SFPQ, a multifunctional nuclear protein, regulates the transcription of PDE3A

Phosphodiesterase 3A (PDE3A), a member of the cGMP-inhibited cyclic nucleotide phosphodiesterase (PDE) family, plays important roles in oocyte maturation and vascular smooth muscle cell proliferation. However, the molecular mechanisms that regulate PDE3A gene expression remain largely unknown. In the present study, we investigated the transcriptional regulation of PDE3A, and found that the splicing factor proline- and glutamine-rich (SFPQ) protein modulated PDE3A mRNA levels. Multiple transcription start sites (TSS1, 2, and 3) were identified within the first exon of PDE3A using 5′-rapid amplification of cDNA ends (RACE). Variable expression levels of three PDE3A variants were also observed in human tissues and HeLa cells. Several putative SFPQ-binding sites were identified upstream of the regulatory region of PDE3A-TSSs using ChIP sequencing (ChIP-seq). Serum-induced PDE3A expression was affected by increasing the amount of SFPQ binding to the upstream regulatory region of PDE3A. In addition, transcription of PDE3A was lower in human cervical adenocarcinoma cells compared with normal cervical tissue. Furthermore, overexpression of PDE3A induced sensitivity to anticancer therapeutic agent, 6-(4-(diethylamino)-3-nitrophenyl)-5-methyl-4,5-dihydropyridazin-3(2H)-one (DNMDP), in HeLa cells. Taken together, these results suggest that SFPQ functions as a transcriptional activator of PDE3A, which is involved in the regulation of DNMDP sensitivity, offering a novel molecular target for the development of anticancer therapies.


Introduction
To maintain the normal cellular physiology, tight regulation of cAMP and cGMP signaling is pivotal, causing serious diseases if disrupted [1,2]. By catalyzing the hydrolysis of cAMP and cGMP, cyclic nucleotide phosphodiesterases (PDEs) regulate their own intracellular concentrations and properly modulate important downstream signaling pathways [1,2]. The PDE superfamily consists of 11 gene families of isoenzymes (PDEs 1-11). Over 100 PDE isoforms are reportedly expressed based upon the analysis of multiple ORFs, promoters, and alternative splicing patterns. PDEs regulate discrete aspects of intracellular signaling through spatial and functional diversification, and their dysfunction affects numerous clinically relevant pathways linked to human diseases [1,3,4]. The localization of individual PDEs to specific intracellular sites or molecular complexes helps regulate the compartmentation of cyclic nucleotide signaling [3,5].
The gene expression cascade from chromatin to mRNA translation to protein decay is a dynamic and complex process in mammalian cells [13,14]. Most mammalian genomes carry fewer genes than other eukaryotes, which indicate that complex mechanisms are likely to exist in mammals to increase the number of functional isoforms originating from a single gene [15]. In mammalian cells, pre-mRNA alternative splicing mechanisms, such as alternative promoter selection, are common ways to generate diversity of gene products [15,16]. Pre-mRNA alternative splicing changes the structure, intracellular localization, and binding partner interactions of proteins [16,17]. A previous study of human myocardial PDE3A [18] demonstrated that two isoforms of PDE3A (hPDE3A) are generated. Specifically, hPDE3A-isoform 1 (NM 000921.4) is the longer transcript, while hPDE3A-isoform 2 (NM 001244683.1) has a truncated 5 -UTR and 5 -coding region. Three hPDE3A protein variants (hPDE3A1/2/3) generated by hPDE3A-isoform 1 have been suggested to have different subcellular localizations in human myocardial cells and distinct roles in selective phosphorylation of protein targets [19]. Recently, it has been reported that splicing and degradation of PDE transcripts can be regulated by the p54 nrb /NONO nuclear protein, which is involved in transcription, splicing, and RNA export [20]. However, the mechanisms regulating the transcription and translation of different PDE3A isoforms, in addition to the function of each variant, are poorly understood.
In the present study, we characterized the pre-mRNA alternative promoter and splicing patterns of PDE3A variants to gain insights into the mechanisms regulating the expression of different isoforms of this gene. Specifically, we examined the mRNA levels of different isoforms in various human tissues and cells, and characterized the SFPQ-mediated transcriptional regulation of PDE3A gene expression. We found that hPDE3A expression was reduced in cervical cancer and the expression of hPDE3A increases the DNMDP sensitivity of cervical cancer cells.

Cell culture and transfection
The human cervical carcinoma HeLa cell line was maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% heat-inactivated FBS (Gibco, NY, U.S.A.) at 37 • C in a humid incubator with 5% CO 2 . Cells were transfected with plasmids and siRNAs using Lipofectamine 2000 (ThermoFisher Scientific, MA, U.S.A.) according to the manufacturer's instructions.

Plasmids, siRNAs, and reagents
The expression constructs for the PDE3A variants TSS1, TSS2, and TSS3 were obtained from cDNA of HeLa cells by PCR amplification. They were introduced into a pcDNA3 vector. The following forward PCR primers were used:

RNA preparation, RT-PCR, and quantitative PCR
Total RNA was isolated from cultured cells using the RNeasy Mini Kit (Qiagen, Hilden, Germany) and reverse transcribed with the SuperScript III Reverse Transcriptase (Thermo Fisher Scientific, MA, U.S.A.) using random hexamers. Reverse-transcription quantitative PCR (RT-qPCR) was performed using FastStart Universal SYBR Green Master Premix (Roche, CA, U.S.A.) and NEXpro TM qRT-PCR Master Premix (Genes Laboratories, Korea). The following primers were used: PDE3A, 5 -GAC AGC GAT GAG TCA GGA GA-3 and 5 -TCT GAA GAG TGC GAC TGA GG-3 ; GAPDH, 5 -CCA TGG AGA AGG CTG GGG-3 and 5 -CAA AGT TGT CAT GGA TGA CC-3 . RT-qPCR was performed using the Applied Biosystems 7900HT Detection System. The specificity of the products generated with each set of primers was examined using agarose gel electrophoresis and confirmed by a melting curve analysis.

Preparation of protein extracts and immunoblot analysis
Whole cell extracts were prepared using M-PER protein extraction reagent (Thermo Fisher Scientific) supplemented with a protease inhibitor cocktail (Roche, CA, U.S.A.). Extracted proteins were denatured and reduced by treatments with SDS and 2-mercaptoethanol, respectively. Proteins were then separated on 4-20% SDS-polyacrylamide gels (Bio-Rad, CA, U.S.A.) and transferred on to nitrocellulose membranes (Thermo Fisher Scientific, MA, U.S.A.). Immunoblot analysis was performed as described recently [21]. Anti-NONO, anti-SFPQ, and anti-ACTIN antibodies were purchased from Santa Cruz Biotechnology (TX, U.S.A.). Anti-PDE3A antibody was used as described previously [22].

ChIP-PCR and electrophoretic mobility shift assay
ChIP was performed using reagents from Upstate Biotechnology Inc. (Lake Placid, NY) according to the manufacturer's protocol. Immunoprecipitated DNA and input DNA were subjected to ChIP-PCR using primer sets that amplified the P1 and P2 regions of the PDE3A promoter, including the SFPQ-binding sites. Primer sequences were as follows: P1 forward, 5 -TTA TTT CTC TGA AAG AAG CAA GG-3 ; P1 reverse, 5 -CTG GGT GGT TGT ACC ACC TTC TAG-3 ; P2 forward, 5 -GGC ACC AGA CAT GTG GAA AC-3 ; P2 reverse: 5 -GA ATC TTT GTA ATT TAC CTG TTG-3 . The SFPQ protein was synthesized in vitro from pCS3 + MT construct using the TNT Coupled Reticulocyte Lysate System for electrophoretic mobility shift assay (EMSA) [16]. ChIP-PCR product of the P1 region was end-labeled using [γ 32 P]ATP and T4 polynucleotide kinase. EMSA was performed as described previously [23].

Cell viability after DNMDP treatment
HeLa cells were seeded at 5000 cells/well in 96-well plates. After incubating overnight, HeLa cells were transfected with PDE3A expression constructs for the TSS1, TSS2, and TSS3 variants. Twenty-four hours after transfection, cells were

PDE3A mRNA expression decreased in SFPQ-depleted cells
While investigating the molecular targets of SFPQ using RNA-seq, PDE3A expression was found to be significantly decreased following SFPQ depletion ( Figure 1A). To confirm this RNA-seq observation, we measured mRNA expression by RT-qPCR after SFPQ depletion or NonO depletion. In agreement with the RNA-seq results, SFPQ depletion decreased PDE3A expression after 24 and 48 h ( Figure 1B). No significant change in the mRNA levels of PDE3A was observed after NonO depletion at 24 h, while the mRNA levels of PDE3A were slightly decreased after NonO depletion at 48 h. Immunoblot analysis confirmed that siRNA-mediated depletion efficiently reduced PDE3A expression ( Figure 1C). These results suggest a role for SFPQ in the regulation of the PDE3A gene.
SFPQ contains regions rich in arginine/glycine and proline/glutamine close to the N-terminus [16]. It also possesses two RNA recognition motifs located in the C-terminus. SFPQ is a multifunctional nuclear protein that has been implicated in various aspects of the regulation of gene expression, including RNA splicing and transcription [24]. Thus, we investigated the molecular mechanisms of regulation of SFPQ-mediated PDE3A gene expression.
After treatment with the transcriptional inhibitor actinomycin D, a reduction in PDE3A protein levels was observed after 9 h (Figure 2A). Therefore, we measured PDE3A mRNA decay rate after depletion of SFPQ or NonO. There was no significant change in PDE3A mRNA decay rate following SFPQ or NonO depletion, suggesting that SFPQ transcriptionally regulates PDE3A ( Figure 2B). To evaluate whether the decrease in PDE3A protein levels was accompanied by a relative change in the amount of SFPQ, PDE3A expression was measured by immunoblot analysis after either SFPQ or NonO depletion. A significant decrease in PDE3A expression was detected in SFPQ-depleted cells, but not in NonO-depleted cells ( Figure 2C). These results suggest that SFPQ has a role in transcriptional regulation of PDE3A.

Identification of the 5 -flanking region of human PDE3A
To investigate the molecular mechanisms of SFPQ-mediated PDE3A transcription, 5 -RACE was carried out to identify a transcription start site (TSS) for human PDE3A using poly-A RNAs prepared from the heart and cervix. In heart, we found that PDE3A contains three TSSs within the first exon, each of which has a different level of gene expression ( Figure 3A,B). Fetal and cardiomyopathy tissues showed stronger expression of TSS1 and TSS2 compared with adult normal cardiac tissues. Additionally, cervix expressed two TSSs, and displayed higher expression of TSS1 in both normal and cancer tissues ( Figure 3C). We speculated that each TSS uses different translation start codons ( Figure 3D).

Binding of SFPQ to the PDE3A promoter
For finding the expression level changes of PDE3A and the regulatory mechanism of SFPQ, we examined the changes of PDE3A expression under different concentrations of FBS. Since FBS provides growth-promoting factors in cell culture and affects the protein expression levels, the changes in PDE3A protein expression were examined by immunoblot analysis after three passages in 1 and 10% FBS treatment. FBS treatment (1%) decreased PDE3A protein levels, but did not affect SFPQ or NonO expression ( Figure 4A,B). This suggests that PDE3A expression is regulated by more than the levels of SFPQ protein.
We hypothesized that SFPQ binding to the PDE3A promoter would be critical for its regulation. Thus, we investigated putative SFPQ binding sites in the PDE3A promoter using ChIP-seq. Three peaks of SFPQ binding within the PDE3A promoter were found at chr12:20673425-20673995 and chr12:20599778-20600539 for P1 and   Figure 4C). To confirm this, ChIP-PCR was carried out. We were unable to detect SFPQ binding at chr12:20673425-20673995 (results not shown). However, SFPQ strongly bound to the P1 region, and 1% FBS treatment reduced this binding ( Figure 4D). The binding affinity of SFPQ to the P2 region was weaker than that to the P1 region in both samples. Further to the less binding of SFPQ to the P2 region was observed in 1% FBS-treated sample than 10% FBS-treated sample. In addition, we explored a physical interaction of SFPQ with P1 region using EMSA. Two slowly migrating complexes were formed with the SFPQ protein. Supershifted band was observed with an addition of anti-SFPQ antibody, suggesting that protein bound to P1 region is indeed SFPQ ( Figure 4E). These results suggested that SFPQ directly binds to the PDE3A promoter and is regulated by serum concentration in the medium. Furthermore, SFPQ is critical for PDE3 gene expression.

PDE3A conferred sensitivity to DNMDP
A recent report showed that PDE3A affects the sensitivity of cancer cell lines to the anticancer therapeutic agent DNMDP [8]. To investigate whether PDE3A gene expression is altered in cervical cancer tissues, the expression levels of PDE3A mRNA were examined using RT-qPCR. The results showed that the relative expression of PDE3A mRNA in cervical cancer tissues was significantly lower compared with normal cervical tissues ( Figure 5A). Next, we examined the effects of PDE3A on DNMDP sensitivity in HeLa cells by measuring cell viability after overexpression of PDE3A. The expression of each and all three PDE3A variants sensitized DNMDP-treated cells, suggesting that the expression of PDE3A indeed regulates DNMDP sensitivity in cervical cancer cells ( Figure 5B).

Discussion
In the present study, we identified SFPQ as a transcriptional regulator of the PDE3A gene. We demonstrated that SFPQ binding to the PDE3A promoter is essential for PDE3A expression and the binding of SFPQ is regulated by serum concentration. Serum contains a number of macromolecular proteins, low molecular weight nutrients, and other compounds necessary for cell growth such as hormones and attachment factors. The factors involved in the regulation of SFPQ binding to the PDE3A promoter remains to be elucidated. We also found that PDE3A expression is decreased in cervical cancer tissues compared with healthy tissues. In addition, we uncovered a role for PDE3A in DNMDP-mediated cell death.
The use of multiple promoters and TSSs is a common means of gene regulation; however, considerable variation and complexity in the patterns of alternative TSS usage exist. Alternative promoter usage can influence gene expression in numerous ways [25]. The overall amount of transcriptional initiation can vary amongst alternative promoters. The turnover or translation efficiency of mRNA variants with different leader exons can vary. Alternative promoters can have cell-type specificity, and can lead to the generation of protein isoforms differing at the N-terminus.
Based on our results, the three TSSs for PDE3A could theoretically generate three different N-termini. The first exon of PDE3A encodes a transmembrane domain that is critical for the protein function [26]. Thus, diverse N-termini in the transmembrane domain could be produced by alternative TSS usage. There is no obvious difference in the expression of the three PDE3A variants or in how they modulate DNMDP sensitivity in cervical cancer cells ( Figure 5B). This suggests that the transmembrane domain of PDE3A is not responsible for regulating DNMDP sensitivity. We observed a similar pattern in PDE3A TSS expression between cervical tissues and HeLa cells (results not shown). Although two TSSs for PDE3A were identified in HeLa cells, immunoblotting with an anti-PDE3A antibody showed mainly one band at approximately 100 kDa. However, determining the mRNA variant of PDE3A and the TSS that produced the 100 kDa protein requires further study.
SFPQ is a ubiquitous and multifunctional nuclear protein essential for life in vertebrates [27,28]. SFPQ is a splicing factor associated with the polypyrimidine tract-binding protein [29]. It is implicated in various nuclear functions including splicing, RNA transport, DNA repair, and transcriptional regulation [24]. SFPQ binds DNA and RNA. It binds to promoters of a number of genes to regulate transcription [30][31][32]. Interestingly, the SFPQ protein expression level is often reduced in many types of cancer tissues when compared with healthy tissues (http://www.proteinatlas. org). In addition, SFPQ interacts with long non-coding RNAs in several cancer tissues [30,[33][34][35]. However, there are no studies of the molecular function of SFPQ in cancer tissues. In this report, we found that serum concentration regulated SFPQ binding to the PDE3A promoter and resulted in decreased PDE3A levels without any change in SFPQ expression. Therefore, we speculate that allosteric modulation of SFPQ amongst other factors may affect its DNA-binding capacity in cancer cells, and changes in SFPQ expression are a critical means of regulating PDE3A expression.
The present study has uncovered a mechanism for PDE3A regulation. Recent reports suggested the potential role of PDE3A in modulating the effectiveness of the therapeutic agent DNMDP in cancer cells. Interestingly, we found that the expression level of PDE3A is lower in cervical cancer tissues compared with healthy control tissues ( Figure  5A). In addition, our observation that PDE3A overexpression induced sensitivity to DNMDP in cervical cancer cells is similar to a recent report [8]. Further study is needed to understand how PDE3A influences DNMDP sensitivity.