Introduction
Proteasome activator 28γ (PA28γ) is a component of the proteasome system, which is one of the most important proteolytic systems in eukaryotes. It binds to the 20S core and mainly promotes proteins degradation in a ubiquitin-and ATP-independent manner (1–4). Strikingly, PA28γ not only plays a role as a proteasome activator, but also has been shown to be involved in cancer progression and virus infection (Fig. 1B). Loss of PA28γ expression in PA28γ-deficient mice results in reduced body size and cell-specific mitotic defects (5–7). It directs degradation of the steroid receptor coactivator SRC-3, which is an oncogene frequently amplified in breast cancer (8). It is also a novel serum marker for human colorectal cancer (CRC) because it can be detected in sera and is significantly elevated in CRC patients compared with healthy donors and patients with benign bowel disease (9). Moreover, overexpression of PA28γ occurs in many different cancer types, including thyroid, colon, liver, lung, ovary and gastric cancer (10–14). PA28γ binds to and regulates the stability and nuclear retention of hepatitis C core protein, contributing to hepatitis C core protein-induced insulin resistance and hepatocarcinoma (15–18). In addition, PA28γ promotes coxsackievirus B3 (CVB3) replication (19–21) and interacts with human T-lymphotropic virus type 1 (HTLV-1) p30 to increase viral spread (22,23). A series of cell cycle, apoptosis, and cancer progression-related PA28γ target proteins have been identified, including p21, p16, p19, p53, Mdm2 (24–26). In recent years, more targets of PA28γ have been identified using antibody array analysis. These proteins include protein kinase A catalytic subuit-α (PKAca), SirT1, and casein kinase (CK)1δ, which play important roles in angiogenesis, hepatic lipid metabolism and premature aging, respectively (27–29).
Alternative splicing of mRNA allows many gene products with different functions to be produced from a single coding sequence according to the cell type, developmental stage, or in response to acute stimuli. It explains how enormous mammalian proteomic diversity can be achieved with the limited number of genes found in higher eukaryotes. Based on deep sequencing of alternative splicing complexity in the human transcriptome, it is estimated that transcripts from ~95% of multiexon genes undergo alternative splicing (30,31). A study involving probabilistic analyses indicated that >60% of human disease-causing mutations affect splicing rather than directly affecting coding sequences. Another study concluded that one-third of all hereditary diseases are likely to have a splicing component (32,33). Abnormally spliced mRNAs are also found in a high proportion of cancerous cells. Combined RNA-Seq and proteomics analyses technology revealed striking divergent expression profile of splice isoforms of key proteins in important cancer pathways. For example, several abnormally spliced DNMT3B mRNAs are found in tumors and cancer cell lines. In two separate studies, expression of two of these abnormally spliced mRNAs in mammalian cells caused changes in the DNA methylation patterns in those cells. Cells with one of the abnormal mRNAs also grew twice as fast as control cells, indicating a direct contribution to tumor development by this product (34–37).
In a previous study, we identified 85 differentially and constantly expressed proteins (>2-fold change, P<0.05) in six pairs of oral leukoplakia tissues with dysplasia and oral squamous cancer tissues via two dimensional electrophoresis (2-DE) followed by ESI-Q-TOF-LC-MS/MS. Among them, three homologs of proteasome activator PA28α, PA28β, and PA28γ were shown to have upregulated mRNA levels in oral squamous cell carcinoma (OSCC) cells relative to oral keratinocytes (38). In this study, we analyzed alternative splicing of PA28γ using the alternative splicing prediction data base ASPicDB (39,40), and found that there are theoretically nearly 50 alternative transcripts. We therefore tried cloning these predicted splice variants. We successfully cloned a novel (the fifth) transcript variant of PA28γ in oral cancer cells. This variant encodes a truncated isoform that retains the most conserved residues of the PA28 family. Furthermore, it is involved in regulation of p53 and Mdm2.
Materials and methods
Bioinformatics prediction
The alternative splicing prediction data base (ASPicDB) is a program designed to provide access to reliable annotations of the alternative splicing pattern of human genes, and to the functional annotation of predicted isoforms. Alternative splicing prediction of PA28γ (PSME3) was performed in ASPicDB.
Cell type and cell culture
Six oral squamous cell carcinomaderived cell lines (HSC-3, HOK, UM1, UM2, Cal27 and HN31) and human embryonic kidney 293 (HEK293) cells were cultured in DMEM with 10% fetal bovine serum.
Gene cloning and DNA sequencing
Total RNA of each cell line was extracted with TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol. The RNA concentration was determined by absorbance at 260 nm with a NanovueTM spectrophotometer (GE Healthcare). Equal amounts of RNA (1.0 μg) were used as template in each reverse transcription reaction (total volume, 30 μl) with the PrimeScript® RT Reagent kit with gDNA Eraser (Takara, Shiga, Japan). The reaction conditions were 42°C for 2 min to remove genomic DNA, 37°C for 15 min, followed by 85°C for 5 sec to obtain total cDNA. The primers designed for PA28γ amplification were: 5′-TTGTATTTCCAGGGCATGGCCTCGTTGCTG-3′ (forward primer) and 5′-CAAGCTTCGTCATCATCAGTA CAGAGTCTC-3′ (reverse primer). PCR was performed for 30 cycles with PrimeSTAR® HS DNA Polymerase and 2 μl total cDNA was used as template in 50 μl reaction volume. The reaction conditions were a denaturation step at 98°C for 10 sec, an annealing step at 55°C for 15 sec, and an elongation step at 72°C for 1 min. PCR product (10 μl) was separated on a 1% agarose gel, and observed under ultraviolet light (Bio-Rad, Hercules, CA, USA). DNA bands were eluted from the agarose gel using a gel extraction kit (Doupson) and cloned into the p15TV-L vector for sequencing.
Plasmid transfection
Plasmids for overexpressing human influenza hemagglutinin A epitope (HA)-tagged PA28γ isoforms were constructed by Chengdu Bio-atom Biotechnology. HEK293 cells were used as experimental cells. The control and positive group cells were transiently transfected with empty vector and HA-PA28γ isoform 5-expressing plasmids, respectively.
Western blotting
Seventy-two hours after treatment, cells were harvested and protein was extracted with RIPA lysis buffer (Beyotime, P0013B). Antibodies to GAPDH (XP® Rabbit mAb, Cell Signaling Technology), p53 (mouse mAb, Cell Signaling Technology), Mdm2 (Phospho-Mdm2 (Ser166) Antibody, Cell Signaling Technology), PA28γ isoform1 (Purified mouse anti-PA28γ, BD Transduction Laboratories), PA28γ isoform 5 (HA-tag antibody, ZSGB-BIO) were used to test the amount of corresponding protein by western blotting (Bio-Rad electrophoresis).
Quantitative PCR
The amount of mRNA of PA28γ transcript variant 1 was measured by the ABI 7500 Real-time PCR (RT-PCR) System with One Step SYBR® PrimeScript™ Plus RT-PCR kit (Takara, RR096A). Gene-specific primers were designed as follows: PA28γ transcript variant 1 forward primer, 5′-ATGGACTGGATGGTCCCACT-3′; PA28γ transcript variant 1 reverse primer, 5′-ACAGCCGGATCTCAGGTTTC-3′; 18S rRNA forward primer, 5′-CTACCACATCCAAGGAA GGCA-3′; 18S rRNA reverse primer, 5′-TTTTTCGTCACTA CCTCCCCG-3′. Gene-specific primers for PA28γ transcript variant 1 were designed crossing the lost sequences which lack in PA28γ transcript variant 5. Mean Ct values for target genes were normalized to mean Ct values for the endogenous control 18S [-ΔCt = Ct (18S)- Ct (target gene)]. The ratio of mRNA expression of target gene versus 18S was defined as 2−ΔCt. All experiments were repeated at least three times.
Results
Alternative splicing prediction of PA28γ via bioinformatics
The genomic location of human PA28γ is at chrosome 17: 40985423–40995777, with a size of 10,354 bp. Expression sequence tags (EST) (1,180) exist in this cluster. Forty-nine alternative transcripts are predicted and 45 of them are protein-coding forms. To date, only four alternative transcripts are reported in GenBank. PA28γ transcript variant 1 (Nucleotide accession: NM_005789.3) encodes the predominant protein with 254 amino acids. Variant 2 (Nucleotide accession: NM_176863.2) uses an alternative in-frame splice site; variant 2 is 13 amino acids longer than variant 1. Variant 3 (Nucleotide accession: NM_001267045.1) is distinguished in the 5′-untranslated region and 5′-coding region, and initiates translation at an alternative start codon. Variant 3 has a distinct N-terminus and is 12 amino acids longer than variant 1. Variant 4 (Nucleotide accession: NR_049772.1) is a long non-coding RNA.
Cloning and identification of a novel PA28γ transcript variant in oral cancer cells
As shown in Fig. 2A, three DNA bands were observed on the agarose gel. According to the DNA markers, from top to bottom, the first band likely corresponds to PA28γ transcript variant 1, and the third band to oligonucleotide primers. In addition, we observed a weak band between the first and third bands. The sequencing data were analyzed by the BLAST tool in NCBI. As expected, the first band was confirmed as PA28γ transcript variant 1. We translated the nucleotide acid sequence of the second band into amino acids and a structure-based sequence alignment was performed. As shown in Fig. 3, the new isoform lacks the ‘homolog-specific insert’ region (41), but it belongs to the PA28γ subfamily according to the high identity in the C-terminal sequence. Therefore, we termed the new alternative splicing as transcript variant 5 and submitted it to GenBank (Nucleotide accession: JX156303.1). Compared with PA28γ variant 1, PA28γ variant 5 lacks the nucleotide acids regions in coding sequence among 185–483 in PA28γ variant 1, which corresponds to the sequence from exon 4 to exon 7 in variant 1 (Fig. 2B). Taken together, these findings indicate that PA28γ transcript variant 5 is a novel transcript variant in the PA28γ subfamily.
PA28γ isoform 5 involved in p53 and Mdm2 regulation
We predicted that PA28γ transcript variant 5 encodes a truncated protein of 170 residues, which is 84 residues shorter than PA28γ isoform1. However, structure-based sequence alignment showed that isoform 5 retains most of the conserved residues and the ‘activation loop’ of PA28 family (Fig. 3). As mentioned in the introduction, PA28γ has a series of target proteins and plays an important role in cell cycle regulation. It was reported that PA28γ regulates p53 by enhancing Mdm2-mediated degradation (25). This prompted us to explore whether PA28γ isoform 5 could play a similar role in this process. As shown in Fig. 4, the positive group overexpressed PA28γ isoform 5, which resulted in significant decrease of p53 and Mdm2 at protein level. Furthermore, we tested the amount of PA28γ isoform1 at both mRNA and protein levels. Strikingly, the mRNA level of PA28γ isoform1 in the positive group was 4.5-fold higher than in the control group. The change in protein levels of PA28γ isoform1 between positive and negative groups was not as great as for mRNA levels, but still increased. To explain these changes, we propose two possibilities: i) increased PA28γ isoform 5 triggers the transcription of PA28γ alternative transcript1, and the increased PA28γ isoform1 regulates the degradation of p53 and Mdm2, as reported before; or ii) overexpressed PA28γ isoform 5 maintains a pool of functional PA28γ, and represses the translation of PA28γ alternative transcript1 as feedback regulation. In either case, the downregulation of p53 and Mdm2 at the protein level resulted from overexpression of PA28γ isoform 5, which indicates that PA28γ isoform 5 is a functional isoform that may play a complementary role in regulating p53 and Mdm2.
Discussion
Based on the results of bioinformatics analysis via ASPicDB, 30 alternative transcripts of PA28α, one alternative transcripts of PA28β, and 49 alternative transcripts of PA28γ were predicted (39,40). We checked the alternative transcripts of these PA28 members deposited in GenBank, and found that four transcript variants of PA28α, one transcript variant of PA28β and four transcript variants of PA28γ have been reported. Therefore, the number of reported transcript variants of PA28β is consistent with the bioinformatics prediction. However, there are many more alternative transcripts of PA28α and PA28γ that are not yet defined. Typically, alternatively spliced transcripts have been found by comparing expression sequence tags (ESTs), but this requires sequencing of very large numbers of ESTs. Most EST libraries come from a very limited number of tissues, so tissue-specific splice variants are likely to be missed in any case. However, high-throughput approaches to examine splicing have been developed, such as DNA microarray-based analyses, RNA-binding assays, and deep sequencing. When combined with splicing assays, including in vivo reporter gene assays, the functional effects of polymorphisms or mutations on the splicing of pre-mRNA transcripts can be analyzed (33,42–44). Herein, utilizing bioinformatics, RT-PCR technology and gene overexpression in model cells, we successfully predicted, cloned, and confirmed a novel transcript variant of PA28γ. Moreover, we found PA28γ isoform 5 lacks long regions compared with other reported isoforms of PA28γ. Therefore, the siRNA target to this region of PA28γ will affect other reported isoforms of PA28γ but not on PA28γ transcript variant 5, which may result in a significant off-target effect of siRNA interference.
Generally speaking, the diversity of transcript variant corresponds to the functional significance of the gene regardless of whether the transcript variant has the same function or not. The variants with the same function will compensate the loss-of-function mutation to maintain homeostasis, whereas patterns of divergent functional variants are likely to be indicators for some disease. In a retrospective analysis of 80 individuals with gastrointestinal sarcoma, predominant expression of the immunosuppressive NKp30c isoform (over the immunostimulatory NKp30a and NKp30b isoforms) was associated with reduced survival of the subjects (45). Previously, we identified three alternative transcripts of oral cancer overexpressed 1 gene (ORAOV1), and an inverse correlation was found between the expression frequency of ORAOV1-A and the degree of differentiation in OSCC (46). In future studies, we will assess the clinical value of PA28γ isoform 5 in tissue samples from normal, precancerous to infiltrative OSCC. Moreover, the relationships of immunostaning with survival rate and recurrence will be analyzed.
Several studies have evaluated the interaction of PA28γ variant 1, p53 and Mdm2. P53 encodes a tumor suppressor protein and responds to diverse cellular stresses to regulate expression of target genes, thereby inducing cell cycle arrest, apoptosis, senescence, DNA repair, or changes in metabolism. Mutations in this gene are associated with a variety of human cancers. Thus, it is a focus of numerous investigations for reversing tumor progression. MDM2 encodes a nuclear-localized E3 ubiquitin ligase. The protein can promote tumor formation by targeting tumor suppressor proteins, such as p53, for proteasomal degradation. Amplification of this locus is detected in a variety of different cancers. Noteworthy, the polymer form of PA28γ variant 1 interacts with both p53 and Mdm2, which facilitates ubiquitination and Mdm2-dependent proteasomal degradation of p53. The decreased p53 attenuated apoptosis stimulation after DNA damage (47,48). In our study, we confirmed that PA28γ isoform 5 also mediates the downregulation of p53 and Mdm2, which may serve as a complementary mechanism for PA28γ isoform1 in regulating p53 and Mdm2. Given that PA28γ is involved in cancer progression, it suggests to examine the PA28γ isoform 5 before we determine PA28γ isoform1 as a therapy target via regulating p53 and Mdm2 pathway.
The proteasome is a primary proteolytic system in eukaryotes. This system is a multi-subunit protease complex composed of 20S catalytic core and proteasome activators (Fig. 1A). The 20S core is a cylindrical stack of four heptameric rings with two outer α rings and two inner β rings. The PA700 (19S) activator binds to the 20S core and primarily mediates degradation of ubiquitinated proteins in ATP-dependent manner. In contrast, the PA28 (11S) activator binds to the 20S core and mainly promotes protein degradation in Ub- and ATP-independent manner (1–4). To date, three classes of PA28 have been identified: PA28α, PA28β, and PA28γ. PA28α and β form a heteroheptamer, which is mainly localized in the cytosol. PA28γ exists as a homoheptamer and is primarily found in the nucleus. PA28α and β mediate proteolytic cleavage after basic, acidic, and hydrophobic residues. PA28γ stimulates proteasomal hydrolysis of peptides with basic residues (41,49–56). Thus, the subcellular location of PA28γ isoform 5 and its effect on proteolytic activity in the proteasome system need to be further studied.
In this study, we made an attempt to transiently transfect HA-tagged PA28γ isoform 5 overexpressing plasmids into oral cancer cells. However, we failed because oral cancer cells are highly keratinized. Although we determined the primary function of PA28γ isoform 5 in HEK293 cells, further investigation of its effects on cell cycle, apoptosis and its correlation with progression of oral squamous cell carcinoma will be performed in oral cancer cells using lentivirus transfection.