The Polypyrimidine Tract-binding Protein (PTB) Is Involved in the Post-transcriptional Regulation of Human Inducible Nitric Oxide Synthase Expression*

Human inducible nitric oxide synthase (iNOS) expression is regulated by transcriptional and post-transcriptional mechanisms. We have recently shown that the multifunctional RNA-binding proteins KH-type splicing regulatory protein and tristetraprolin are critically involved in the post-transcriptional regulation of human iNOS expression. Several reports have shown that KH-type splicing regulatory protein colocalizes with the polypyrimidine tract-binding protein (PTB), and both RNA-binding proteins seem to interact with the same mRNAs. Therefore we analyzed the involvement of PTB in human iNOS expression. In human DLD-1 cells, cytokine incubation necessary to induce iNOS expression did not change PTB localization or expression. However, intracellular binding of PTB to the human iNOS mRNA increased after cytokine stimulation. Overexpression of PTB resulted in enhanced cytokine-induced iNOS expression. Accordingly, small interfering RNA-mediated knock down of PTB reduced cytokine-dependent iNOS expression. Recombinant PTB displayed binding to an UC-rich sequence in the 3′-untranslated region of the human iNOS mRNA. Transfection experiments showed that PTB mediates its effect on iNOS expression via binding to this region. The underlying mechanism is based on a modulation of iNOS mRNA stability. In summary, human iNOS is the first example of a human pro-inflammatory gene regulated by PTB on the level of mRNA stability.

Post-transcriptional mechanisms represent an important part of the regulation of gene expression. In particular, genes whose expression has to be controlled precisely (e.g. oncogenes, pro-inflammatory genes) are regulated at the post-transcriptional level, mostly by modulation of mRNA stability (1,2). Inherently unstable mRNAs that code for cytokines, transcription factors, proto-oncogenes, and pro-inflammatory mediators often contain AU-rich elements (AREs) 3 in their 3Ј-un-translated regions (3Ј-UTRs). These AREs are targets for transacting proteins regulating mRNA stability and translation (1,3). The family of ARE-binding proteins includes the embryonic lethal abnormal vision protein family members (most importantly HuR), the ARE/poly(U)-binding/degradation factor 1 (AUF-1, also named hnRNP D), the KH-type splicing regulatory protein (KSRP), tristetraprolin, the T cell-restricted intracellular antigen (TIA-1), and the T cell-restricted intracellular antigen-related protein (1,3). In mammalian cells, the AREsequences mediate mRNA decay mainly by recruitment of the exosome (a multisubunit particle with 3Ј to 5Ј nuclease activity) to the mRNAs, thereby promoting their rapid degradation. However, the mammalian exosome does not seem to recognize the ARE-containing RNAs on its own but requires certain AREbinding proteins (like KSRP or tristetraprolin) for this interaction (4,5).
The polypyrimidine tract-binding protein (PTB), also known as hnRNP I, is a major hnRNP protein with multiple roles in mRNA metabolism, including regulation of alternative splicing (6), internal ribosome entry site-driven translation (7), hepatitis C virus replication (8), mRNA localization (9), and polyadenylation (10). Recent reports also showed stabilization of the CD154, insulin, and vascular endothelial growth factor mRNA by PTB (11)(12)(13)(14)(15). PTB has strong RNA binding activity since it possesses four tandem RNA recognition motif domains (16). A preferred RNA-binding site of PTB is a UCUU sequence flanked by pyrimidines (17). PTB has been shown to colocalize with KSRP in murine and human cells (18,19). In the case of alternative splicing of the neuronal-specific N1 exon of c-Src opposite, roles for PTB (exclusion of the N1 exon) and KSRP (inclusion of the N1 exon) have been described (18).
The inducible isoform of nitric oxide synthase (iNOS) is an important enzyme involved in multiple physiologic and pathophysiologic pathways (20,21). Its expression is controlled in large part by post-transcriptional mechanisms. The 3Ј-UTR of the human iNOS mRNA contains five ARE sequences and destabilizes the mRNA of a heterologous reporter gene in human A549 or DLD-1 cells (22). In recent studies, we showed that KSRP, HuR, and tristetraprolin are essentially involved in the post-transcriptional regulation of human iNOS expression in a complex manner (22)(23)(24)(25).
We were interested in analyzing the involvement of PTB in the regulation of human iNOS expression because KSRP seems to colocalize with PTB and binding of both proteins to the same mRNA has been described (18,19,26). Moreover, in the murine system, an interaction of PTB with the iNOS mRNA has been reported (27).
In DLD-1 cells, cytokine incubation did not change PTB expression or localization. However, intracellular binding of PTB to the human iNOS mRNA was markedly increased after cytokine stimulation. Overexpression or siRNA-mediated down-regulation of PTB demonstrated that PTB enhances cytokine-dependent iNOS induction. In vitro binding studies revealed an interaction of PTB with an UC-rich sequence in the 3Ј-UTR of the human iNOS mRNA. We could show that PTB exerts its effect on iNOS expression via this binding by enhancing iNOS mRNA stability.

EXPERIMENTAL PROCEDURES
Cell Culture, Cytokine Treatment, and RNA Isolation-Human alveolar epithelial A549/8 cells and human colon carcinoma DLD-1 cells were grown in Dulbecco's modified Eagle's medium with 2 mM L-glutamine, penicillin, and streptomycin (all from Sigma, Deisenhofen, Germany) and 5 or 10% heatinactivated fetal bovine serum (from PAN-Systems, Nürnberg, Germany), respectively. Eighteen hours before cytokine induction, the cells were washed with phosphate-buffered saline and incubated with Dulbecco's modified Eagle's medium containing 2 mM L-glutamine in the absence of serum and phenol red. iNOS expression in cells was induced with a cytokine mixture (CM) containing human interferon-␥ (100 units/ml), interleukin-1␤ (50 units/ml), and tumor necrosis factor-␣ (10 ng/ml, all cytokines from Strathmann, Hannover, Germany) for the corresponding time periods depending on the experiment. Afterward, supernatant of the cells (50 l) was used to measure NO 2 Ϫ by the Sievers NOA 280 nitric oxide analyzer (ADInstruments, Spechbach, Germany), and cells were processed for RNA isolation by guanidinium thiocyanate/phenol/chloroform extraction as described (22) or for protein extraction as described below.
Oligonucleotide 3 Each experimental reaction was performed in triplicate. All primer/probe sets had efficiencies of 100% (Ϯ10%).
To calculate the relative expression of iNOS or PTB mRNA, the 2 (Ϫ⌬⌬C(T )) method (29) was used. The values of untreated cell samples were set at 100%, and the percentage of iNOS or PTB mRNA expression was calculated.
DRB Experiments-To analyze the effect of experimental interventions on iNOS mRNA stability, cells were incubated as indicated, and iNOS expression was induced by cytokines for 4 h. Then 25 g/ml 6-dichloro-1-ribofuranosylbenzimidazole (DRB, Sigma) was added, and RNAs were prepared 0 -4 h thereafter. Relative iNOS and GAPDH mRNA amounts were determined by qRT-PCR, and iNOS mRNA was normalized to GAPDH mRNA. The relative amount of iNOS mRNA at 0 h DRB was set at 100%. Curve fittings of the resulting DRB time curves were performed by non-linear regression using Graph-Pad Prism 3.0 (GraphPad Software, San Diego, CA).
Analysis of Human iNOS Promoter Activity in Stably Transfected Cells-To investigate the effect of PTB overexpression on cytokine-induced iNOS promoter activity, DLD-1-EGFP⅐ PTB or DLD-1-EGFP-C1 cells were transiently transfected by lipofection with FuGENE according to the manufacturer's recommendations. After overnight incubation, cells were incubated with or without CM. Then the cells were lysed in 1ϫ passive lysis buffer (Promega, Heidelberg, Germany), and firefly and Renilla luciferase activities were determined. The light units of the firefly luciferase were normalized by those of Renilla luciferase after subtraction of extract background.
Western Blot Experiments-To study protein expression in DLD-1 or A549/8 cells, total cell proteins or cytoplasmic and nuclear proteins were separated on SDS-polyacrylamide gels and transferred to nitrocellulose membranes by semidry electroblotting. All further steps were performed as described (22). For the detection of KSRP, iNOS, PTB, TATA box-binding factor, or ␤-tubulin monoclonal antibodies (anti-iNOS, R&D Systems, Wiesbaden, Germany; anti-PTB, Zymed Laboratories Inc., San Francisco, CA; anti-TATA box-binding factor, BD Transduction Laboratories, Heidelberg, Germany; anti-␤-tubulin, Sigma) were used. The immunoreactive proteins on the blots were visualized by the enhanced chemiluminescence detection system (ECL; Amersham Biosciences, Freiburg, Germany).
Analysis of mRNA Stability Using a Tetracycline-inducible Expression Vector-DLD-1-TR7 cells constitutively expressing a tetracycline repressor were a kind gift of Dr. M. Weitering, Utrecht, The Netherlands. Two different pcDNA4/TO-based luciferase expression vectors with or without the human iNOS 3Ј-UTR sequence were generated. DNA sequences of the clones were determined using the dideoxy chain termination method with a sequencing kit from Amersham Biosciences.
To analyze the effect of PTB on the expression of these luciferase reporter mRNAs, DLD-1-TR7 cells were transiently transfected by lipofection with FuGENE according to the manufacturer's recommendations. To induce luciferase expression, cells were incubated with 10 ng/ml doxycycline for 24 h. Then cells were lysed in 1ϫ passive lysis buffer, and firefly and Renilla luciferase activities were determined as described above.
Purification of GST⅐PTB Proteins-A procaryotic expression vector (pGEX2T-PTB) coding for a GST⅐PTB fusion protein was generated. Purified GST or GST⅐PTB fusion proteins were prepared using the plasmids pGEX2T (Amersham Biosciences) and pGEX2T-PTB as described (22). The yield of the purification procedure was determined by comparison with a BSA standard on Coomassie Blue-stained SDS-PAGE.
Down-regulation of PTB Expression by RNA Interference-To generate an expression vector enabling intracellular short hairpin RNA (shRNA) synthesis, a double-stranded oligonucleotide (5Ј-AAGGAACTTCCATCATTCCAGAGAACTTGC-TTCTTCTCTGGAATGATGGAAGTTCCTATAGTGA-3Ј; sequence of the siRNA repeats directed against the human PTB mRNA underlined) was cloned into the BbsI sites of psiRNA-hH1-GFPzeo (InvivoGen, San Diego, CA) to generate psiRNA-hH1-GFPzeo-PTB. The DNA sequence of the construct was determined using the dideoxy chain termination method with a sequencing kit from Amersham Biosciences. To generate A549/8 cells stably expressing shRNAs directed against the human PTB mRNA, cells were transfected with psiRNA-hH1-GFPzeo-PTB by lipofection with FuGENE according to the manufacturer's recommendations. Stable transfectants were selected with zeocin (200 g/ml, InvivoGen) and for GFP expression by fluorescence-activated cell sorting. As a control, A549/8 cells stably transfected with the psiRNA-hH1-GFPzeo vector were generated as well.
Immunoprecipitation-qRT-PCR Assay-For determination of intracellular protein-RNA interactions, DLD-1 cells were incubated for 4 h with or without the cytokine mixture. All further steps were performed as described before (24).
Statistics-Data represent means Ϯ S.E. Statistical differences were determined by factorial analysis of variance followed by Fisher's protected least-significant-difference test for comparison of multiple means.

Cytokine Incubation Does Not Change PTB Expression or
Localization in DLD-1 Cells-In previous studies, we demonstrated that KSRP is critically involved in the post-transcriptional regulation of human iNOS expression in interplay with other RNA-binding proteins (24,25). In murine and human cells, KSRP colocalizes with PTB (18,19), and opposite functions of both proteins in hnRNA splicing have been described (18). Therefore we were interested in whether PTB as well as KSRP fulfills a role in the regulation of iNOS expression.
Induction of iNOS in human cells requires stimulation with a complex CM containing interleukin-1␤, interferon-␥, and tumor necrosis factor-␣ (23). First we analyzed the influence of this cytokine mixture on the expression of PTB. Therefore DLD-1 cells were incubated with or without CM for 2-24 h, and RNA and total cell proteins were isolated. As shown before for KSRP (24), cytokine incubation did not affect PTB protein (Fig.  1A) or mRNA expression (Supplemental Fig. 1).
Then the cellular localization of PTB and KSRP was determined by Western blot experiments using nuclear and cyto- plasmic extracts from DLD-1 cells. These analyses showed that treatment with cytokines for 4 h did not change the cellular distribution of PTB or KSRP (Fig. 1B). Also, shorter time periods of CM incubation had no effect on the localization of both proteins (data not shown).
Cytokine Incubation Enhances Intracellular Binding of PTB to the Human iNOS mRNA-To analyze whether PTB binds to the human iNOS mRNA in intact cells and whether CM incubation modifies this interaction, we performed immunoprecipitation-qRT-PCR assays using specific PTB (and as control) KSRP antibodies. Interestingly, the intracellular binding of PTB to the iNOS mRNA was enhanced up to 5-fold after cytokine induction (Fig. 2). As published before (24), cytokine incubation reduced the intracellular binding of KSRP to the human iNOS mRNA. Thus both proteins showed opposite binding properties. Altogether, these results imply an important role for PTB in the regulation of human iNOS expression.
Overexpression of PTB Enhances Cytokine-induced iNOS Expression-To determine the effect of PTB on human iNOS expression, we generated pools of stably transfected DLD-1 cells, which constitutively express an EGFP⅐PTB fusion protein, or as control, the EGFP protein (Fig. 3A). These cells were incubated in the presence or absence of CM for different time periods, and total cell proteins and RNA were isolated. Also, cytokine-induced NO production was analyzed by measuring the nitrite content in the supernatants of the cells.
These analyses showed a PTB-dependent enhancement of human iNOS expression. In DLD-1-EGFP⅐PTB cells, we detected a similar increase of cytokine-induced iNOS mRNA (Fig. 3B) and protein expression (Fig. 3C) as well as iNOS-mediated NO production (Fig. 3D) when compared with control cells (DLD-1-EGFP).
Down-regulation of PTB Reduces Cytokine-induced iNOS Expression in A549/8 and DLD-1 Cells-To confirm these results, we tested whether down-regulation of endogenous PTB expression using the RNA interference technique resulted in reduced iNOS expression. Therefore A549/8 cells (known to be able to express iNOS after cytokine incubation (30)) were stably transfected with an expression vector (psiRNA-hH1-GFPzeo-PTB) leading to constitutive intracellular expression of shRNAs directed against the PTB coding region (indicated on figures as siPTB). As a control, cells stably transfected with the expression vector (indicated on figures as siGFP) were generated as well. These cell pools were incubated in the presence or absence of the cytokine mixture, and total protein and RNA were isolated. NO production was determined as described above.
In A549/8 cells stably transfected with psiRNA-hH1-GFPzeo-PTB (siPTB) PTB protein (Fig. 4A) and mRNA (Supplemental Fig. 2), expression was significantly reduced when compared with control cells (siGFP). Analysis of cytokine-induced iNOS mRNA in these cells showed that down-regulation of PTB by siPTB clearly reduced iNOS mRNA expression (Fig.  4B). Also, cytokine-induced iNOS protein expression (Fig. 4C) and iNOS-dependent NO production (Fig. 4D) were reduced to a similar degree in these cells. Transient transfection of anti-PTB shRNAs into DLD-1 cells resulted in an analogous effect (see Supplemental Fig. 3). In summary, the data indicate that PTB up-regulates cytokine-induced human iNOS expression.
PTB Binds to the 3Ј-UTR of the Human iNOS mRNA-The human iNOS 3Ј-UTR sequence contains several putative PTBbinding sites (Fig. 5A, UCUU_1 to UCUU_3). To investigate the definite binding site of PTB in the human iNOS 3Ј-UTR, recombinant GST⅐PTB fusion protein was incubated with 32 Plabeled transcripts comprising different nucleotide regions (Fig. 5A). Then PTB-RNA interaction was assayed by UV crosslinking experiments. We detected complex formation between recombinant GST⅐PTB protein and the whole 3Ј-UTR transcript. This binding activity was not observed with the GST protein (Fig. 5B, panel I, 3Ј-UTR). To localize the PTB-binding site within the iNOS 3Ј-UTR, the region was first dissected into two subfragments. One was the non-AU fragment without AU repeats containing one UCUU sequence (UCUU_1). The other was the AU fragment containing the AU repeats and two UCUU sequences (UCUU_2 and UCUU_3). Only the AU subfragment interacted with the GST⅐PTB protein (Fig. 5B, panel  I). Subsequently, the AU fragment was dissected into three subfragments: subfragment A (232-329, with the UCUU_2 sequence), subfragment B (327-428, with the UCUU_3 sequence), and subfragment C (387-477, with the UCUU_3 sequence). As shown in Fig. 5B (panel II), GST⅐PTB displayed a marked binding activity to both fragment B and fragment C. These results suggested that the PTB-binding site in the human iNOS mRNA 3Ј-UTR is located in the overlapping sequence of fragments B and C (Fig. 5A, B_C_overlap). Also, these data imply that the UCUU_2 sequence is not important for the binding of PTB to the iNOS 3Ј-UTR. In accordance, direct mutation of the UCUU_2 sequence (UCUU 3 AGAA) did not result in reduced binding to the human iNOS 3Ј-UTR (data not shown). As shown in Fig. 5B (panel III), deletion of the UCUU_3 sequence (indicated as AU-⌬Nco) in the context of the AU fragment or direct mutation of this sequence (UCUU 3 AGAA; data not shown) did not modify the binding of GST⅐PTB. However, deletion of the B_C-overlap sequence in the context of the AU fragment (Fig. 5, panel III, AU-⌬BsrGI) resulted in a complete loss of the binding of GST⅐PTB.
All the data described above showed that none of the UCUU sequences are important for the binding of PTB to the human iNOS 3Ј-UTR. However, the UC-rich sequence 5Ј-UAACACCC-AGUCUGUUCCCCAUGG-3Ј (positions 389 -412 of the 3Ј-UTR; Fig.  5A, F_PTB_s) seems to be essential for this PTB-RNA interaction.
PTB Enhances the Expression of a Luciferase mRNA Containing the Human iNOS mRNA 3Ј-UTR-To analyze whether the enhancing effect of PTB on human iNOS mRNA expression in intact cells results from its binding to the 3Ј-UTR, we performed Tet-On analyses. We generated expression vectors with (pcDNA4/TO-Luc-3Ј-UTR; indicated on figures as Luc-UTR) or without (pcDNA4/TO-luc; indicated on figures as Luc) the iNOS 3Ј-UTR cloned behind the firefly luciferase reporter gene under the control of a tetracycline responsible promoter. These vectors together with pEGFP-C1 or pEGFP-PTB and pRL-SV40 (Renilla luciferase expression vector) were transiently transfected into DLD-1-TR7 cells constitutively expressing a tetracycline repressor. To induce luciferase expression, cells were incubated for 24 h with doxycycline (Dox). Then cells were lysed, and relative luciferase activities were determined. As shown in Fig. 6, Dox incubation resulted in a marked enhancement of firefly luciferase activity (compare columns 1 or 2 with columns 5 and 6, respectively). The presence of the 3Ј-UTR of the human iNOS mRNA 3Ј to the firefly luciferase stop codon markedly reduced luciferase expression (comparison of Luc with Luc-UTR in pEGFP-C1-transfected cells; column 5 versus column 6). Overexpression of PTB in this system resulted in clearly enhanced firefly luciferase activity in cells transfected with pcDNA4/TO-Luc-UTR (comparison of Luc-UTR in pEGFP-C1 and pEGFP-PTB-transfected cells; column 6 versus column 8). Therefore the effect of PTB on the human iNOS mRNA is mediated by its binding to the human iNOS 3Ј-UTR sequence.
Overexpression of PTB Modulates Human iNOS mRNA Stability-To test whether the enhancing effect of PTB on iNOS expression resulted from PTB-mediated changes in the stability of the human iNOS mRNA, we performed experiments using DRB to block the RNA polymerase II-dependent transcription. DLD-1-EGFP-C1 (indicated on figures as EGFP) or DLD-1-EGFP⅐PTB (indicated on figures legends as EGFP⅐PTB) cells were incubated with CM for 4 h. Then DRB (25 g/ml) was added to stop transcription, and RNA was iso- lated after 0, 2, and 4 h. Expression of iNOS mRNA in comparison with GAPDH was determined by qRT-PCR. As shown in Fig. 7A, overexpression of PTB resulted in a nearly 2-fold enhancement of human iNOS mRNA stability (t1 ⁄ 2 ϭ 5.9 Ϯ 0.9 h) when compared with pEGFP-C1 cells (t1 ⁄ 2 ϭ 3.3 Ϯ 0.5 h).
As several RNA-binding proteins have been described also to bind to DNA and regulate promoter activity (31-34), we analyzed whether increase of PTB expression changes human iNOS promoter activity. We transiently transfected pNOS2(16)Luc (containing a 16-kb fragment of the human iNOS promoter (35)) and pRL-SV40 into DLD-1-EGFP-C1 (Fig. 7, EGFP) and DLD-1-EGFP⅐PTB (Fig. 7, EGFP⅐PTB) cells. After transfection, the cells were incubated for different time periods in the presence or absence of the cytokine mixture, and firefly and Renilla luciferase activity were analyzed in extracts of these cells. As shown in Fig. 7B, the human iNOS promoter displayed a nearly identical behavior in both cell types with a maximal 8-fold increase of promoter activity after a 6-h incu-bation with CM. Therefore the effect of PTB on iNOS expression does not involve regulation of human iNOS promoter activity. In summary, the enhancing effect of PTB on human iNOS mRNA expression results from increased mRNA stability.

DISCUSSION
The individual rate of mRNA decay is determined by cis-acting sequences within the mRNA, which are recognized by trans-acting factors (1,2). Dysregulation of these mRNA protein interactions and thereby dysregulation of mRNA stability has been associated with different human diseases like chronic inflammatory diseases, ␣-thalassemia, cancer, cardiovascular diseases, and Alzheimer disease (1,36).
We have shown that the human iNOS expression is critically regulated on the level of mRNA stability (22,24,25). The 3Ј-UTR of the human iNOS mRNA contains five AREs (Fig. 5A) and has been shown to destabilize reporter mRNAs (22). Of several proteins known to bind to such AREs, the embryonic lethal abnormal vision protein HuR and KSRP bind to this 3Ј-UTR sequence and regulate the stability of the human iNOS mRNA (22,24).
Beyond this, several other RNAbinding proteins have been characterized that are involved in the regulation of RNA metabolism. One of them is the PTB also known as hnRNP I. PTB has initially been described as a splicing repressor that affects splicing of many alternative exons (6). However, recent evidence suggests that PTB is a major hnRNP protein with multiple roles in mRNA metabolism, including the stabilization of mRNAs like the CD154, insulin, and vascular endothelial growth factor mRNA (11)(12)(13)(14)(15). Moreover, PTB has been implicated in the regulation of the murine iNOS expression (27,37).
Previous reports have shown that PTB interacts and colocalizes with KSRP in murine and human cells (18,19). As we have demonstrated that KSRP is critically involved in regulation of human iNOS expression (24,25), we analyzed the expression and cellular localization of PTB and KSRP in DLD-1 cells. As shown by Hall et al. (19) for HeLa cells, we could detect a very similar distribution of PTB and KSRP in the nucleus and the cytoplasm of DLD-1 cells by Western blot (Fig. 1B) and immunofluorescence analyses (data not shown). The localization and expression of both proteins did not change by cytokine treatment of the cells (Fig. 1A) (24).

. Down-regulation of PTB expression by siRNAs reduces cytokine induced iNOS expression.
A549/8 cells were stably transfected with psiRNA-hH1-GFPzeo-PTB (siPTB). In these cells, shRNAs were generated that were directed against the human PTB mRNA. Stable transfectants were selected for zeocin resistance and GFP expression. As a control, A549/8 cells stably transfected with the psiRNA-hH1-GFPzeo vector were generated as well (siGFP). The cells were preincubated for 18 h in medium without fetal calf serum and phenol red before being exposed to cytokines (CM) for 6 or 8 h. Then total RNA or protein extracts were prepared. To determine iNOS-mediated NO production, cells were incubated for 24 h with or without CM, and the supernatant of the cells was analyzed for nitrite content. A, analysis of PTB and ␤-tubulin (␤-Tub) protein expression in extracts from A549/8-psiRNA-hH1-GFPzeo-PTB (siPTB) or -psiRNA-hH1-GFPzeo In contrast to the unchanged localization and expression of PTB and KSRP, cytokine incubation of DLD-1 cells resulted in a markedly enhanced intracellular binding of PTB to the human iNOS mRNA (Fig. 2). As published before (24), the KSRP binding was reduced in the same cells. As both proteins are described to have contrary roles in the splicing of the c-Src mRNA (18) and PTB has been shown to enhance the stability of different mRNAs (12,14,15,26,38), we hypothesized that PTB presumably regulates human iNOS expression in an opposite manner than KSRP.
Indeed, our data confirmed this hypothesis. In contrast to KSRP, which destabilizes iNOS mRNA (24), PTB up-regulates iNOS expression as shown by overexpression (Fig. 3) or siRNA-mediated down-regulation of PTB (Fig. 4). As the effect of PTB on iNOS mRNA equals those on iNOS protein and iNOS-dependent NO production, PTB seems to influence mainly iNOS mRNA expression. The effect of PTB on iNOS mRNA translation seems of minor importance but cannot be excluded absolutely by the data presented. Since these effects were observed in DLD-1 and A549/8 cells, PTB-mediated modulation of iNOS expression appears to be a more general feature of human cells.
PTB possesses a strong RNA binding activity, and in vitro studies determined UCUU flanked by pyrimidines as a preferred PTB-binding site (17). Analysis of the human iNOS 3Ј-UTR sequence showed the existence of three of these motifs (Fig. 5A, UCUU_1 to UCUU_3). Analysis of PTB interaction with the human iNOS mRNA 3Ј-UTR surprisingly showed that PTB binding does not seem to depend on one of these UCUU sequences. Our data strongly indicate that PTB binds to the sequence 5Ј-UAACACCCAG-UCUGUUCCCCAUGG-3Ј (positions 389 -412 of the human iNOS 3Ј-UTR, indicated in Fig. 5A as F_PTB_s). This sequence is also relatively UC-rich but does not contain the "classical" UCUU-motif. In summary, besides the ARE sequences already proven to be essential for iNOS mRNA stability (22)(23)(24)(25), these data reveal a second important sequence motif necessary for the post-transcriptional regulation of human iNOS expression. To prove the importance of PTB binding to the iNOS 3Ј-UTR in intact cells, we analyzed the effect of PTB on the expression of a luciferase reporter mRNA with (indicated as Luc-UTR) or without (indicated as Luc) the iNOS 3Ј-UTR under control of a tetracycline-inducible promoter. With this method, we detected the formerly described (22) destabilizing effect of the human iNOS mRNA 3Ј-UTR (Fig. 6, compare column 5 and column 6). More importantly, overexpression of PTB in this system markedly enhanced firefly luciferase activity after Dox treatment in the Luc-UTR-transfected cells (Fig. 6, compare column 6 and column 8). So as shown for the endogenous iNOS expression, PTB also enhances expression of an unrelated reporter mRNA containing the human iNOS 3Ј-UTR. Combined with our immunoprecipitation-qRT-PCR and UV cross-linking experiments, this indicates that the PTB effect on human iNOS expression is mediated by the PTB-binding site (Fig. 5A, F_PTB_s) located in the 3Ј-UTR.
All the data shown above support our hypothesis that PTB posttranscriptionally affects human iNOS expression. Indeed, as shown in Fig. 7, PTB enhanced human iNOS mRNA stability but did not change human iNOS promoter activity at any time point analyzed.
The above data indicate that after cytokine incubation, PTB increases the stability of the human iNOS mRNA by enhanced binding to its 3Ј-UTR. In this way, PTB contrib-utes to the cytokine-dependent induction of human iNOS expression and iNOS-mediated NO production.
In the RNA-binding protein network regulating iNOS mRNA stability, KSRP and PTB display antagonistic effects,  3, 4, 7, and 8). The Renilla luciferase expression plasmid pRL-SV40 was cotransfected for normalization. After transfection, cells were treated with 10 ng/ml Dox (columns 5-8) to induce firefly luciferase expression. After 24 h, cells were lysed, and firefly luciferase and Renilla luciferase activity were determined. Columns (mean Ϯ S.E.) represent the relative doxycycline-induced firefly luciferase activities expressed as a percentage of Dox-induced normalized luciferase activity in cells transfected with pcDNA4/TO-Luc and pEGFP-C1 (***, p Ͻ 0.001; ns ϭ not significant versus Dox-induced pEGFP-C1 and pcDNA4/TO-Luc-transfected cells; ###, p Ͻ 0.001 versus pEGFP-C1 and pcDNA4/TO-Luc-UTR-transfected cells). Then 25 g/ml DRB was added, and RNAs were prepared after 0 -4 h. iNOS mRNA and GAPDH mRNA concentrations were determined by qRT-PCR, and iNOS mRNA was normalized to GAPDH mRNA. A summary of six qRT-PCR analyses is shown. Data (means Ϯ S.E.) represent relative iNOS mRNA levels (*, p Ͻ 0.05; **, p Ͻ 0.01; ns ϭ not significant versus 0 h DRB). Curve fitting was performed using GraphPad Prism for Macintosh. B, DLD-1-EGFP-C1 control cells (EGFP) or DLD-1-EGFP⅐PTB cells (EGFP⅐PTB) were transfected with a luciferase reporter gene under the control of the 16-kb human iNOS promoter. After transfection, cells were incubated with or without a mixture of cytokines (CM) for 2-24 h. Columns (mean Ϯ S.E.) represent the cytokine-induced activation of the iNOS promoter expressed as a percentage of the corresponding basal values determined in the absence of cytokines (***, p Ͻ 0.001; *, p Ͻ 0.05; ns ϭ not significant versus untreated pEGFP-C1 control cells; ns # ϭ not significant versus pEGFP-C1 cells incubated for the same time period with or without CM).
which parallels the described opposite roles of both proteins in c-Src-mRNA splicing (18). As published before, KSRP and HuR also regulate iNOS mRNA expression in an opposite manner by competing for the same binding site (24). In contrast, the binding site of PTB in the iNOS 3Ј-UTR differs from that one of KSRP. Thus the antagonistic effect of both proteins does not depend on direct competition for the same RNA sequence. Finally, PTB plays an essential role in the post-transcriptional regulation of human iNOS expression and therefore could also be important for the regulation of other pro-inflammatory human genes whose expression is controlled by KSRP.