Significant differences (P < 0.05; paired,two-sided t-test) between SmE treated and control cells were labelled with asterisks.

Next we examined several G2/M check point related proteins to elucidate the molecular events leading to SmE-induced G2 arrest. Cyclin B1 binds to CDC2 and is expressed in late S and G2 phase, but cyclin B1/CDC2 complexes remain inactive until late G2 when their activation is required for entry into mitosis. As shown in Figure 5A (central panel), SmE expression resulted in a significant decrease of total CDC2 and its inactive Tyr-15 phosphorylated form (p-CDC2). However, the ratio of p-CDC2 to total CDC2 increased from 0.94 to 1.47 following infection with Ad-SmE, suggesting that the dephosphorylated active form of CDC2 kinase was reduced. Consistent with the suppression of active CDC2, the expression of phosphatase CDC25C, which is responsible for CDC2 activation by removing the inhibitory phosphorylation at Tyr-15 was significantly diminished. A significant repression of CDC2, CDK2 and CDC25C gene expression by SmE was also found on RNA level as shown by quantitative RT-PCR (Fig.5B).

In addition to Sm proteins, the small nuclear ribonucleoproteins U1, U2, U4/U6 and U5 are essential components of the spliceosome that catalyses pre-mRNA splicing [29]. To investigate whether the RNA splicing activity of SmE is related to its cell growth regulatory function, we analysed coexpression of other components of the functional spliceosome such as U1 SnRNP70. No obvious alteration was observed between cells infected with the control vector and SmE expressing cells (Fig.5A, bottom panel), suggesting that SmE induced cell growth inhibition is independent from its pre-mRNA splicing activity.

Knock down of endogenous SmE promotes cell cycle progression

To study the effect of endogenous SmE on cell cycle progression, H1299 cells were grown under starvation conditions for 48 hrs after infection with an Ad vector expressing shRNA against SmE (Ad-shSmE). Subsequently cell growth was promoted by addition of DMEM/20% FCS in the presence of 1 μM BrdU for 1 hr, and DNA synthesis was measured by flow cytometry. Compared to control vector (Ad-shGFP) infected cells, the amount of BrdU-positive cells corresponding to S phase slightly increased (34–42%, P= 0.047) in cells in which SmE was specifically knocked down by overexpression of shSmE. Concomitantly, the percentage of cells arresting in G2/M phase declined from 12 to 8% after inactivation of SmE (P= 0.0026; Fig.6), indicating that endogenous SmE protein contributes to cell growth regulation in proliferating cells.

Figure 6

Suppression of endogenous SmE protein expression reverses its cell cycle inhibitory effect. Determination of percent H1299 cells in S and G2/M phase after infection with adenoviral vector encoding shRNA against SmE (grey bar) or GFP (black bar). 48 hrs after viral infection, BrdU was applied to the cells for 1 hr, and then cells were fixed and processed for flow cytometric analysis (left panel). The mean percentage of cells in S phase and G2/M phase is as indicated (Ad-shGFP versus Ad-shSmE treated cells, P = 0.047 and 0.0026 for S and G2/M phase, respectively). The degree of SmE knock down by Ad-shSmE is shown on protein level compared to Ad-shGFP (right panel).

Discussion

SmE was originally identified as an integral component of the splicesomal complex that is involved in pre-mRNA processing. Increasing evidence, however, suggests that proteins of this family may participate in other biological processes independent of their RNA splicing function. For example, Sm proteins are necessary for transcriptional silencing in germ cell precursors and SmE is essential to maintain the expression of germ cell-specific proteins [18]. In this study we have shown that ectopic expression of SmE can efficiently reduce cell viability of several cancer cell lines independent of their endogenous p53 status via suppression of cell cycle progression. Interestingly, in agreement to previous results described for the tumour suppressor p53 [29], karyotypically normal non- tumourigenic human fibroblasts were completely insensitive to SmE-mediated cell growth inhibition. Considering that most human cancers harbour aberrations of cell cycle control, which result in a deregulated and elevated activity of the cell cycle promoting transcription factor E2F1 [30], SmE's cell cycle inhibitory effect is likely related to the higher activity of E2F1 in transformed cells.

Our data indicate that the SmEΔ1 of the SmE protein is essential for its growth arresting function. Whereas the truncated mutant SmEΔ1 exhibited a greatly reduced anti-proliferative effect in short and long-term assays, the SmEΔ2 mutant showed a similar suppressive effect on cell growth as full-length SmE. To further investigate the relevance of the SmEΔ1 in modulation of cell growth, we sought to find other proteins with a homologous sequence by NCBI protein database searches. Structural analysis revealed two other proteins with a Sm-1 like motif including the GSK3, which shares 80% identity in the Sm-1 motif. The cytoplasmic serine/threonine protein kinase that was first described in a metabolic pathway of glycogen synthase regulation was recognized by several studies as a key component modulating cell growth regulatory processes and tumourigenesis. GSK3β contributes to both cell death and survival. It has been shown that GSK3 promotes apoptosis under a variety of conditions, such as trophic factor withdrawal, toxicity induced by Alzheimer's disease amyloid ß-peptide (Aß), ceramide, heat shock, platelet activating factor and mitochondr-ial toxins [31]. Moreover, GSK3ß activity was linked to apopto-sis induced by p53 following DNA damage [32], hypoxia [33], prion toxicity [34] and endoplasmic reticulum stress [35]. Another candidate gene with a 64% identity in the SmEΔ1 belongs to EF hand calcium binding proteins, shown to be involved in malignant transformation [36]. Together these findings support the hypothesis that the SmEΔ1 plays an essential role in regulation of cellular proliferation.

The transition of one cell cycle phase to another occurs in an orderly fashion and is regulated by different cellular proteins including cyclins and CDKs, a family of serine/threonine protein kinases that are activated at specific cell cycle checkpoints, their inhibitory proteins p27Kip or p21Cip which counteract CDK activity, and the upstream acting CDC25 phosphotase family that activates CDK by removing inhibitory phospho-tyrosine residues in the cdc2 molecule [37]. Our data support the notion that SmE-mediated cell growth arrest is associated with a distinct regulation of these cell checkpoint proteins, suggesting a potential mechanism shown in Figure 7. It has been known from previous studies that cyclin E associates with CDK2 to regulate progression from G1 into S phase [38]. We observed a pronounced decrease of cyclin E and CDK2 expression following SmE overexpression accompanied by up-regulation of the CDK2 inhibitor p27Kip, indicating that SmE-induced inhibition of DNA synthesis occurs through expression alteration of these proteins. After completion of DNA-synthesis, the timely regulation of CDC2 kinase activity is a prerequisite in allowing cell entry into mitosis. Of central importance for the proper activation of CDC2 is its dephosphorylation, for example, at Tyr15 [39]. Importantly, the G2 arrest mediated by SmE was associated with a clear reduction in CDC2 expression and a concomitant decrease of its activator CDC25C. The increase in the ratio of phosphorylated to total CDC2 revealed a relative decrease in its active dephosphorylated form, which is in accordance with the lower expression of CDC25C phosphatase. In addition, consistent with an earlier observation demonstrating that p53 depleted HCT116 cells, which preferentially arrest in G2 phase upon p14ARF expression, exhibit increasing levels of cyclin B1 [40], cyclin B1 was also found up-regulated upon SmE expression in p53 deficient H1299 cells that underwent a prolonged G2 arrest. The detailed contribution of cyclin B1 to the SmE mediated G2 arrest, however, is unclear.

Figure 7

Mechanism of cell cycle inhibition by SmE. Increased expression of SmE blocks DNA synthesis by down-regulation of cyclin E/CDK2 and arrests cells in G2 phase. SmE activity at the G2 checkpoint directly correlates with the repression of CDC25C and CDC2, resulting in a relative reduction of its dephosphorylated (active) form.

Cell cycle arrest in response to DNA damage is an important mechanism for maintaining genomic integrity. Like apoptosis, cell cycle arrest is essential to suppress the propagation of damaged DNA and to inhibit transformation and tumour progression. p53 plays an essential role in G1 and G2 arrest. The loss of p53 eliminates G1 arrest in response to DNA damage, but cells lacking p53 still rapidly arrest in G2, implicating that p53-independent pathways also play an important role [41]. However, p53 is necessary to sustain long-term G2 arrest. The mechanism by which p53 maintains G2 arrest has been attributed to transcriptional induction of the p21, gadd45 and 14–3-3° genes [42], finally resulting in the reduction of cyclin B/CDC2 kinase activity. In contrast, the contribution of p53-independent pathways to G2 arrest has not been extensively investigated. It is suggested that this pathway is controlled by the ATM kinase and the ATM and RAD3-related (ATR) kinase, which phosphorylates and activates CHK1 and CHK2 serine kinases in response to DNA damage. Compared to p53-dependent pathways, down-regulation of dephosphorylated CDC2 and reduction of the CDC25C phosphatase seems to be a common mechanism for both pathways. Cells which have initiated a G2/M arrest in response to genotoxic stress can succumb to a variety of fates, including apoptosis [43], prolonged permanent arrest [44], recovery after repair of DNA damage or adaptation to the damage, thereby allowing progression through the cell cycle with damaged DNA that initially evoke the arrest. Our experiments, however, suggest that cells arresting in G2 upon SmE overexpression can not re-enter into the cell cycle. This conclusion is validated by long-term colony formation assays, demonstrating that cell proliferation is greatly inhibited even two weeks after Ad-SmE infection.

In summary, our data provide evidence that small nuclear ribonucleoproteins containing the SmEΔ1 play a role in p53-independent cell cycle arrest pathways. Although SmE acts by modulating the expression of key molecules of cell cycle progression, suggesting a non-spliceosomal or splicing-independent mechanism, we can not exclude the involvement of the RNA splicing machinery for the cell growth inhibitory function of SmE. Considering that a large portion of human cancers are defective in p53 activity and given the role of SmE in p53-independent growth arrest of tumour cells shown here, targeting SmE in cancer gene therapy will be of particular interest.