Abstract
Expression of the human papillomavirus 16 E6 oncogene interferes with several vital cellular processes, including the p53-dependent response to DNA damage. To assess the influence of E6 on the early response to DNA damage, we analyzed gene expression following mitomycin C–induced genotoxic stress in human E6–expressing U2OS cells (U2OSE64b) as well as in p53-expressing control cells (U2OSE6AS) by comparative global expression profiling. As expected, genes involved in p53-dependent pathways were activated in p53-expressing cells. In the U2OSE64b cells, however, a largely nonoverlapping group of genes was identified, including two splicing factors of the SR family. Immunoblot analysis revealed increased expression of several SR proteins during the early response to DNA damage, which was accompanied by activation of alternative splicing activity. Disruption of splicing activity by treatment with small interfering RNA directed against splicing factor SRp55 resulted in the increased viability of p53-deficient cells following DNA damage. To determine whether the transient activation of splicing activity was due to E6-mediated degradation of p53, or was due to some other activity of E6, we compared the early response of the p53 wild-type and p53−/− isogenic HCT116 cell lines, and found that the increase in splicing activity was observed only in the absence of p53. Finally, both the U2OSE64b and the p53−/− cells showed altered splicing patterns for the CD44 receptor. Together, these data show that cells lacking p53 can activate alternative splicing following DNA damage. [Cancer Res 2007;67(16):7621–30]
Introduction
Genotoxic stresses from chemical, physical, or viral agents initiate damage response pathways that lead to cell cycle arrest, DNA repair, and apoptosis (1, 2). The mechanisms by which these cellular responses are triggered are not fully understood, but it is clear that the p53 tumor repressor is one of the most important checkpoint proteins, influencing multiple response pathways to a diversity of stressors (3).
However, p53 is not indispensable for the induction of apoptosis (4, 5). This has clinical relevance, as more than half of human tumors lack wild-type p53. Therefore, if apoptosis is to be induced in these cells, p53-independent pathways must be engaged. This is particularly true in cervical carcinoma cases caused by human papillomavirus (HPV) 16 infection, as the viral E6 protein effectively eliminates cellular p53 by accelerating its proteolysis (6). E6-expressing cells, therefore, can function as a model system for discovering and studying p53-independent, DNA damage–triggered apoptotic pathways.
In contrast to the well-established role of transcriptional regulation, alternative splicing is only now emerging as an important mechanism for regulating gene function. It has been estimated that at least 74% of multi-exon genes are alternatively spliced (7), and these isoforms frequently have opposing functions. This may have a significant effect on a variety of cell processes, including apoptosis (8).
There is now ample evidence that aberrations of alternative splicing are widespread in tumors (9). Although the relationship between changes in splicing and tumor progression has not yet been well-defined, condition-dependent splicing alterations in such genes as FAS, RBM9, CD44, hnRNPA/B, APLP2, and MYL6 were found in breast cancer cells by using splicing-sensitive microarrays (10). Despite growing interest in splicing regulation and the discovery of a large number of genes involved in pre–RNA splicing, the mechanisms governing this process are not well-understood (11). It is believed that alternative splicing activity is regulated by the combinatorial actions of a fairly large number of regulatory factors, including two families of proteins that generally function in an antagonistic manner, heterogeneous ribonucleoproteins and SR (serine/arginine-rich) proteins (12).
SR proteins participate in both constitutive and alternative splicing. Although they seem to be functionally redundant in constitutive splicing, a growing body of evidence shows that they have distinctive and not fully overlapping functions in alternative splicing. Each particular SR protein is able to bind to a specific splicing enhancer and activate splicing of weak splicing sites, leading to altered splicing patterns of target genes (12).
In this report, we have investigated the early response of the osteosarcoma U2OSE64b (E6-expressing) and colon carcinoma HCT116 p53−/− cell lines to DNA damage caused by the genotoxic drug mitomycin C. The results from this study provide strong evidence that alternative splicing activity can be involved in regulating the stress response in the absence of the classical p53-dependent pathways.
Materials and Methods
Cell lines and cell culture. U2OS cells, derived from a human osteosarcoma, were obtained from the American Type Culture Collection. HCT116 human colorectal cancer cells and p53-null derivatives (13) were supplied by Dr. B. Vogelstein. They were cultured in McCoy's 5A medium (Invitrogen) supplemented to contain 10% fetal bovine serum (Invitrogen), penicillin (100 μg/mL), and streptomycin (100 μg/mL; Sigma). Construction of the pHA-E6S and pHA-E6AS plasmids, which respectively, contain either the sense or the antisense versions of epitope-tagged E6 (HA-E6) under the control of the CMV promoter, and establishment of the U2OSE64b and U2OSE6AS cell lines by stable transfection with these plasmids, have been described previously (14).
Measurement of mitomycin C cytotoxicity. To measure cell survival following mitomycin C treatment, cells in exponential growth were exposed to mitomycin C (Roche) at the concentrations noted for the individual experiments. After incubation for the indicated time, the number of viable cells was quantified by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay as described earlier (14). Three independent experiments were each conducted in triplicate and significance was analyzed by Student's t test.
Immunoblot analysis. Proteins were extracted from cells (106) in 100 μL of Laemmli's buffer. Lysates were sonicated and protein concentration was determined by the DC protein assay (Bio-Rad). Lysates (15 μg) were separated on SDS-polyacrylamide gels and transferred onto Immobilon P membranes (Millipore). Membranes were probed with primary antibodies directed against p53 (NCL-p53-DO7; Novocastra), poly(ADP-ribose) polymerase-1 (PARP-1; Ab-2; EMD Biosciences), phosphorylated Chk1 (phospho-Chk1-Ser345; Cell Signaling Technology), β-actin (Sigma), or pan-SR (1H4; American Type Culture Collection). After washing, membranes were incubated with the appropriate horseradish peroxidase–conjugated secondary antibodies (Sigma), and bands were detected using the SuperSignal West Substrate System (Pierce). Densitometry measurements were done using the ChemImager Imaging System and processed with the AlphaEase software package (Alpha Innotech).
Semiquantitative reverse transcriptase-PCR. Reverse transcriptase-PCR (RT-PCR) analysis was used to estimate transcript levels in total RNA samples isolated from cell cultures. Ten micrograms of total RNA was used as a template for cDNA synthesis with SuperScript III reverse transcriptase and an oligo(dT) primer according to the manufacturer's instructions (Invitrogen). The cDNA obtained was normalized by PCR with primers specific to cofilin1 (CFL1), a gene that showed no change in expression in microarray experiments (CCTTCCCAAACTGCTTTTGAT and CTGGTCCTGCTTCCATGAGTA). Serial dilutions of the normalized cDNA samples were used to find the linear range of amplification for gene-specific DNA fragments. PCR was done at 94°C for 30 s, at 57°C for 30 s, and at 72°C for 1 min for 35 cycles using Taq DNA polymerase and the recommended protocol (NEB). For the SRFS6 gene, primers were CAGGTCGAGTTCCAGAGATTA and TCAAACTGCAATTTCAACTCA. CD44-specific primers were taken from ref. (15).
Expression profiling by microarray analysis. Total RNA was extracted using Trizol Reagent (Invitrogen) and additionally purified using the RNeasy kit (Qiagen) according to the manufacturer's instructions. RNA samples from three independent replicates for each time point were further processed and hybridized to Affymetrix Human Genome U133A and U133B expression arrays according to standard recommended protocols (Affymetrix) at the DNA MicroArray Facility, University of California (Irvine, CA). To extract and analyze microarray data, we used the Microarray Suite 5.0, GCOS 1.2, and Data Mining Tools 3.0 programs (Affymetrix). To compare data from different arrays, the signal intensities of the arrays were globally scaled to 500 and normalization was done using a probe set of 100 constitutively expressed transcripts provided by Affymetrix. The signal log ratio was calculated by comparing transcripts between mitomycin C–treated and untreated cells. Generation of detected (present or absent) and changed (increased or decreased) calls was done using the Wilcoxon test and default variables of P value cutoffs. During the first step to identify genes with altered expression, we discarded genes that had changed calls in less than six out of nine comparative files (Rank test) as well as genes that had “absent” calls in all microarrays. The remaining genes were further analyzed using both a Student's t test and the Mann-Whitney test (P < 0.05 cutoffs). Genes were considered up-regulated if they passed the Mann-Whitney and Rank tests, and down-regulated if they passed the Rank and t tests.
Quantitative in vivo analysis of alternative splicing activity. Cells (5 × 105) were transfected with the pLuc14 plasmid containing a tau minigene fused to luciferase (16) as described previously (17) and split into a 96-well plate 24 h later. At 48 h posttransfection, cells were treated with 2 μg/mL of mitomycin C for the indicated amount of time and harvested in luciferase lysis buffer (Promega). Luciferase activity was quantified using a MicroLumatPlus microplate reader (Berthold Technology). Three independent experiments, were each conducted in triplicate, and significance was analyzed by one-way ANOVA. Differences were considered significant at a 0.05 level of confidence.
Inhibition of SFRS6 gene expression by small interfering RNA. Cells were transfected with predesigned small interfering RNAs (siRNA) for the human SFRS6 gene from Ambion or from Santa Cruz Biotechnology, or with control siRNA using the siRNA transfection kit as recommended by the manufacturer (Santa Cruz Biotechnology). Levels of silencing of the SFRS6 gene were evaluated by RT-PCR and immunoblot analyses as described above. Twenty-four hours posttransfection, cells were treated with mitomycin C for 24 h, followed by quantification of viable cells by the MTT assay (14). Three independent experiments were done, each was conducted in quadruplicate, and significance was analyzed by one-way ANOVA. Differences were considered significant at a 0.05 level of confidence.
Results
E6 decreases the level of cellular p53, but does not inhibit cell death following DNA damage. Our laboratory has previously shown that in U2OS cells in which E6 expression is regulated by a tetracycline-responsive promoter, increased amounts of E6 led to a dose-dependent decrease in p53 as measured by p53 ELISA (17). This suggested that E6 might decrease apoptosis triggered by genotoxic stress. To test this possibility, cells expressing E6 (U2OSE64b) and control cells transfected with an antisense variant of E6 (U2OSE6AS) were treated with mitomycin C, a potent genotoxic agent (18). Surprisingly, the two cell lines responded in a very similar manner to mitomycin C treatment, either as assessed by changing the dose of drug or by changing the duration of treatment (Fig. 1A). At the same time, the p53 immunoblot analysis clearly showed that although the concentration of p53 in U2OSE64b cells did increase somewhat during treatment, the overall level of p53 was reduced by at least 10-fold as compared with the control U2OSE6AS cells (Fig. 1B). In fact, the level of p53 in U2OSE64b cells after 5 h of treatment was comparable to that seen in the untreated U2OSE6AS cells.
HPV16 E6 decreases cellular p53 levels but does not protect cells from mitomycin C. A, U2OSE64b (—▪—) and U2OSE6AS cells (—♦—) were incubated for 16 h with the indicated concentrations of mitomycin C (left) or treated with 2 μg/mL of mitomycin C for the indicated times (right) and subjected to the MTT test. Measurements were made in triplicate; bars, SD. B, C, and D, U2OSE64b and U2OSE6AS cells were incubated for the indicated times with 2 μg/mL of mitomycin C and analyzed by immunoblot with antibodies directed against p53 (B), the phosphorylated form of CHK1 (C), and PARP-1 (D). The β-actin antibodies were used for re-blotting to verify uniformity of sample loading.
HPV16 E6 decreases cellular p53 levels but does not protect cells from mitomycin C. A, U2OSE64b (—▪—) and U2OSE6AS cells (—♦—) were incubated for 16 h with the indicated concentrations of mitomycin C (left) or treated with 2 μg/mL of mitomycin C for the indicated times (right) and subjected to the MTT test. Measurements were made in triplicate; bars, SD. B, C, and D, U2OSE64b and U2OSE6AS cells were incubated for the indicated times with 2 μg/mL of mitomycin C and analyzed by immunoblot with antibodies directed against p53 (B), the phosphorylated form of CHK1 (C), and PARP-1 (D). The β-actin antibodies were used for re-blotting to verify uniformity of sample loading.
Next, we monitored CHK1 kinase activation, one of the first steps of the DNA damage response (19), using an antibody specific for its phosphorylated form. Accumulation of activated CHK1 kinase was observed in both cell lines, showing that these cell lines responded similarly to the genotoxic stress at this level (Fig. 1C). To evaluate whether mitomycin C treatment of U2OS cells causes cell death through apoptosis and to evaluate its progression, we analyzed PARP-1 cleavage in the treated cells. PARP-1 is activated in response to DNA damage by cleavage into two apoptosis-specific fragments of 24 and 89 kDa (20, 21). Analysis of PARP-1 cleavage during mitomycin C treatment revealed the presence of the apoptosis-specific 89 kDa fragment within the first 24 h of treatment with 2 μg/mL of mitomycin C in both cell lines, indicating that they underwent apoptosis (Fig. 1D).
E6 alters the set of genes affected by DNA damage. The survival experiments described above indicated that the cytotoxicity of mitomycin C for both U2OS-derived cell lines was essentially equivalent. However, the differences in p53 levels suggested that the pathways involved might be quite different. Because we were interested in the early response of cells to DNA damage, we decided to evaluate the global changes in expression after 5 h of treatment with 2 μg/mL of mitomycin C. Based on the time of CHK1 kinase activation, at this drug concentration and time of treatment, the cells should have begun to respond to the insult. RNA samples were extracted from both cell lines, with and without mitomycin C treatment, and subjected to microarray analysis. According to the criteria outlined in Materials and Methods, we found 38 genes in U2OSE6AS cells and 22 genes in U2OSE64b cells that showed statistically significant up-regulation or down-regulation (Tables 1 and 2).
List of genes with altered expression in U2OSE6AS cells
Gene name* . | Gene title . | Gene ontology biological process† . | Fold change (mean ± SD) . | |||
---|---|---|---|---|---|---|
UP-REGULATED GENES | ||||||
p53-dependent genes | ||||||
MDM2 | Mdm2, p53-binding protein | Cell growth, negative regulation of cell proliferation | 2.46 ± 0.75Ê | |||
SESN2 | Sestrin 2 | Cell cycle arrest | 2.33 ± 0.69 | |||
BTG2 | BTG family, member 2 | DNA repair | 2.11 ± 0.25 | |||
ATF3 | Activating transcription factor 3 | Regulation of transcription | 1.65 ± 0.22 | |||
PPM1D | Protein phosphatase 1D magnesium-dependent, δ | Negative regulation of cell proliferation, cell growth, and response | 1.44 ± 0.07 | |||
TP53INP1 | Tumor protein p53-inducible nuclear protein 1 | Apoptosis, p53 regulation | 1.43 ± 0.20 | |||
PMAIP1 | NOXA | Regulation of apoptosis | 1.37 ± 0.41 | |||
Other genes with known functions | ||||||
TOB1 | Transducer of ERBB2, 1 | Negative regulation of cell proliferation | 2.23 ± 0.69 | |||
FBXW7 | F-box and WD-40 domain protein 7 | Ubiquitin cycle | 1.95 ± 0.59 | |||
RALGDS | Ral guanine nucleotide dissociation stimulator | Small GTPase-mediated signal transduction | 1.92 ± 0.56 | |||
CPM | Carboxypeptidase M | Aromatic compound metabolism, morphogenesis, and proteolysis | 1.71 ± 0.21Ê | |||
GPR87 | G protein–coupled receptor 87 | G-protein coupled receptor protein signaling pathway | 1.51 ± 0.16 | |||
MCM10 | MCM10 mini-chromosome maintenance–deficient 10 | DNA replication | 1.43 ± 0.54 | |||
Genes with unidentified functions | ||||||
C6orf69 | Chromosome 6 ORF 69 | 1.67 ± 0.61 | ||||
TTC17 | Tetratricopeptide repeat domain 17 | 1.61 ± 0.49 | ||||
DOWN-REGULATED GENES | ||||||
Signal transduction | ||||||
RAPGEF2 | Rap guanine nucleotide exchange factor (GEF) 2 | MAPKKK cascade, cyclic AMP–mediated signaling | Ê−1.75 ± 0.12 | |||
TIAM1 | T cell lymphoma invasion and metastasis 1 | Intracellular signaling cascade | −1.67 ± 0.14 | |||
TRIO | Triple functional domain (PTPRF interacting) | Protein tyrosine phosphatase signaling pathway | −1.54 ± 0.18 | |||
Regulation of transcription, RNA processing | ||||||
CUTL1 | Cut-like 1, CCAAT displacement protein | Negative regulation of transcription | −1.69 ± 0.10 | |||
HNRPA2B1 | Heterogeneous nuclear ribonucleoprotein A2/B1 | RNA processing | −1.77 ± 0.10 | |||
GTF3C1 | General transcription factor IIIC, polypeptide 1 | Regulation of transcription | −1.66 ± 0.16 | |||
CTBP2 | COOH-terminal binding protein 2 | Transcriptional repressor, l-serine biosynthesis | −1.44 ± 0.19 | |||
DDX17 | DEAD (Asp-Glu-Ala-Asp) box polypeptide 17 | RNA processing | −1.43 ± 0.19 | |||
Chromosome organization | ||||||
CENPA | Centromere protein A (17 kDa) | Chromosome organization and biogenesis, nucleosome assembly | Ê−1.55 ± 0.15 | |||
TUBGCP3 | Tubulin, γ complex-associated protein 3 | Microtubule nucleation | −1.60 ± 0.08 | |||
Genes with known functions | ||||||
ITM2B | Integral membrane protein 2B | Neurogenesis | −1.58 ± 0.44Ê | |||
IFRD1 | IFN-related developmental regulator 1 | Myoblast cell fate determination | −1.57 ± 0.10 | |||
Genes with unidentified functions | ||||||
DKFZp434H205 | mRNA; cDNA DKFZp434H205 | −2.53 ± 0.58 | ||||
TncRNA | Trophoblast-derived noncoding RNA | −1.93 ± 0.26 | ||||
C20orf129 | Chromosome 20 ORF 129 | −1.80 ± 0.25 | ||||
FRAS1 | Fraser syndrome 1 | −1.75 ± 0.48 | ||||
ATXN1 | Ataxin 1 | −1.69 ± 0.27 | ||||
RAI17 | Retinoic acid–induced 17 | −1.69 ± 0.24 | ||||
DKFZp762E131 | Hypothetical protein DKFZp762E1312 | −1.61 ± 0.06 | ||||
TANC | TPR domain, ankyrin-repeat and coiled-coil | −1.61 ± 0.12 | ||||
MGC57827 | Similar to RIKEN cDNA 2700049P18 gene | −1.54 ± 0.11 | ||||
NSUN2 | NOL1/NOP2/Sun domain family 2 | −1.49 ± 0.29 | ||||
RBAF600 | Retinoblastoma-associated factor 600 | −1.49 ± 0.16 |
Gene name* . | Gene title . | Gene ontology biological process† . | Fold change (mean ± SD) . | |||
---|---|---|---|---|---|---|
UP-REGULATED GENES | ||||||
p53-dependent genes | ||||||
MDM2 | Mdm2, p53-binding protein | Cell growth, negative regulation of cell proliferation | 2.46 ± 0.75Ê | |||
SESN2 | Sestrin 2 | Cell cycle arrest | 2.33 ± 0.69 | |||
BTG2 | BTG family, member 2 | DNA repair | 2.11 ± 0.25 | |||
ATF3 | Activating transcription factor 3 | Regulation of transcription | 1.65 ± 0.22 | |||
PPM1D | Protein phosphatase 1D magnesium-dependent, δ | Negative regulation of cell proliferation, cell growth, and response | 1.44 ± 0.07 | |||
TP53INP1 | Tumor protein p53-inducible nuclear protein 1 | Apoptosis, p53 regulation | 1.43 ± 0.20 | |||
PMAIP1 | NOXA | Regulation of apoptosis | 1.37 ± 0.41 | |||
Other genes with known functions | ||||||
TOB1 | Transducer of ERBB2, 1 | Negative regulation of cell proliferation | 2.23 ± 0.69 | |||
FBXW7 | F-box and WD-40 domain protein 7 | Ubiquitin cycle | 1.95 ± 0.59 | |||
RALGDS | Ral guanine nucleotide dissociation stimulator | Small GTPase-mediated signal transduction | 1.92 ± 0.56 | |||
CPM | Carboxypeptidase M | Aromatic compound metabolism, morphogenesis, and proteolysis | 1.71 ± 0.21Ê | |||
GPR87 | G protein–coupled receptor 87 | G-protein coupled receptor protein signaling pathway | 1.51 ± 0.16 | |||
MCM10 | MCM10 mini-chromosome maintenance–deficient 10 | DNA replication | 1.43 ± 0.54 | |||
Genes with unidentified functions | ||||||
C6orf69 | Chromosome 6 ORF 69 | 1.67 ± 0.61 | ||||
TTC17 | Tetratricopeptide repeat domain 17 | 1.61 ± 0.49 | ||||
DOWN-REGULATED GENES | ||||||
Signal transduction | ||||||
RAPGEF2 | Rap guanine nucleotide exchange factor (GEF) 2 | MAPKKK cascade, cyclic AMP–mediated signaling | Ê−1.75 ± 0.12 | |||
TIAM1 | T cell lymphoma invasion and metastasis 1 | Intracellular signaling cascade | −1.67 ± 0.14 | |||
TRIO | Triple functional domain (PTPRF interacting) | Protein tyrosine phosphatase signaling pathway | −1.54 ± 0.18 | |||
Regulation of transcription, RNA processing | ||||||
CUTL1 | Cut-like 1, CCAAT displacement protein | Negative regulation of transcription | −1.69 ± 0.10 | |||
HNRPA2B1 | Heterogeneous nuclear ribonucleoprotein A2/B1 | RNA processing | −1.77 ± 0.10 | |||
GTF3C1 | General transcription factor IIIC, polypeptide 1 | Regulation of transcription | −1.66 ± 0.16 | |||
CTBP2 | COOH-terminal binding protein 2 | Transcriptional repressor, l-serine biosynthesis | −1.44 ± 0.19 | |||
DDX17 | DEAD (Asp-Glu-Ala-Asp) box polypeptide 17 | RNA processing | −1.43 ± 0.19 | |||
Chromosome organization | ||||||
CENPA | Centromere protein A (17 kDa) | Chromosome organization and biogenesis, nucleosome assembly | Ê−1.55 ± 0.15 | |||
TUBGCP3 | Tubulin, γ complex-associated protein 3 | Microtubule nucleation | −1.60 ± 0.08 | |||
Genes with known functions | ||||||
ITM2B | Integral membrane protein 2B | Neurogenesis | −1.58 ± 0.44Ê | |||
IFRD1 | IFN-related developmental regulator 1 | Myoblast cell fate determination | −1.57 ± 0.10 | |||
Genes with unidentified functions | ||||||
DKFZp434H205 | mRNA; cDNA DKFZp434H205 | −2.53 ± 0.58 | ||||
TncRNA | Trophoblast-derived noncoding RNA | −1.93 ± 0.26 | ||||
C20orf129 | Chromosome 20 ORF 129 | −1.80 ± 0.25 | ||||
FRAS1 | Fraser syndrome 1 | −1.75 ± 0.48 | ||||
ATXN1 | Ataxin 1 | −1.69 ± 0.27 | ||||
RAI17 | Retinoic acid–induced 17 | −1.69 ± 0.24 | ||||
DKFZp762E131 | Hypothetical protein DKFZp762E1312 | −1.61 ± 0.06 | ||||
TANC | TPR domain, ankyrin-repeat and coiled-coil | −1.61 ± 0.12 | ||||
MGC57827 | Similar to RIKEN cDNA 2700049P18 gene | −1.54 ± 0.11 | ||||
NSUN2 | NOL1/NOP2/Sun domain family 2 | −1.49 ± 0.29 | ||||
RBAF600 | Retinoblastoma-associated factor 600 | −1.49 ± 0.16 |
Gene names represent gene symbols according to the HUGO Gene Nomenclature Committee.
Gene annotation is based on the GeneOntology Consortium (http://www.geneontology.org/) and Affymetrix NetAffx Analysis Center (http://www.affymetrix.com/analysis/netaffx/index.affx).
List of genes with altered expression in U2OSE64b cells
Gene name* . | Gene title . | Gene ontology biological process† . | Fold change (mean ± SD) . | |||
---|---|---|---|---|---|---|
UP-REGULATED GENES | ||||||
Cell cycle regulation | ||||||
CDC25A | Cell division cycle 25A | Regulation of CDK activity, mitosis | 1.93 ± 0.48 | |||
SESN2 | Sestrin 2 | Cell cycle arrest | 1.81 ± 0.33 | |||
CDC42 | Cell division cycle 42 | Small GTPase mediated signal transduction | 1.75 ± 0.36 | |||
Regulation of transcription, RNA processing | ||||||
PEG3 | Paternally expressed 3 | Regulation of transcription, DNA-dependent | 2.01 ± 0.55 | |||
SFRS6 | Splicing factor, arginine/serine-rich 6 | mRNA splice site selection | 1.92 ± 0.43 | |||
TULP4 | Tubby like protein 4 | Regulation of transcription | 1.84 ± 0.92 | |||
SFRS7 | Splicing factor, arginine/serine-rich 7 | mRNA splicing | 1.52 ± 0.20 | |||
THOC4 | THO complex 4 | mRNA-nucleus export, mRNA splicing | 1.46 ± 0.17 | |||
Other genes with known functions | ||||||
BRD2 | Bromodomain containing 2 | Spermatogenesis, mitogen-activated kinase | 1.68 ± 0.46 | |||
PMAIP1 | NOXA | Regulation of apoptosis | 1.65 ± 0.21 | |||
Genes with unidentified functions | ||||||
MALAT-1 | Metastasis-associated lung adenocarcinoma transcript 1 | 2.22 ± 0.88 | ||||
LOC162073 | Hypothetical protein LOC162073 | 2.17 ± 0.97 | ||||
IRF2BP2 | IFN regulatory factor 2 binding protein 2 | 1.76 ± 0.49 | ||||
PNAS-4 | CGI-146 protein | 1.73 ± 0.56 | ||||
VMP1 | Likely orthologue of rat vacuole membrane | 1.60 ± 0.19 | ||||
KIAA1171 | KIAA1171 protein | 1.49 ± 0.26 | ||||
WIPI49-like | DKFZP434J154 protein | 1.39 ± 0.12 | ||||
C14orf4 | Chromosome 14 open reading frame 4 | 1.34 ± 0.08 | ||||
DOWN-REGULATED GENES | ||||||
Regulation of transcription, RNA processing | ||||||
CUTL1 | Cut-like 1, CCAAT displacement protein | Negative regulation of transcription | −1.77 ± 0.20 | |||
Chromosome organization | ||||||
CENPA | Centromere protein A (17 kDa) | Chromosome organization and biogenesis, nucleosome assembly | Ê−1.64 ± 0.10 | |||
Genes with unidentified functions | ||||||
DKFZp762E13 | Hypothetical protein DKFZp762E1312 | −1.44 ± 0.18 | ||||
C20orf129 | Chromosome 20 open reading frame 129 | −1.41 ± 0.04 |
Gene name* . | Gene title . | Gene ontology biological process† . | Fold change (mean ± SD) . | |||
---|---|---|---|---|---|---|
UP-REGULATED GENES | ||||||
Cell cycle regulation | ||||||
CDC25A | Cell division cycle 25A | Regulation of CDK activity, mitosis | 1.93 ± 0.48 | |||
SESN2 | Sestrin 2 | Cell cycle arrest | 1.81 ± 0.33 | |||
CDC42 | Cell division cycle 42 | Small GTPase mediated signal transduction | 1.75 ± 0.36 | |||
Regulation of transcription, RNA processing | ||||||
PEG3 | Paternally expressed 3 | Regulation of transcription, DNA-dependent | 2.01 ± 0.55 | |||
SFRS6 | Splicing factor, arginine/serine-rich 6 | mRNA splice site selection | 1.92 ± 0.43 | |||
TULP4 | Tubby like protein 4 | Regulation of transcription | 1.84 ± 0.92 | |||
SFRS7 | Splicing factor, arginine/serine-rich 7 | mRNA splicing | 1.52 ± 0.20 | |||
THOC4 | THO complex 4 | mRNA-nucleus export, mRNA splicing | 1.46 ± 0.17 | |||
Other genes with known functions | ||||||
BRD2 | Bromodomain containing 2 | Spermatogenesis, mitogen-activated kinase | 1.68 ± 0.46 | |||
PMAIP1 | NOXA | Regulation of apoptosis | 1.65 ± 0.21 | |||
Genes with unidentified functions | ||||||
MALAT-1 | Metastasis-associated lung adenocarcinoma transcript 1 | 2.22 ± 0.88 | ||||
LOC162073 | Hypothetical protein LOC162073 | 2.17 ± 0.97 | ||||
IRF2BP2 | IFN regulatory factor 2 binding protein 2 | 1.76 ± 0.49 | ||||
PNAS-4 | CGI-146 protein | 1.73 ± 0.56 | ||||
VMP1 | Likely orthologue of rat vacuole membrane | 1.60 ± 0.19 | ||||
KIAA1171 | KIAA1171 protein | 1.49 ± 0.26 | ||||
WIPI49-like | DKFZP434J154 protein | 1.39 ± 0.12 | ||||
C14orf4 | Chromosome 14 open reading frame 4 | 1.34 ± 0.08 | ||||
DOWN-REGULATED GENES | ||||||
Regulation of transcription, RNA processing | ||||||
CUTL1 | Cut-like 1, CCAAT displacement protein | Negative regulation of transcription | −1.77 ± 0.20 | |||
Chromosome organization | ||||||
CENPA | Centromere protein A (17 kDa) | Chromosome organization and biogenesis, nucleosome assembly | Ê−1.64 ± 0.10 | |||
Genes with unidentified functions | ||||||
DKFZp762E13 | Hypothetical protein DKFZp762E1312 | −1.44 ± 0.18 | ||||
C20orf129 | Chromosome 20 open reading frame 129 | −1.41 ± 0.04 |
Gene names represent gene symbols according to the HUGO Gene Nomenclature Committee.
Gene annotation is based on the GeneOntology Consortium (http://www.geneontology.org/) and Affymetrix NetAffx Analysis Center (http://www.affymetrix.com/analysis/netaffx/index.affx).
Only five genes with altered expression had similar expression profiles in both cell lines. Three genes were down-regulated: centromere protein A, transcription regulator CUTL1, and a gene encoding the putative protein DKFZp762E1312. Two genes were found to be up-regulated in both cell lines: NOXA, a proapoptotic member of the Bcl-2 family (22), and the stress-activated gene encoding sestrin 2 (23).
Fifteen genes were identified as up-regulated in U2OSE6AS cells by microarray analysis, with 13 of them representing genes with identified functions (Table 1). Among these genes, seven have well-documented relationships to p53-dependent apoptosis, including MDM2, ATF3, BTG2, PPM1D, NOXA, SESN2, and TP53INP1 (24–30). These data show that, as expected, DNA damage activated the p53-mediated response in U2OSE6AS cells.
A separate group of 22 genes with statistically significantly altered expression was identified in U2OSE64b cells treated with mitomycin C (Table 2). Interestingly, unlike the group of genes up-regulated in U2OSE6AS cells, more than half of the genes with increased expression in U2OSE64b cells (10 out of 18) are poorly characterized or entirely unknown.
Among the genes with known functions that are up-regulated, three participate in cell cycle regulation and five participate in the regulation of transcription and RNA processing. Interestingly, three genes are directly involved in RNA splicing, and two of them, SFRS6 and SFRS7, belong to the SR splicing factor family (31).
Splicing factors and alternative splicing activity are induced in U2OSE64b cells following DNA damage. To determine whether up-regulated transcription of genes encoding the splicing factors also corresponded to an increase at the protein level, we did immunoblot analysis using the pan-SR protein antibody, 1H4, which has been shown to recognize four human members of this family (32). In addition to increased protein levels for SRp55 encoded by SFRS6, we also found increased protein levels for the splicing factors SRp30 and SRp40 (Fig. 2A). Only one of the splicing factors reactive with this antibody, SRp75, did not show an obvious change following treatment, whereas the other three SR splicing factors had a peak of expression at 5 h posttreatment. To determine whether this decrease was due to the lack of p53 or to some other activity of E6, we monitored SR protein expression in two isogenic cell lines: HCT116, which expresses wild-type p53, and HCT116 p53−/−. Results from the immunoblot analysis clearly showed that increased expression of the SRp55 splicing factor was seen only in the cells lacking p53 (Fig. 2B), indicating that it was the p53-degradatory activity of E6 that was responsible for SRp55 up-regulation.
Mitomycin C treatment leads to an induction of splicing factor expression and an increase in alternative splicing activity in U2OS cells expressing E6. A, Western blot analysis of total cellular proteins extracted from U2OSE6AS and U2OSE64b cells treated with mitomycin C (2 μg/mL). The levels of splicing factors were determined by immunoblotting with monoclonal antibody 1H4. β-Actin antibodies were used for re-blotting to verify the uniformity of sample loading. B, Western blot analysis of protein lysates of HCT116 and HCT116 p53−/− cells treated with mitomycin C (2 μg/mL) using 1H4, p53, and β-actin antibodies. C, scheme of LucM14 construct (16) used to evaluate alternative splicing activity in vivo. In this system, luciferase activity is dependent on the exclusion of exon 10 of the tau minigene. D, increased alternative splicing activity as measured by a luciferase system based on exclusion of exon 10 of tau gene. Left, luciferase activity in U2OSE64b (—♦—), and in U2OSE6AS (—▪—). Right, increased luciferase activity in HCT116 p53−/− cells after 5 h of treatment. Cells were transfected with the LucM14 construct, treated with mitomycin C (2 μg/mL), and subjected to luciferase analysis. Columns, difference in luciferase activity between treated (t = 5) and untreated cells (t = 0). Measurements were made in triplicate; bars, SD. **, P = 0.99, level of confidence.
Mitomycin C treatment leads to an induction of splicing factor expression and an increase in alternative splicing activity in U2OS cells expressing E6. A, Western blot analysis of total cellular proteins extracted from U2OSE6AS and U2OSE64b cells treated with mitomycin C (2 μg/mL). The levels of splicing factors were determined by immunoblotting with monoclonal antibody 1H4. β-Actin antibodies were used for re-blotting to verify the uniformity of sample loading. B, Western blot analysis of protein lysates of HCT116 and HCT116 p53−/− cells treated with mitomycin C (2 μg/mL) using 1H4, p53, and β-actin antibodies. C, scheme of LucM14 construct (16) used to evaluate alternative splicing activity in vivo. In this system, luciferase activity is dependent on the exclusion of exon 10 of the tau minigene. D, increased alternative splicing activity as measured by a luciferase system based on exclusion of exon 10 of tau gene. Left, luciferase activity in U2OSE64b (—♦—), and in U2OSE6AS (—▪—). Right, increased luciferase activity in HCT116 p53−/− cells after 5 h of treatment. Cells were transfected with the LucM14 construct, treated with mitomycin C (2 μg/mL), and subjected to luciferase analysis. Columns, difference in luciferase activity between treated (t = 5) and untreated cells (t = 0). Measurements were made in triplicate; bars, SD. **, P = 0.99, level of confidence.
The results obtained by both microarray and immunoblot analyses showed enhanced expression of splicing factors in p53-depleted cells following genotoxic stress. To determine whether this increase led to a change in splicing activity, we measured alternative splicing activity using a cell-based analysis developed by Yu et al. (16). This analysis allows quantification of exclusion of the alternate exon 10 of the human tau gene by measurement of luciferase activity. The luciferase gene in this construct is fused to the mini-tau gene and is in-frame only when exon 10 is spliced out (Fig. 2C). This system provided a particularly attractive model for testing our results, as it has previously been shown that overexpression of splicing factors SRp20, SRp40, and SRp55 significantly increased the exclusion of exon 10 from the construct (16). Therefore, to evaluate alternative splicing activity after DNA damage, we transfected LucM14 into the U2OSAS, UOSE64b, HCT116, and HCT116 p53−/− cell lines. Transfected cells were treated with mitomycin C, and luciferase activity in treated versus untreated cells was compared. The cell lines deficient in p53, U2OSE64b, and HCT116 p53−/−, but not the control cell lines expressing p53, U2OSE6AS, and HCT116, showed enhanced luciferase activity at 5 h posttreatment (Fig. 2D) due to increased exclusion of exon 10 from mRNA by alternative splicing. At its peak, following 5 h of treatment, alternative splicing activity in U2OSE64b cells increased by almost 40% compared with untreated cells. However, this increase was time-dependent, and after 7.5 h of treatment, alternative splicing activity was found to be essentially the same in both U2OS cell lines. A sharp increase in luciferase activity was also detected following mitomycin treatment in p53-deficient but not in p53-proficient HCT116 cells.
Silencing the SFRS6 gene increases the survival of U2OSE64b cells following mitomycin C treatment. To explore the possibility that SRp55 activity might affect the overall cellular response to genotoxic stress, its expression was silenced by siRNA inhibition and the resulting sensitivity to mitomycin C treatment was evaluated by the MTT assay. To ensure that the silencing effect was specific, we did these survival experiments using two independently designed siRNAs. Transfection of U2OSE64b cells with each of these two siRNAs led to a sharp decrease in the corresponding mRNA level as estimated by semiquantitative RT-PCR (Fig. 3A) and was accompanied by a substantial decrease in the SRp55 protein concentration (Fig. 3B). SRp55-depleted cells were then tested for cell viability following mitomycin C treatment. With both siRNA species, p53-deficient U2OSE64b, but not U2OSE6AS, cells showed a significantly higher resistance to drug treatment in the absence of SRp55 expression (Fig. 3C). The same tendency was observed with the HCT116 cells. Depletion of SRp55 activity in HCT116 p53−/− cells resulted in an increased resistance to mitomycin C at concentrations between 2 and 10 μg/mL, whereas it did not affect the sensitivity of the cells with wild-type p53 (Fig. 3D).
Inhibition of SFRS6 gene expression increases resistance of U2OSE64b cells to mitomycin C. A, treatment of U2OSE64b cells with SFRS6-specific siRNA effectively decreases the SFRS6 mRNA level. The level of SFRS6 mRNA was determined by RT-PCR (top bands). The cofilin1 (CFL1) gene was used as a control. Lane 1, RNA isolated from U2OSE64b cells; lane 2, RNA isolated from U2OSE64b cells transfected with SFRS6 siRNA (Ambion); lane 3, RNA isolated from U2OSE64b cells transfected with SFRS6 siRNA (Santa Cruz Biotechnology). B, inhibition by siRNA leads to a decrease in the level of the SRp55 protein. Immunoblot of protein lysates isolated from control U2OSE64b cells (lane 1), cells transfected with SFRS6 siRNA (Ambion; lane 2), and cells transfected with SFRS6 siRNA (Santa Cruz Biotechnology; lane 3) and treated with mitomycin C. Arrow, the SRp55 protein band. β-Actin antibodies were used for re-blotting to verify the uniformity of sample loading. C, inhibition of SFRS6 expression protects U2OSE64b, but not U2OSE6AS cells from cell death induced by mitomycin C treatment. Both cell lines were transfected with either control siRNA or siRNA directed against SRp55, then incubated with 5 μg/mL of mitomycin C for 24 h. Viable cells were quantified by the MTT assay. Columns, means from triplicate measurements; bars, SD. * and **, P = 0.95 and P = 0.99 confidence levels, respectively. D, inhibition of SFRS6 expression protects HCT116 p53−/−, but not HCT116 p53+/+ cells against cell death induced by mitomycin C treatment. Both cell lines were transfected with either control siRNA or siRNA directed against SRp55 (Santa Cruz Biotechnologies), then incubated with 0, 2, 4, and 10 μg/mL of mitomycin C for 24 h. Viable cells were quantified by the MTT assay. Columns, means from triplicate measurements; bars, SD.
Inhibition of SFRS6 gene expression increases resistance of U2OSE64b cells to mitomycin C. A, treatment of U2OSE64b cells with SFRS6-specific siRNA effectively decreases the SFRS6 mRNA level. The level of SFRS6 mRNA was determined by RT-PCR (top bands). The cofilin1 (CFL1) gene was used as a control. Lane 1, RNA isolated from U2OSE64b cells; lane 2, RNA isolated from U2OSE64b cells transfected with SFRS6 siRNA (Ambion); lane 3, RNA isolated from U2OSE64b cells transfected with SFRS6 siRNA (Santa Cruz Biotechnology). B, inhibition by siRNA leads to a decrease in the level of the SRp55 protein. Immunoblot of protein lysates isolated from control U2OSE64b cells (lane 1), cells transfected with SFRS6 siRNA (Ambion; lane 2), and cells transfected with SFRS6 siRNA (Santa Cruz Biotechnology; lane 3) and treated with mitomycin C. Arrow, the SRp55 protein band. β-Actin antibodies were used for re-blotting to verify the uniformity of sample loading. C, inhibition of SFRS6 expression protects U2OSE64b, but not U2OSE6AS cells from cell death induced by mitomycin C treatment. Both cell lines were transfected with either control siRNA or siRNA directed against SRp55, then incubated with 5 μg/mL of mitomycin C for 24 h. Viable cells were quantified by the MTT assay. Columns, means from triplicate measurements; bars, SD. * and **, P = 0.95 and P = 0.99 confidence levels, respectively. D, inhibition of SFRS6 expression protects HCT116 p53−/−, but not HCT116 p53+/+ cells against cell death induced by mitomycin C treatment. Both cell lines were transfected with either control siRNA or siRNA directed against SRp55 (Santa Cruz Biotechnologies), then incubated with 0, 2, 4, and 10 μg/mL of mitomycin C for 24 h. Viable cells were quantified by the MTT assay. Columns, means from triplicate measurements; bars, SD.
Mitomycin C treatment of U2OSE64b cells leads to alternative splicing of the CD44 gene. In spite of substantial progress in our understanding of alternative splicing regulation, only a few genes have been identified as direct targets of splicing activation (33). CD44, a cell surface glycoprotein with more than 30 identified splice forms, has 10 variable exons, which are located between constant exons 5 and 6 (Fig. 4A). The splicing pattern of inclusion/exclusion of variable exons has been shown to be sensitive to the activities of several splicing factors including 9G8, ASF/SF2, and SRp20 (34). To test whether CD44 was sensitive to the activity of the SRp55 splicing factor, we compared the expressions of several known splice forms of this gene in control and SRp55-depleted U2OS cells using combinations of forward and reverse primers that were able to recognize several CD44 isoforms (Fig. 4A). This comparison showed that isoforms containing v7 and v10 exons were very sensitive to the absence of SRp55 activity (Fig. 4B,, top). Monitoring of CD44 splicing following mitomycin C treatment of U2OSE64b cells revealed that the predominant standard CD44 isoform that lacks all variable exons (S11) as well as splice variants with the v6 exon were up-regulated after 5 and 10 h of treatment, whereas levels of splice variants with the v7 and v10 exons were lower after 5 h of exposure to the drug (Fig. 4B,, bottom). Interestingly, this time period coincided with the increased alternative splicing activity detected by in vivo measurement (Fig. 2D).
CD44 splicing patterns change during the early response of U2OSE64b cells to mitomycin C in a SRp55-dependent manner. A, structure of the variable part of the CD44 receptor gene. Rectangles, constant exons 5 and 6; hexagons, variable exons. Arrows, the primers used for RT-PCR (15). B, top, RNA isolated from cells treated with control (negative) siRNA (C) or with siRNA specific to SRp55 (si) were analyzed using primers specific to the constant exon S11 and the variable exons v6, v7, and v10. Bottom, expression of CD44 isoforms in U2OSE64b cells following mitomycin C treatment (2 μg/mL) for the indicated times. C, top, comparison of splice patterns of CD44 expression in U2OS and HCT116 cells. RNA isolated either from U2OS cells (U), or from HCT116 cells (H), were subjected to RT-PCR analysis using the CD44-specific primers S11, v6, v17, and V10. Bottom, RT-PCR analysis of RNA samples isolated from HCT116 p53−/− cells treated with mitomycin C (2 μg/mL) for 0 and 5 h.
CD44 splicing patterns change during the early response of U2OSE64b cells to mitomycin C in a SRp55-dependent manner. A, structure of the variable part of the CD44 receptor gene. Rectangles, constant exons 5 and 6; hexagons, variable exons. Arrows, the primers used for RT-PCR (15). B, top, RNA isolated from cells treated with control (negative) siRNA (C) or with siRNA specific to SRp55 (si) were analyzed using primers specific to the constant exon S11 and the variable exons v6, v7, and v10. Bottom, expression of CD44 isoforms in U2OSE64b cells following mitomycin C treatment (2 μg/mL) for the indicated times. C, top, comparison of splice patterns of CD44 expression in U2OS and HCT116 cells. RNA isolated either from U2OS cells (U), or from HCT116 cells (H), were subjected to RT-PCR analysis using the CD44-specific primers S11, v6, v17, and V10. Bottom, RT-PCR analysis of RNA samples isolated from HCT116 p53−/− cells treated with mitomycin C (2 μg/mL) for 0 and 5 h.
The set of CD44 isoforms presented by the HCT116 cells differs in several ways from that seen in the U2OS cells (Fig. 4C,, top). Variants that incorporate exon v7 sequences were completely absent in HCT116 cells; the major isoform containing the v6 exon in HCT116 cells belonged to the type 15 variant (545 bp band) compared with the 188 bp major band in U2OS cells (compatible with variant types 6, 11, and 14); and for exon 10, the major HCT116-specific variant was the epithelial variant type 5 (396 bp band), whereas the lower 204 bp band corresponding to the type 2 and 3 variants was more abundant in U2OS cells (nomenclature of CD44 isoforms is according to ref. 35). Monitoring CD44 splicing following mitomycin C treatment of HCT116 p53−/− cells revealed that as in the case of U2OSE64b, this pattern changed at the time of increased alternative splicing activity. We found that after 5 h of treatment, isoforms that incorporate v6 are up-regulated whereas v10-containing variants become less abundant (Fig. 4C , bottom), showing a tendency similar to that observed in U2OSE64b cells.
Discussion
The overall aim of most chemotherapeutic and radiotherapeutic approaches to cancer treatment is to induce apoptosis in tumor cells. The p53-mediated pathway is clearly an important mechanism for DNA damage–initiated apoptotic signaling, yet the correspondence between the p53 status of tumors and their response to such therapies is far from perfect. For this reason, and because more than half of human tumors lack functional p53, it is essential to understand p53-independent pathways to apoptosis and how they are triggered. One system in which this understanding is particularly critical includes those tumors (primarily cervical carcinomas) that express HPV16 E6. E6 is best known for its ability to degrade p53, although our laboratory and others have shown multiple additional functions (see ref. 36 and references therein). The aim of this present study was to evaluate the early response to genotoxic stress when the p53 pathway is compromised by either the absence of p53 or by the presence of HPV16 E6.
We compared the genotoxic stress response in two pairs of cell lines. The first pair consisted of human osteosarcoma U2OS cells that stably expressed E6, U2OSE64b, or its antisense construct, U2OSE6AS, which was used as a control. To determine whether the changes in response were due the effects of E6 on p53 stability or to some other E6 activity, we employed a second pair of cell lines, HCT116, expressing wild-type p53 and its isogenic derivative HCT116 p53−/− (13). Immunoblot analysis of the U2OS-derived cells treated with mitomycin C showed that, as expected, the presence of E6 caused a sharp decrease in both baseline and induced levels of p53, (Fig. 1B), and in the HCT116 cells, a complete absence of p53 expression was detected in the null mutant (Fig. 2B). Survival experiments showed that expression of E6 in the U2OS cell lines did not alter sensitivity to treatment with mitomycin C, suggesting that U2OSE64b cells were able to effectively use p53-independent apoptotic pathways (Fig. 1A).
To examine changes at an early stage of the genotoxic stress response, we chose the 5 h exposure to 2 μg/mL of mitomycin C for expression analysis because CHK1 activation was already observed by this time, although the treated cells had not yet experienced PARP-1 cleavage, a hallmark of apoptosis (Fig. 1D). Consistent with the idea that an early response was being observed in our analysis, we found the global changes in transcription to be modest, both with respect to the number of identified genes with altered expression and with the magnitude of the differences.
Comparative global expression profiling confirmed the idea that p53 depletion by E6 expression significantly changes the set of genes engaged in the early response to DNA damage (Tables 1 and 2). Intriguingly, almost half of the identified up-regulated genes in U2OSE64b cells were genes with unidentified functions, suggesting the involvement of new and uncharacterized pathways and mechanisms in these cells. One of the most striking and unexpected observations of the microarray analysis was the induction of the SFRS6 and SFRS7 genes in E6-expressing cells. These genes code for proteins that belong to a family of 16 currently known human SR proteins which play a pivotal role in both constitutive and alternative regulation of RNA splicing (31). Immunoblotting using a pan-SR antibody (32) not only confirmed the up-regulation of the SRp55 protein, encoded by the SFRS6 gene, but also revealed the up-regulation of two other splicing factors, SRp30 and SRp40, in U2OS-derived cells (Fig. 2A). SRp55 was also observed to be up-regulated in HCT116 p53−/− cells (Fig. 2B). In addition, the observed increase of up to 50% in alternative splicing activity for the U2OS-derived cells (Fig. 2D), as well as a significant up-regulation in the HCT116 cells, shows that these changes in transcript and protein levels have a functional consequence. It is of interest to note that the profile of SR protein expression in U2OSE64b cells coincides with that of alternative splicing activity, suggesting that alternative splicing activation may be due to up-regulation of SR protein expression. The fact that similar responses were seen in the U2OS- and HCT116-derived cell lines indicates that the observed transient increase of alternative splicing activity was not specific to U2OS cells. It also points out that the response is dependent on the p53 status of a cell, and that it is therefore most likely the E6-p53 interaction that causes the increase in alternative splicing, rather than some other activity of E6.
A growing number of new findings including those involving MDM2 and MDM4 (37), SRPK1 (38), and TAF1 (39) show not only the existence of connections between the DNA damage response and changes in splicing, but also that these connections are functional. The results from this present study add another example to the literature regarding alternative splicing regulation, by showing a clear connection between DNA damage, splicing factor induction, and activation of alternative splicing.
The hypothesis that the activation of alternative splicing activity during the early response results in altered splicing patterns of target genes was verified by our identification of the CD44 gene as a target for which the splicing profiles changed dramatically within the first 5 h of drug treatment. CD44, a cell surface glycoprotein involved in cell-cell and cell-matrix interactions, is an example of a gene with multiple splice variants that are thought to affect events connected to cell proliferation and survival as well as adhesion and motility (40).
We found that CD44 isoforms containing v7 and v10 exons were highly responsive to the depletion of SRp55 activity in U2OS cells (Fig. 4B). Silencing of the SRp55 gene promotes an increase in the v7 360 bp band, which is compatible with CD44 variant types 7 (pmeta-2), 11, and 14 (35), and this variant decreases during the first 5 h of mitomycin treatment of U2OSE64b cells. Interestingly, the disappearance of this CD44v7 isoform after 5 h of treatment, and its reappearance after 10 h, coincides with the profile of SRp55 expression in these mitomycin C–treated cells, implying that its activity may be directly involved in v7-specific splicing regulation. Comparison of CD44 expression in U2OS and HCT116 cells revealed that its splicing patterns were cell type–specific. CD44 isoforms containing the v7 exon, which are induced in SRp55-depleted U2OS cells, are practically absent in HCT116, whereas the pattern of HCT116-specific isoforms detected by the v10 primer was very similar to that observed in U2OS cells following SRp55 depletion (Fig. 4B and C). Nevertheless, in spite of different baseline splicing patterns, the tendency of mitomycin C treatment to up-regulate the v6 isoforms and down-regulate the v10 isoforms is the same in both p53-deficient cell lines (Fig. 4A and C).
Finally, experiments using siRNA to inhibit SFRS6 gene expression show that the contributions of this SR protein to alternative splicing activity modify the genotoxic stress response in p53-deficient cells. Depletion of the SRp55 splicing factor in both U2OSE64b and HCT116 p53−/− cells resulted in a significant increase in cell viability following treatment with the genotoxic drug (Fig. 3C). This increase is particularly remarkable given the fact that the up-regulation of alternative splicing activity was time-dependent and was observed only within the first 8 h of treatment (Fig. 2D). These data suggest that in the absence of a p53-mediated response, splicing can be heavily involved in the modulation of cellular responses.
Alternative splicing is being increasingly recognized as a fundamental mechanism for the generation of protein diversity and as an important source of variability and complexity. It provides an additional mechanism for genes to be differentially expressed throughout development, and the loss of splicing fidelity, which often occurs during tumorigenesis, can result in the production of aberrant and alternative splice isoforms (9). It is worth noting that more than half of all human tumors lack wild-type p53, and the observed influence of p53 status on alternative splicing activity regulation in response to low doses of DNA-damaging agents may contribute to an increased loss of alternative splicing regulation in tumors. In this context, it is worth noting that p53 status has been linked to the presence or absence of aberrant transcripts in cancer cells (41, 42). The results of this report, together with observations described previously (37, 39), show the involvement of alternative splicing activity in modulating the response to genotoxic stress and provide new insights into the functions of alternative splicing.
Acknowledgments
Grant support: National Cancer Institute grant R01 CA095461 from the NIH and by a Seed Grant from the Loma Linda University School of Medicine.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Dr. J. Zhou (Department of Medicine and Program in Neuroscience, University of Massachusetts Medical School, Worcester, MA) for the LucM14 construct and Dr. B. Vogelstein (The Howard Hughes Medical Institute and the Ludwig Center for Cancer Genetics and Therapeutics, Johns Hopkins Kimmel Comprehensive Cancer Center, Baltimore, MD) for the IICT116 cell lines.