p54^nrb/NONO Regulates Cyclic AMP-Dependent Glucocorticoid Production by Modulating Phosphodiesterase mRNA Splicing and Degradation

Abstract

Glucocorticoid production in the adrenal cortex is activated in response to an increase in cyclic AMP (cAMP) signaling. The nuclear protein p54^nrb/NONO belongs to the Drosophila behavior/human splicing (DBHS) family and has been implicated in several nuclear processes, including transcription, splicing, and RNA export. We previously identified p54^nrb/NONO as a component of a protein complex that regulates the transcription of CYP17A1, a gene required for glucocorticoid production. Based on the multiple mechanisms by which p54^nrb/NONO has been shown to control gene expression and the ability of the protein to be recruited to the CYP17A1 promoter, we sought to further define the molecular mechanism by which p54^nrb/NONO confers optimal cortisol production. We show here that silencing p54^nrb/NONO expression in H295R human adrenocortical cells decreases the ability of the cells to increase intracellular cAMP production and subsequent cortisol biosynthesis in response to adrenocorticotropin hormone (ACTH) stimulation. Interestingly, the expression of multiple phosphodiesterase (PDE) isoforms, including PDE2A, PDE3A, PDE3B, PDE4A, PDE4D, and PDE11A, was induced in p54^nrb/NONO knockdown cells. Investigation of the mechanism by which silencing of p54^nrb/NONO led to increased expression of select PDE isoforms revealed that p54^nrb/NONO regulates the splicing of a subset of PDE isoforms. Importantly, we also identify a role for p54^nrb/NONO in regulating the stability of PDE transcripts by facilitating the interaction between the exoribonuclease XRN2 and select PDE transcripts. In summary, we report that p54^nrb/NONO modulates cAMP-dependent signaling, and ultimately cAMP-stimulated glucocorticoid biosynthesis by regulating the splicing and degradation of PDE transcripts.

INTRODUCTION

The non-POU domain-containing octamer-binding protein p54^nrb/NONO is a member of the Drosophila behavior/human splicing (DBHS) family of RNA-binding proteins and is comprised of two tandem RNA recognition motif (RRM) domains followed by a charged protein-protein interaction module (1–3). DBHS proteins have been described as multifunctional nuclear proteins, with p54^nrb/NONO being implicated in numerous nuclear processes, including transcription regulation, mRNA splicing, nuclear retention, and subnuclear body formation (4–7). In addition to binding to DNA elements such as intracisternal A particles (8), p54^nrb/NONO is also known to promote the binding of other transcription factors to their response elements (9). The polypyrimidine tract-binding protein-associated splicing factor (PSF)/p54^nrb/NONO heterodimer regulates the transcriptional activity of several nuclear receptors (NRs), including the thyroid hormone receptor (10) and the androgen receptor (11). This heterodimeric complex also promotes nuclear retention of hyperedited RNAs by binding to the inner nuclear matrix structural protein matrin 3 (12, 13) and stimulates topoisomerase I activity (14). A heterodimer of p54^nrb/NONO and paraspeckle protein 1 has been shown to be targeted to paraspeckles in an RNA-dependent manner (15), and p54^nrb/NONO has been shown to be necessary for cyclic AMP (cAMP)-dependent activation of cAMP response element binding protein (CREB) target genes in vivo (16). Finally, the presence of variants of p54^nrb/NONO in human breast cancer (17) and interactions with transcription factors such as PU.1 in erythroleukemia (18) link p54^nrb/NONO to the etiology of pathophysiological processes.

Adrenocorticotropin hormone (ACTH) stimulates glucocorticoid biosynthesis by inducing the expression of steroidogenic genes, via a signaling pathway that is mediated by the activation of adenylyl cyclase, increased intracellular cAMP, and the activation of cAMP-dependent protein kinase (PKA) (19, 20). We have shown previously that p54^nrb/NONO can regulate the transcription of CYP17A1 (21, 22), a steroidogenic enzyme that functions at a key branch point in biosynthesis of glucocorticoids and adrenal androgens.

The cyclic nucleotide phosphodiesterases (PDEs) play an integral role in modulating cellular signaling by degrading the second messenger cAMP, an intracellular activator of steroidogenesis. Mammalian PDEs are encoded by 21 different genes and are subdivided into 11 families (PDE1 to -11), with more than 90 alternative splice variants generated from multiple alternative transcriptional start sites and alternative splicing (23, 24). Because of their fundamental role in degrading cAMP (and cGMP), PDEs have been linked with varied endocrine pathophysiologies. Indeed, a whole-genome association study identified that cAMP-selective PDEs are involved in adrenocortical tumorigenesis (25).

Given that p54^nrb/NONO modulates several nuclear processes, coupled with the role of the protein in regulating CYP17A1 gene expression, we generated a stable cell line where p54^nrb/NONO was stably silenced in order to further explore the role of this multifunctional protein in steroidogenesis and adrenocortical function. Significantly, we observed a global reduction in steroidogenic capacity by silencing p54^nrb/NONO, which was caused by altered PDE expression. Repression of p54^nrb/NONO induced expression of multiple spliced PDE isoforms. Furthermore, p54^nrb/NONO was found to interact with exoribonuclease XRN2 to regulate the mRNA turnover of select PDE transcripts.

MATERIALS AND METHODS

Reagents.

Dibutyryl cAMP (Bt₂cAMP), 3-isobutyl-1-methylxanthine (IBMX), and actinomycin D were obtained from Sigma-Aldrich (St. Louis, MO). ACTH [1-39] was purchased from American Peptide, Inc. (Sunnyvale, CA) and BC11-38 from Tocris Bioscience (Minneapolis, MN).

Cell culture.

H295R adrenocortical cells were generously donated by William E. Rainey (University of Michigan, Ann Arbor, MI) and cultured in Dulbecco modified Eagle medium (DMEM)–F-12 medium (Mediatech, Inc., Manassas, VA) supplemented with 10% Nu-Serum I (BD Biosciences, Palo Alto, CA), 1% ITS Plus (BD Biosciences), and antibiotics.

Generation of the H295R p54^nrb/NONO^KD stable cell line.

p54^nrb/NONO knockdown (p54^nrb/NONO^KD) cell lines were generated by transfecting H295R cells using GeneJuice (EMD Bioscience) with short hairpin RNA (shRNA) plasmids (in the pGFP-V-RS HuSH vector [Origene, Rockville, MD]) containing the oligonucleotides shRNA 1 (5′-TGG CAT TCG GCA GCC AAT AGA ATC TAA GA) and shRNA 4 (5′-TCT GAC AGT AGC TCT TAG ACT CGC CTA TC). Stable clones were selected using 10 μg/ml puromycin (Mediatech, Inc.). For rescue experiments, p54^nrb/NONO^KD cells were transfected with either the pCR3.1 vector or pCR3.1-HA-p54^nrb (provided by Philip W. Tucker, University of Texas, Austin, TX). Cells were then treated with 0.4 mM dibutyryl cAMP and then harvested for analysis of RNA and protein expression.

Cortisol and DHEA assays.

H295R wild-type (WT) or p54^nrb/NONO^KD cells were subcultured into 12-well plates and treated with 0.4 mM Bt₂cAMP, 100 μM IBMX, or 10 μM BC11-38 for 48 h before medium collection. Cortisol and DHEA released into the medium were determined in triplicate against cortisol or DHEA standards made up in DMEM–F-12 medium using a 96-well plate DHEA assay or a cortisol enzyme-linked immunosorbent assay (ELISA) (Assay Designs, Inc., Ann Arbor, MI). Steroid hormone amounts were normalized to the total cellular protein content, as determined using the bicinchoninic acid (BCA) protein assay (Pierce, Rockford, IL).

Microarray analyses.

H295R WT or p54^nrb/NONO^KD cells were subcultured into 100-mm dishes and treated with 0.4 mM Bt₂cAMP for 18 h. Total RNA was isolated using the RNeasy kit (Qiagen, Valencia, CA), and gene expression profiling was done by Phalanx Biotech Group, Inc. (Palo Alto, CA) using the human whole-genome OneArray DNA microarray (HOA_004). Scatter plotting, volcano plotting, and hierarchical clustering visualize significant changes in signal between conditions, significant fold changes between conditions, and groups of genes based on their expression levels, respectively. Differentially expressed genes were subjected to gene ontology (GO) enrichment analysis using the controlled vocabulary of the Gene Ontology Consortium (54). The gene sets are based on GO terms and their association with human genes. Shown are the number of genes and the ratio of genes changed in each set for the top classes of gene sets (Molecular Function or Biological Process) when comparing mRNA expression in the p54^nrb/NONO^KD cell line to that in the wild-type H295R cell line. The significantly enriched GO terms of the target gene sets were detected compared to the genome-wide background (selection criteria, log₂ ratio of ≥1.0 or ≤−1.0 and P < 0.05).

Wild-type or p54^nrb/NONO^KD cells were subcultured into 100-mm dishes. Total RNA was isolated using an RNeasy minikit (Qiagen), and quality assessment was conducted using an RNA 6000 Nano lab chip (Bioanalyzer; Agilent) and with a NanoDrop spectrophotometer (Thermo). Affymetrix human transcriptome array 2.0 arrays were hybridized according to Affymetrix recommendations using the Ambion WT protocol (Life Technologies) and Affymetrix labeling and hybridization kits. One hundred nanograms of total RNA was processed in parallel with an external MAQC A RNA to control robustness of data. Transcriptomic array data set analysis was performed with Transcriptome Analysis Console (TAC) software. One-way analysis of variance (ANOVA) was performed to compare gene intensities between WT and p54^nrb/NONO^KD cells. Genes were considered significantly differentially expressed when the fold change (linear) was <−2 or >2 and the ANOVA P value (condition pair) was <0.05.

Alternative splicing analysis.

Analysis at the splicing level was also performed with TAC software. The criteria for the splicing index (SI) were as follows: (i) a transcript cluster gene is expressed under both conditions, and (ii) a probe selection region (PSR) or junction can be analyzed by SI if it expresses in at least one condition. Normalized intensities were compared using one-way ANOVA for the PSRs and junctions within a gene. After running ANOVA, multitesting correction was performed using the Benjamini-Hochberg step-up false-discovery rate (FDR)-controlling procedure for all the expressed genes and expressed PSRs/junctions (expressed under at least one condition). Results were considered significantly different when the SI (linear) was <−2 or >2 and the FDR P value was ≤0.05. Partek Genomic Suite software was also used to perform the alternative splicing analysis.

Quantification of intracellular cAMP.

WT or p54^nrb/NONO^KD H295R cells were cultured into 6-well plates and treated with 50 nM ACTH, 100 μM IBMX, or 10 μM BC11-38 for 10 min. Cells were lysed in 0.1 M HCl for 20 min at room temperature, collected by scraping, and centrifuged at 4,000 rpm for 10 min. The supernatant was collected, acetylated, and used in a direct cyclic AMP correlate-EIA (Assay Designs, Inc.) following the manufacturer's instructions. Data were normalized to the total protein amount of each sample and are expressed as picomoles per microgram of total protein.

Coimmunoprecipitation.

H295R cell lysates were isolated and precleared by incubation for 30 min with rabbit IgG (Millipore, Temecula, CA) and protein A/G Plus-agarose (Santa Cruz Biotechnology). The precleared lysates were immunoprecipitated overnight at 4°C using 5 μg of anti-p54^nrb/NONO antibody (Millipore, Temecula, CA) and protein A/G Plus-agarose. The agarose beads were washed three times with radioimmunoprecipitation assay (RIPA) buffer and three times with phosphate-buffered saline (PBS) and then subjected to SDS-polyacrylamide gel electrophoresis (PAGE) and Western blotting using anti-XRN2 antibody.

Western blotting.

H295R WT or p54^nrb/NONO^KD cell lysates were collected in RIPA buffer (1× PBS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 1× protease inhibitor cocktail set I [EMD Biosciences]). Cells were then lysed by sonication (one 2-s burst) followed by incubation on ice for 30 min. Lysates were centrifuged at 12,000 rpm for 15 min at 4°C and the supernatant collected for analysis by SDS-PAGE. Aliquots of each sample (30 μg of protein) were resolved on 8% SDS-polyacrylamide gels and transferred to polyvinylidene difluoride (PVDF) membranes (Millipore, Temecula, CA). Blots were probed with the antibodies listed in Table 1, and expression was detected using an ECF Western blotting reagent kit (GE Healthcare, Piscataway, NJ) and visualized by scanning the blots on a VersaDoc 4000 imager (Bio-Rad, Hercules, CA). Protein amounts were determined using the BCA protein assay (Pierce, Rockford, IL).

TABLE 1.

List of antibodies

Target protein	Source (catalog no., vendor)
p54nrb/NONO	05-950, Millipore
MC2R	sc-13107, Santa Cruz
MRAP	ab103319, Abcam
P450scc (CYP11A1)	Donated by Michael R. Waterman, Vanderbilt University
P450c11B (CYP11B1/2)	sc-28205, Santa Cruz
P450c17 (CYP17A1)	sc-66849, Santa Cruz
P450c21B (CYP21A2)	sc-134566, Santa Cruz
STAR	sc-28205, Santa Cruz
SF-1	07-618, Millipore
NUR77	SAB2101650, Sigma-Aldrich
NURR1	N6538, Sigma-Aldrich
NOR1	sc-100906, Santa Cruz
XRN2	A301-103A, Bethyl
PDE11A	ab154273, Abcam
GAPDH	sc-25778, Santa Cruz

RNA isolation and real-time RT-PCR.

H295R WT or p54^nrb/NONO^KD cells were subcultured into 6-well plates. Total RNA was isolated using Isol-RNA lysis reagent (5 Prime, Inc., Gaithersburg, MD). Nuclear and cytoplasmic RNAs were isolated using the Ambion Paris system (Life Technologies). RNA was then reversed transcribed and amplified using a one-step SYBR green reverse transcription-PCR (RT-PCR) kit (Thermo Scientific, Inc., Waltham, MA) and the primer sets listed in Table 2. Gene expression was normalized to β-actin (ACTB) mRNA content and calculated using the ΔΔC_T method.

TABLE 2.

Real time RT-PCR primer sequences

Gene symbol	Primer (5′→3′)
	Forward	Reverse
ACTB	ACGGCTCCGGCATGTGCAAG	TGACGATGCCGTGCTGCATG
GAPDH	GTCTCCTCTGACTTCAACAGCG	ACCACCCTGTTGCTGTAGCCAA
p54nrb/NONO	ATCAACATCACCACCAGCAG	AGTGATGTCGGGAGGAAGAT
MC2R	GTGGTGCTTACGGTCATCTGGA	AGGCACAGGATGAAGACCAGCA
MRAP	TTCGTGGTGCTGCTCTTCCTCA	CAGGCACTTCTGGATGCAGAGG
CYP11A1	CGTGGAGTCGGTTTATGTC	CTCTGGTAATACTGGTGATAGG
CYP11B1	TAGACTGGGTTGCTGACAAAC	GAATGGAACTGGCGTCCTTAT
CYP17A1	CTCTTGCTGCTTCACCTA	TCAAGGAGATGACATTGGTT
CYP21A2	TGTGGAACTGGTGGAAGC	GGTGGAGCCTGTAGATGG
STAR	GCTCTCTACTCGGTTCTC	GCTGACTCTCCTTCTTCC
NR4A1	GGACAACGCTTCATGCCAGCAT	CCTTGTTAGCCAGGCAGATGTAC
NR4A2	AAACTGCCCAGTGGACAAGCGT	GCTCTTCGGTTTCGAGGGCAAA
NR4A3	ACTGCCCAGTAGACAAGAGACG	GTTTGGAAGGCAGACGACCTCT
NR5A1	TGCAGAATGGCCGACCAG	TGGCGGTAGATGTGGTC
NR5A2	TACCGACAAGTGGTACATGGAA	CGGCTTGTGATGCTATTATGGA
PDE1B	CGTTCAGTGCTGGAGAATCACC	CCAACACCATCTCAATGACCAGG
PDE1C	TCGCTGGACAATGTCACTCCTG	GTGACTGAGCAACCATAGTGGAC
PDE2A	GCTGGTGAACAAGATCAATGGGC	GCTGCGATACTGAGCCTCATTC
PDE3A	AGAGCCTCTGAGGAAAGCATCG	GCTGCTCCATAATTGAGTCCAGG
PDE3B	AAAGTTCGAGACTTGCATTTGA	AATGGACTGATGGGCAGAC
PDE4A	CTGCGACATCTTCCAGAACCTC	GCTGGTCACTTTCTTGGTCTCC
PDE4B	TAGTCAGCCTCCTGTCTCCAGA	GAAGCCATCTCACTGACAGACC
PDE4C	AGGTCACTACCACGCCAATGTG	CAGCCAGGATTTCCAAGTCTGTG
PDE4D	GGACACTTTGGAGGACAATCGTG	CCTTTTCCGTGTCTGACTCAC
PDE7A	TCAGGCTTATTCTCACATCTGCC	TTCTGGCGACTGATGTCTGTGG
PDE7B	CAGCCTGGTAACACTGTTGTGC	GTGATACGGGTTTTGGCTGTGG
PDE8A	GCTGATGTGCTTCATGCCACTG	CACAGGAAGGAGTTGGTTCTCC
PDE8B	GAGATTTTACGGACCACAGAACTG	TCCTGACAGTCCTTCTCAAGCCG
PDE10A	GAAGAGTGGCAAGGTGTCATGC	ATGTCCCACAGGACCGATGAAC
PDE11A	GTAGGTGGCTTTTGACAGTGAGG	CTCGGTCAGAATGTCTTGAAACC
PDE4A1	CGTCCTTCCATGGCACT	GGCCAGGAGATGGTGTC
PDE4A2	GACGCCTTCGGCTTCTC	GCCAGGAGATGGTGTCG
PDE4A3	CCTTCAGCGATGAGGACAC	GTGATGGGATCTCCACTTCATTC
PDE4A4	CTCTAAGCCTTGGCTGGTG	CCTGCTCATTTCTGACAGGT
PDE4A5	CGCTAGGTTGGCGAGATG	GTTGGCACGGTCGGATT
PDE3A1	CAAGTCCCATCGGAGGAC	CTCGTTTGTGGTCCCATTCT
PDE3A2	CTCATTACTGATTATACATGGTGACG	CGGCGATGTCCACAGTAATA
PDE2A1	TCCTCAAGCCGGACGAG	CTCCAGTCGGTCGTCTCTAC
PDE2A3	CATGCGGCCACTCCATC	TCGTCCGGCTTGAGGAA
PDE4D1	GGGTATGGCAGGATGGC	TTTCTGGTAGGCCTCCTCT
PDE4D2	GAAGGAGCAGCCCTCAT	GTCCAGACACCAGTCCAG
PDE4D3	GACAGAAGATCTGCGAACATGA	TCATGGGATCCAAGGGACT
PDE4D4	ATGTCCTGGCCCTCCTC	TCATGGGATCCAAGGGACT
PDE4D5	GCGCAGAATGCTTCTCA	CGCAGATGTGCCATTGT
PDE4D6	ACTATTTACTGTCAGTGTCTTGGG	CCGACTCATTTCAGAGAGATGG
PDE4D8	GTGCCAGGACCATCTACAA	GTGCCATTGTCCACATCAAA
PDE4D9	AGATCCCGATCTACAAGTTCCC	GACTCCGTCCCGCAGAT
PDE11A1	GACATGCATGGCTTGGTTT	CATCAGCACTGCACTTTGG
PDE11A2	ATGAAGGGCTTTGAGTGTAGG	AGCGTTAGATATGGCGATTCC
PDE11A3	AGGTGCTCTTTCTGGATCG	AAGGTCTTCTTGCCTGCTT
PDE11A4	GGGAGCATGGAGAAACGG	TGTCTTGTATCCAGTTAGCTTGTC
LDLR	ACGGTGGAGATAGTGACAATG	AGACGAGGAGCACGATGG
LIPE	CACTACAAACGCAACGAGAC	CCAGAGACGATAGCACTTCC

Isolation of poly(A)⁺ RNA in cellular fractions.

Poly(A)⁺-enriched RNA was isolated from total RNA using oligo(dT) resin in the form of Dyna l beads according to the kit protocol (PolyA Spin mRNA isolation kit; New England BioLabs, Beverly, MA). Nuclear and cytoplasmic RNA samples were extracted by using a Paris kit (Life Technologies) according to the manufacturer's protocol.

RNA immunoprecipitation.

RIPA was performed according to the kit protocol (Magna RIP RNA-binding protein immunoprecipitation kit; Millipore, Billerica, MA). Briefly, cells were lysed in RIP lysis buffer. Antibodies against p54^nrb/NONO, XRN2, or mouse IgG were used for RIPA. Target RNAs were isolated and reverse transcribed using a one-step SYBR green RT-PCR kit (Thermo Scientific, Inc., Waltham, MA). Specific primer pairs were used to detect CYP17A1, PDE2A, PDE3A, PDE3B, PDE4A, PDE4D, PDE11A, ACTB, and GAPDH (glyceraldehyde-3-phosphate dehydrogenase).

Stability of mRNA.

Cells were treated with actinomycin D (5 μg/ml), and the mRNA level in select genes after normalization over the time period was then analyzed by the one-phase exponential decay of nonlinear regression with GraphPad Prism version 5.0 (GraphPad Software, Inc., San Diego, CA).

Statistical analysis.

One-way ANOVA, Tukey-Kramer multiple comparison, and unpaired Student t tests were performed using GraphPad Prism version 5.0. Significant differences from a compared value were defined as a P value of <0.05. Hierarchical clustering analysis of DNA microarray studies was done by performing observation and variable-tree computation using complete linkage clustering and correlation distance matrix with robust center-scale normalization.

RESULTS

Silencing p54^nrb/NONO leads to aberrant steroid hormone production.

We have shown previously that p54^nrb/NONO can regulate the transcription of CYP17A1 in H295R cells (22); however, the precise mechanism by which this nuclear protein regulates steroidogenic capacity is unknown. Thus, we examined the effect of silencing p54^nrb/NONO expression in H295R cells. As shown in Fig. 1A, p54^nrb/NONO mRNA and protein levels were reduced by 73% and 81%, respectively. Microarray analysis identified 351 genes that were upregulated, while 163 genes were downregulated in p54^nrb/NONO^KD cells (Fig. 1B). GO analysis revealed that many of these altered genes are involved in cellular processes, such as steroid metabolism, amino acid metabolism, nucleotide PDE activity, and lipid metabolic processes (Fig. 1C). To further validate the role of p54^nrb/NONO in regulating steroidogenesis, we next measured the levels of cortisol and dehydroepiandrosterone (DHEA). While observing a reduction in the secretion of DHEA and cortisol in unstimulated p54^nrb/NONO^KD cells, we found that silencing p54^nrb/NONO diminished the effect of the cAMP analogue dibutyryl cAMP (Bt₂cAMP) on steroid hormone secretion (Fig. 2A).

Fig 1.

Global gene expression analyses of the effect of silencing p54^nrb/NONO. (A) The H295R p54^nrb/NONO^KD cell line, which stably harbors p54^nrb/NONO-targeting shRNA, was established. Total RNA and proteins were isolated, and the mRNA and protein levels of p54^nrb/NONO were quantified by qRT-PCR and Western blotting, respectively. **, P < 0.01. (B) Microarray data comparing global gene expression profiles of WT and p54^nrb/NONO^KD cells. Upper left panel, scatter plot presenting the distribution of mRNA expressions from WT (x axis) and p54^nrb/NONO^KD (y axis) cells. Upper right panel, volcano plot comparison of gene expression patterns between WT and p54^nrb/NONO^KD cells. The x axis indicates the differential expression profiles, plotting the fold induction ratios on a log₂ scale. The y axis indicates the statistical significance of the difference in expression (P value from a t test) on a log₁₀ scale. Lower left panel, hierarchical cluster heat map revealing global molecular fingerprinting performed on WT and p54^nrb/NONO^KD cells. Horizontal stripes represent genes, and columns show experimental samples. Relative increased and decreased expression are depicted in red and green, respectively. Lower right panel, number of genes differentially altered after silencing p54^nrb/NONO. (C) Multilevel GO analysis for the most regulated gene categories of their corresponding Biological Process and Molecular Function terms within the microarray data set. The vertical axis represents the GO terminology. The upper lateral axis represents the number of the related genes within each category. The lower lateral axis represents the proportion of the related genes within each category.

FIG 2.

Ablated cAMP-responsive DHEA and cortisol production and steroidogenic gene expression by silencing p54^nrb/NONO. (A) Total DHEA and cortisol levels in cells treated with 0.4 mM Bt₂cAMP were quantified by ELISA and normalized to the total cellular protein content of each sample. Asterisks indicate statistically significant differences from the untreated controls within each group. (B and C) Expression of indicated steroidogenic genes (B) and nuclear receptors (C) was analyzed after treatment with 0.4 mM Bt₂cAMP in H295R WT and p54^nrb/NONO^KD cells. The mRNA and protein levels for the indicated genes were determined by quantitative RT-PCR (qRT-PCR) and Western blotting. Data are graphed as mean ± standard error of the mean (SEM) from three separate experiments, each done in triplicate. *, P < 0.05; **, P < 0.01 (versus respective nontreatment control). #, P < 0.05 (versus WT control). (D) The mRNA levels for MC2R and MRAP were determined by qRT-PCR.

Repressing p54^nrb/NONO ablates cAMP-responsive steroidogenic gene expression.

To further investigate the effect of silencing of p54^nrb/NONO on cAMP-stimulated steroid hormone production, we next analyzed the expression of genes required for cortisol production. As shown in Fig. 2B, real-time RT-PCR data revealed that Bt₂cAMP was able to induce the expression of CYP17A1, CYP11A1, CYP11B1, CYP21A2, and StAR by 4.4-, 2.9-, 3.5-, 4.6-, and 3.2-fold in WT cells, respectively, but the inductive effect was suppressed in p54^nrb/NONO^KD cells. Analysis of protein levels by Western blotting yielded similar results (see the densitometric analysis in Fig. S1 in the supplemental material). Since the nuclear receptor (NR) steroidogenic factor 1 (SF1) plays a pivotal role in regulation of steroidogenic gene expression (26, 27) and binds to p54^nrb/NONO (21, 22), we then examined the expression of SF1 and found no effect of silencing p54^nrb/NONO on the expression of this receptor (Fig. 2C). Moreover, no significant effect of p54^nrb/NONO silencing on the basal expression of other nuclear receptors that regulate the expression of steroidogenic genes was found (28, 29), including NR4A1, NR4A2, and NR4A3. However, silencing of p54^nrb/NONO reduced cAMP-dependent induction of some of these immediate early genes (Fig. 2C).

As mentioned above, ACTH is the primary initiator of cAMP-mediated glucocorticoid biosynthesis. Thus, the lack of a significant increase in cAMP-stimulated steroidogenic gene expression prompted us to examine the effect of p54^nrb/NONO silencing on the expression of ACTH receptor (melanocortin 2 receptor [MC2R]) and MC2R accessory protein (MRAP). Notably, the expression of both MC2R and MRAP was increased in the p54^nrb/NONO^KD cell line, suggesting that the dampened cAMP-stimulated steroidogenesis in knockdown cells is not due to the reduction of the expression of these cell surface proteins (Fig. 2D).

Role of PDEs in the impaired cAMP response in p54^nrb/NONO^KD cells.

We next tested whether silencing of p54^nrb/NONO affects the ability of ACTH to increase intracellular cAMP. As shown in Fig. 3A, whereas WT cells exhibited a 1.7-fold increase in intracellular cAMP after an acute stimulation with ACTH, p54^nrb/NONO^KD cells were unresponsive but exhibited a 0.4-fold reduction in the basal cAMP level compared with WT cells under control conditions. Since PDEs control the degradation rates of the second messenger cAMP, we next treated cells with the nonselective PDE inhibitor 3-isobutyl-1-methylxanthine (IBMX) and found that IBMX increased cAMP production in p54^nrb/NONO^KD cells. Coadministration of IBMX and ACTH almost completely restored the capacity of p54^nrb/NONO^KD cells to generate cAMP in response to ACTH stimulation (Fig. 3A) and the ability to produce cortisol and DHEA in response to cAMP stimulation (Fig. 3B and C), suggesting that PDEs of the cAMP pathway are potential targets of p54^nrb/NONO in modulating cAMP-induced steroidogenic processes.

FIG 3.

IBMX rescues the cAMP response in p54^nrb/NONO^KD cells. (A) Intracellular cAMP levels in cells treated with 50 nM ACTH and/or 100 μM IBMX were quantified by ELISA. cAMP amounts are normalized to total cellular protein content, and the data graphed represent the mean ± SEM from three experiments, each done in triplicate. (B and C) Total cortisol (B) and DHEA (C) secreted into the media collected from cells treated with 0.4 mM Bt₂cAMP and/or 100 μM IBMX were quantified by ELISA and normalized to the total cellular protein content of each sample. *, P < 0.05; **, P < 0.01 (versus WT control). #, P < 0.05; ##, P < 0.01 (versus p54^KD control).

Elevated PDE expression in p54^nrb/NONO^KD cells.

As discussed above, mammalian PDEs are composed of 21 genes and are categorized into 11 families based on sequence homology, enzymatic properties, and sensitivity to inhibitors (23, 24). Thus, we assessed the expression of the PDE isoforms that degrade cAMP. Consistent with microarray studies (Fig. 1C), real-time RT-PCR analysis revealed increased expression of the PDE2A, PDE3A, PDE3B, PDE4A, PDE4D, and PDE11A isoforms in the p54^nrb/NONO^KD cell line (Fig. 4A). Given that PDEs are subjected to alternative splicing, coupled with the fact that p54^nrb/NONO is implicated in splicing (4), we next analyzed the expression levels of PDE splice variants. As shown in Fig. 4B to F, there was a marked induction in the expression of several variants in p54^nrb/NONO^KD cells compared to WT cells. While most of the PDE4A variants were increased in the knockdown cell line (Fig. 4B), the expression of PDE4D5 was increased, with most of the other variants exhibiting reduced expression when the expression of p54^nrb/NONO was silenced (Fig. 4E). Notably, the expression of PDE11A1 was increased by more than 50-fold in p54^nrb/NONO^KD cells (Fig. 4F).

FIG 4.

Silencing of p54^nrb/NONO induces PDE isoforms and splicing variants. (A) The mRNA levels of multiple PDE isoforms in H295R WT and p54^nrb/NONO^KD cells treated without or with 0.4 mM Bt₂cAMP for 24 h were quantified by qRT-PCR and normalized to the mRNA expression of β-actin. (B to F) mRNA expression of splicing variants for the indicated PDE isoforms in H295R WT and p54^nrb/NONO^KD cells treated without or with 0.4 mM Bt₂cAMP for 24 h were quantified by qRT-PCR. The data graphed represent the mean ± SEM from three experiments, each performed in triplicate. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

PDE transcript variants are alternatively spliced in p54^nrb/NONO^KD cells.

To further investigate a role for p54^nrb/NONO in splicing, we performed Affymetrix whole-transcriptome array analysis as described in Materials and Methods. This analysis revealed that 798 coding genes were expressed in WT cells but not in p54^nrb/NONO^KD cells, and 67 coding genes existed in p54^nrb/NONO^KD cells but not WT cells (Fig. 5A). Of note, we also found that 594 coding genes have at least one probe selection region (PSR) with a splicing index (SI) that was changed more than ∼2-fold. Significantly, PDE2A, PDE3B, and PDE11A are among the genes that are alternatively spliced.

FIG 5.

Analysis of alternative splicing by silencing of p54^nrb/NONO. (A) Alternative splicing analysis summary based on algorithm parameters applied during the analysis. Affymetrix human transcriptome array 2.0 was used to perform alternative splicing profiling in H295R WT and p54^nrb/NONO^KD cells. This was followed by global analysis of alternative splicing using Transcriptome Analysis Console (TAC) software. The group of genes listed below (798, 627, 67, and 55) represents the genes that are differentially expressed. (B) Structural views of variants PDE2A, PDE3B, and PDE11A by TAC software. All probe selection regions (PSRs) and junctions are represented in the structure view with boxes that have same size. An inclusion junction detects 2 neighboring PSRs. An exclusion junction detects PSRs that are apart from each other. A crossed-out box represents a PSR/junction that does not contain data. A diagonally crossed-out box represents a PSR/junction that is not expressed under at least one condition. A spacer represents a TC PSR where a selection of probes is not possible. (C) Gene views of PDE2A, PDE3B, and PDE11A by Partek Genomic Suite. Gene structures of splicing variants of specific genes are shown. Summarized exon expression levels are drawn below, with red lines representing KD cells and blue lines representing WT cells.

Visualization of exons and junctions in PDE2A, PDE3B, and PDE11A was shown in the splicing viewer of TAC (Fig. 5B). We found 13 exons/junctions in the PDE2A gene and 14 exons/junctions in PDE11A gene with an SI of greater than 2 (Table 3). Since there is only one splice variant in the current RefSeq database for PDE3B, we were unable to analyze the splicing pattern of this isoform. To gain more insight into the whole-transcriptome array data, we used the Partek Genomic Suite platform to analyze the expression level of each exon within a transcription cluster. Although parallel exon expression patterns were observed in the PDE3B gene in WT and p54^nrb/NONO^KD cells, we found elevated exon expression in the first exons of the 5′ ends of PDE2A2, PDE2A1, and PDE11A4 in p54^nrb/NONO^KD cells, further supporting that alternative splicing was increased when p54^nrb/NONO was suppressed (Fig. 5C).

TABLE 3.

Splicing indices in all PSRs/junctions in PDE2A, PDE3B, and PDE11A

Transcript cluster	Gene symbol	PSR/junction	Splicing index (linear)
TC11002059.hg.1	PDE2A	JUC11012999.hg.1	−2.77
		PSR11024624.hg.1	−2.78
		PSR11024612.hg.1	−2.91
		PSR11024595.hg.1	−2.96
		PSR11024635.hg.1	−3.02
		JUC11013015.hg.1	−3.08
		JUC11013019.hg.1	−3.59
		PSR11024633.hg.1	−3.69
		PSR11024629.hg.1	−3.70
		PSR11024634.hg.1	−3.83
		PSR11024632.hg.1	−3.86
		PSR11024638.hg.1	−4.28
		PSR11024639.hg.1	−5.16
TC11000214.hg.1	PDE3B	PSR11002859.hg.1	−2.57
TC02002570.hg.1	PDE11A	JUC02021461.hg.1	4.57
		PSR02040790.hg.1	2.94
		JUC02021467.hg.1	2.54
		PSR02040789.hg.1	2.05
		PSR02040763.hg.1	−2.84
		JUC02021456.hg.1	−4.21
		PSR02040800.hg.1	−4.38
		PSR02040807.hg.1	−4.39
		PSR02040762.hg.1	−4.45
		PSR02040802.hg.1	−4.55
		JUC02021482.hg.1	−4.63
		PSR02040808.hg.1	−6.35
		PSR02040773.hg.1	−7.41
		JUC02021477.hg.1	−8.24

To provide further evidence for the role of PDE11A in regulating cAMP-mediated steroidogenesis, we then treated cells with the selective PDE11A inhibitor BC11-38. As shown in Fig. 6, combining treatment of cells with BC11-38 and ACTH partially restored the ability of p54^nrb/NONO^KD cells to generate cAMP in response to ACTH stimulation and the capacity to produce cortisol and DHEA in response to cAMP stimulation, lending support for the role of increased PDE11A expression in mediating disrupted cAMP signaling in the p54^nrb/NONO^KD cell line.

FIG 6.

Partial rescue of the impaired cAMP response in p54^nrb/NONO^KD cells. (A) Intracellular cAMP levels in cells treated with 50 nM ACTH and/or 10 μM BC11-38 were quantified by ELISA. cAMP amounts are normalized to total cellular protein content, and the data graphed represent the mean ± SEM from three experiments, each done in triplicate. (B and C) Total DHEA and cortisol levels in cells treated with 0.4 mM Bt₂cAMP and/or 10 μM BC11-38 were quantified by ELISA and normalized to the total cellular protein content of each sample. *, P < 0.05; **, P < 0.01 (versus WT control). #, P < 0.05; ##, P < 0.01 (versus p54^KD control).

Rescue of p54^nrb/NONO expression restores PDE11A expression.

The role of PDE11A in adrenocortical steroid hormone biosynthesis and, more notably, adrenocortical dysfunction (30–32) prompted us to determine if reintroduction of p54^nrb/NONO into the knockdown cell line would restore the expression levels of PDE11A to those observed in the wild-type cell line. As shown in Fig. 7A, Bt₂cAMP decreases the expression of PDE11A in the wild-type H295R cell line, likely to ensure increased intracellular cAMP and activated PKA-dependent steroidogenesis. Bt₂cAMP also decreased the mRNA expression of PDE2A variants but not PDE4A variants in wild-type H295R cells (data not shown). Transfecting the p54^nrb/NONO^KD cell line with a vector expressing hemagglutinin (HA)-tagged p54^nrb/NONO reduced the mRNA expression of PDE11A. Significantly, the protein expression of PDE11A was consistent with the transcript results, where increased protein levels were found in the knockdown cell line but were reduced to the expression levels found in the wild-type cells with the reexpression of p54^nrb/NONO (Fig. 7B).

FIG 7.

Expression of p54^nrb/NONO restores PDE11A expression. (A) p54^nrb/NONO^KD cells were transfected with pCR3.1-HA-p54^nrb or the pCR3.1 empty vector. At 48 h after transfection, cells were treated with 0.4 mM Bt₂cAMP for 16 h, RNA was isolated, and the mRNA expression of PDE11A was compared to that in the wild-type H295R cells. PDE11A expression is normalized to the mRNA expression of ACTB, and data are graphed as mean ± SEM from two independent experiments, each done in triplicate. (B) Wild-type or p54^nrb/NONO^KD cells that were transfected with an expression plasmid for HA-tagged p54^nrb/NONO were treated with Bt₂cAMP and whole-cell lysates isolated for SDS-PAGE and Western blotting. Blots were incubated with anti-PDE11A1, anti-HA, or GAPDH. Shown are representative blots from one independent experiment, where three separate experiments were carried out, each done in duplicate.

Silencing of p54^nrb/NONO increases mature PDE transcripts in cellular fractions.

We performed oligo(dT) affinity chromatography to selectively enrich poly(A)⁺ mRNA and reexamined the changes in PDE variants by real-time RT-PCR analysis. Consistent with the results shown in Fig. 4, the expression levels of PDE2A, PDE3A, PDE3B, PDE4A, PDE4D, and PDE11A poly(A)⁺ mRNAs were all significantly induced by silencing p54^nrb/NONO (Fig. 8A). In addition to analyzing the changes in PDE transcripts, we also assessed the expression of the low-density lipoprotein (LDL) receptor, which imports circulating cholesterol as LDL particles for steroidogenesis (33), the LIPE gene, coding for the hormone-sensitive lipase (HSL), which catalyzes the hydrolysis of fatty acid esters (34), and the NR5A2 gene, coding for the liver receptor homolog-1 (LRH1), which also plays a critical role in steroidogenesis (35). However, we did not observe altered expression of LDLR, LIPE, or NR5A2 (see Fig. S2 in the supplemental material), suggesting that the effect of silencing p54^nrb/NONO does not globally impact steroidogenesis-related genes.

FIG 8.

Analyses of cytoplasmic and nuclear poly(A)⁺ mRNAs by silencing of p54^nrb/NONO. (A) poly(A)⁺ mRNAs of the indicated PDEs were isolated and quantified by qRT-PCR. (B) poly(A)⁺ mRNA levels of the indicated PDE transcripts were analyzed in both cytoplasmic and nuclear fractions. (C) The nuclear/cytoplasmic mRNA ratio of indicated PDE transcripts was calculated from panel B. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

p54^nrb/NONO and PSF have been recently shown to promote the export of spliceosomal U small nuclear RNA (snRNA) (36). Thus, we next examined the subcellular distribution of PDE transcripts and found that while silencing p54^nrb/NONO caused the accumulation of the poly(A)⁺ forms of several PDE isoforms in both cytoplasmic and nuclear compartments (Fig. 8B), when comparing the ratio of poly(A)⁺ mRNA expression in the nucleus and cytoplasm (nuc/cyto ratio) in WT and p54^nrb/NONO^KD cells, we observed a decreased nuc/cyto ratio in the PDE3A transcript. Notably, we found an increase in the nuc/cyto ratios of PDE2A, PDE3B, PDE4A, PDE4D, and PDE11A in the knockdown cell line (Fig. 8C), suggesting selective accumulation of PDE isoforms in the nucleus.

PDE transcripts interact with p54^nrb/NONO and XRN2.

To further probe the mechanism by which p54^nrb/NONO regulates PDEs, we examined the interaction of endogenous p54^nrb/NONO with PDE mRNAs and found significant enrichments of multiple PDE mRNAs but no enrichment of PDE4B mRNA (Fig. 9A). This RNA immunoprecipitation analysis also revealed an interaction between p54^nrb/NONO and CYP17A1 mRNA. Given our previous findings demonstrating that p54^nrb/NONO regulates the transcription of CYP17A1 by binding to SF1 at the promoter of this gene (22), our current findings may indicate that p54^nrb/NONO may serve to bridge transcription and splicing of CYP17A1.

FIG 9.

p54^nrb/NONO and XRN2 bind to select PDE isoforms. (A and B) Fold enrichment of the indicated transcripts after radioimmunoprecipitation with an anti-p54^nrb/NONO antibody of H295R lysates (A) or an anti-XRN2 antibody relative to control IgG in H295R WT and p54^nrb/NONO^KD cell lysates (B). (C) H295R cell lysates were subjected to immunoprecipitation using an anti-p54^nrb/NONO antibody and protein A/G-agarose. Immobilized proteins were washed, separated by SDS-PAGE, and analyzed by Western blotting. Output blots were incubated with an anti-XRN2 antibody. Ten percent inputs were subjected to SDS-PAGE and Western blotting using an antibody against p54^nrb/NONO. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

The increase in nuclear PDE RNA (Fig. 8) prompted us to determine if p54^nrb/NONO regulated PDE expression by modulating RNA decay. p54^nrb/NONO can recruit 5′-3′ exoribonuclease 2 (XRN2), the main exoribonuclease that is responsible for RNA degradation in the nucleus, to facilitate RNA processing (37, 38). Therefore, we determined if XRN2 interacted with PDE transcripts and if this interaction was altered in the p54^nrb/NONO^KD cell line. Although we did not detect the enrichment of CYP17A1 mRNA (not shown), we detected significant enrichments of PDE2A, PDE4A, and PDE11A mRNAs bound to XRN2 in WT cells but not in p54^nrb/NONO^KD cells (Fig. 9B), suggesting that the interaction between XRN2 and distinct PDE mRNA is dependent on the presence of p54^nrb/NONO. Coimmunoprecipitation assays confirmed that XRN2 copurifies with p54^nrb/NONO, supporting a role for p54^nrb/NONO in RNA metabolism through the modulation of exoribonuclease XRN2-mediated PDE mRNA decay.

Suppressing p54^nrb/NONO reduces RNA degradation in PDEs.

To further explore the mechanism by which p54^nrb/NONO regulated PDE mRNA degradation, actinomycin D was used to inhibit transcription in WT p54^nrb/NONO^KD cells, and RNA decay was followed over time. Silencing of p54^nrb/NONO stabilized spliced mRNAs of PDE2A, PDE3A, PDE3B, PDE4A, PDE4D, and PDE11A (Fig. 10), consistent with the upregulation of these transcripts in p54^nrb/NONO^KD cells (Fig. 4). In contrast, the stability of PDE4B mRNA was not affected by p54^nrb/NONO repression, whereas silencing p54^nrb/NONO decreased the stability of CYP17A1 mRNA (Fig. 10), consistent with the downregulation in expression of this steroidogenic enzyme (Fig. 2B).

FIG 10.

Silencing of p54^nrb/NONO alters RNA decay. Wild-type and p54^nrb/NONO^KD H295R cells were treated with actinomycin D, and total RNAs were harvested at the indicated times. The mRNA decay of CYP17A1, PDE2A, PDE3A, PDE3B, PDE4A, PDE4B, PDE4D, and PDE11A was determined by real-time RT-PCR and normalized to ACTB mRNA. The averages from three independent experiments are plotted, with standard deviations denoted by the error bars. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

DISCUSSION

Previously we identified p54^nrb/NONO as a component in a complex containing SF1, PSF, and mSin3A that regulated CYP17A1 gene transcription (22). Consistent with a role in cAMP signaling, p54^nrb/NONO has been shown to link cAMP response element binding protein (CREB)/transducer of regulated cAMP response element binding protein 2 (TORC2) and RNA polymerase II (Pol II) (16). p54^nrb/NONO was also shown to be involved in a multiprotein complex with SF1, PSF, LRH1, and high-mobility group proteins that regulate cAMP-stimulated retinol-binding protein transcription (39). In addition, p54^nrb/NONO can bind to Rasd1 to regulate the cAMP signaling cascade (40). In this study, we identified p54^nrb/NONO as a novel regulator of PDE expression by controlling the degradation of select PDE isoforms. Since PDEs play a pivotal role in controlling the intracellular levels of cAMP, and thus cAMP-dependent signaling, p54^nrb/NONO acts to regulate the ability of ACTH to activate cortisol biosynthesis.

As discussed above, p54^nrb/NONO is a member of the DBHS family of proteins. Members of this family of RNA-binding proteins have been implicated in both transcriptional and posttranscriptional gene regulation, from transcription activation/repression to RNA splicing to RNA retention (3–6). Previous studies have shown that p54^nrb/NONO can bind to β-globin pre-mRNA (41) and the intronic pyrimidine-rich sequence in β-tropomysin pre-mRNA (18). p54^nrb/NONO also associates with the activated carboxy-terminal domain (CTD) of Pol II to facilitate pre-mRNA splicing (42). Other studies have revealed that p54^nrb/NONO can be recruited to both 5′ and 3′ splice sites and promote intron excision during pre-mRNA splicing (43). Additionally, p54^nrb/NONO can interact with splicing silencer sequences to regulate alternative RNA splicing (4, 44, 45), thus controlling isoform and splice variant expression. Our study provides further support for the role of p54^nrb/NONO in regulating RNA processing. Whole-genome exon array analysis identified alternative splicing of 594 coding genes when the expression of p54^nrb/NONO was silenced (Fig. 5). To our knowledge, this is the first exon-level analysis demonstrating the fundamental role for p54^nrb/NONO in alternative splicing. Among these 594 coding genes, we found distinct PDE isoforms, the products of which hydrolyze cAMP or cGMP and maintain intracellular cyclic nucleotide levels. As mentioned above, PDEs are encoded by at least 21 genes that are categorized, based on sequence similarity, mode of regulation, and preference for cAMP or cGMP as the substrate, into 11 gene families. Over 90 splice variants are generated from different transcription initiation sites and differential splicing of their mRNAs (23, 24). The observation that p54^nrb/NONO repression caused the upregulation of spliced PDE RNAs in select isoforms and splice variants suggested that p54^nrb/NONO is not uniformly required for the pre-mRNA splicing of all PDEs (Fig. 4).

Bioinformatic analysis of the whole-genome exon array data identified alternative splicing existing in PDE2A and PDE11A in p54^nrb/NONO knockdown cells (Fig. 5). By analyzing results obtained from real-time RT-PCR together with exon-level expression analysis, we not only found the upregulated expression of six isoforms of PDEs, PDE2A, PDE3A, PDE3B, PDE4A, PDE4D, and PDE11A but also provide evidence that the splice variants PDE2A2 and PDE11A4 might be preferably spliced when p54^nrb/NONO is silenced. Support for the interaction of p54^nrb/NONO with select PDE transcripts, including PDE2A, PDE3A, PDE3B, PDE4A, PDE4D, and PDE11A, was further confirmed by the detection of a protein-RNA complex after immunoprecipitation of p54^nrb/NONO (Fig. 9).

We observed an increase in the ratio of nuclear to cytoplasmic transcript levels for PDE2A, PDE4D, and PDE11A (Fig. 8), suggesting that p54^nrb/NONO may act to regulate the export of transcripts from the nucleus. However, our finding that these isoforms interact with XRN2 in a p54^nrb/NONO-dependent manner (Fig. 9B) points to a role for p54^nrb/NONO in controlling the degradation of specific PDE transcripts. p54^nrb/NONO has recently been shown to bind to A-kinase anchoring protein 3 (AKAP3) and other sperm-specific mRNAs during the elongation stage of spermiogenesis, where cAMP signaling increases AKAP3 translation without affecting transcription of the gene (46). It was previously shown that p54^nrb/NONO can recruit XRN2 to facilitate transcription termination and pre-mRNA 3′ processing, which links RNA degradation to RNA export (37). We show here that a protein-mRNA complex exists between XRN2 and PDE transcripts, including PDE2A, PDE4A, and PDE11A, which can be disrupted by suppressing p54^nrb/NONO. Interaction between XRN2 and p54^nrb/NONO (Fig. 9C) lends support to the notion that p54^nrb/NONO acts as a bridge in facilitating XRN2-catalyzed mRNA turnover. Consistent with this role for p54^nrb/NONO in RNA metabolism, we show that silencing of p54^nrb/NONO increases the half-lives of several PDE transcripts (Fig. 10).

Collectively, our findings identified a key role for p54^nrb/NONO in regulating cAMP/cGMP modulator PDEs by altering pre-mRNA splicing, alternative splicing, and mRNA decay through XRN2. Previous studies have suggested that PDE2A, PDE8B, and PDE11A are associated with endocrine diseases in pituitary, thyroid, and adrenal glands and testis (47–51). PDE11A was the first PDE to be associated with adrenocortical tumorigenesis, and the inactivating mutations of PDE11A4 have been identified in a subgroup of patients with Cushing's syndrome due to micronodular adrenocortical hyperplasia (25, 30, 52, 53). Consistent with a role for aberrant function of this variant in adrenal cells, PDE11A4 is one of the variants that is alternatively spliced when we silence p54^nrb/NONO in the H295R adrenal carcinoma cell line. The splicing and degradation of PDE2A were also found to be tightly controlled by p54^nrb/NONO. Therefore, our studies provide novel insights into the pivotal role for p54^nrb/NONO in controlling the splicing and degradation of PDEs. The ability of p54^nrb/NONO to interact with PDEs suggests that in addition to glucocorticoid biosynthesis, the nuclear protein may exert regulatory effects on varied cyclic nucleotide-dependent physiological processes in other cell types.