HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 51, 50143-50154, December 20, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/51/50143    most recent M207409200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Muh, S. J. Articles by Carstens, R. P. Search for Related Content PubMed PubMed Citation Articles by Muh, S. J. Articles by Carstens, R. P. Social Bookmarking                      What's this?

A Non-sequence-specific Double-stranded RNA Structural Element Regulates Splicing of Two Mutually Exclusive Exons of Fibroblast Growth Factor Receptor 2 (FGFR2)^*

Stephanie J. Muh, Ruben H. Hovhannisyan, and Russ P. Carstens

From the Renal-Electrolyte and Hypertension Division, Department of Medicine, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania 19104-6144

Received for publication, July 23, 2002, and in revised form, October 8, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Alternative splicing of fibroblast growth factor receptor 2 (FGFR2) mutually exclusive exons IIIb and IIIc represents a tightly regulated and functionally relevant example of post-transcriptional gene regulation. Rat prostate cancer DT3 and AT3 cell lines demonstrate exclusive selection of either exon IIIb or exon IIIc, respectively, and have been used to characterize regulatory FGFR2 RNA cis-elements that are required for splicing regulation. Two sequences termed ISE-2 and ISAR are located in the intron between exons IIIb and IIIc and are required for cell-type specific exon IIIb. Previous studies suggest that the function of these elements involves formation of an RNA stem structure, even though they are separated by more than 700 nucleotides. Using transfected minigenes, we performed a systematic analysis of the sequence and structural components of ISE-2 and ISAR that are required for their ability to regulate FGFR2 splicing. We found that the primary sequence of these elements can be replaced by completely unrelated sequences, provided that they are also predicted to form an RNA stem structure. Thus, a nonsequence-specific double stranded RNA stem constitutes a functional element required for FGFR2 splicing; suggesting that a double-stranded RNA binding protein is a component of the splicing regulatory machinery.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Alternative splicing represents an important mechanism of modulating the expression of gene transcripts (1-5). It is estimated that at least 35% of human genes are alternatively spliced, although this is based on conservative estimates and it may be that alternative splicing is more the rule than the exception in the post-transcriptional processing of pre-mRNA transcripts (6, 7). Studies of the mechanisms employed by mammalian cells to differentially regulate splicing indicate that the selection of splice sites is influenced by proteins that bind to non-splice site RNA cis-elements and recruit or block spliceosome assembly. Such RNA cis-elements include exonic and intronic splicing enhancers (ESEs¹ or ISEs) as well as exonic and intronic splicing silencers (ESSs or ISSs) that enhance or block splicing to neighboring splice sites. A number of ubiquitous RNA-binding proteins have been demonstrated to alter alternative splice site choice. Examples include the SR proteins (8-11), heterogeneous nuclear ribonucleoproteins (hnRNPs) (12-26), KSRP (KH-type splicing regulatory protein) (27), and TIA-1 (28, 29). While most proteins demonstrated to modulate splicing of mammalian transcripts are not differentially expressed, several mammalian tissue-specific RNA-binding proteins that function as splicing regulators have recently emerged (30-33). Alternatively spliced mammalian transcripts generally contain several cis-elements in introns and exons with positive or negative functions (enhancers and silencers) that modulate splicing. Such observations have led to models of combinatorial control, whereby regulation of alternatively spliced mammalian transcripts is achieved through the net influence of several proteins that bind to different cis-elements flanking regulated splice sites (1, 5). While some of the regulatory proteins implicated in splicing regulation are ubiquitously expressed, it appears that cell-type specific regulatory factors may often tip the balance in favor of specific splicing patterns in cells that express them.

A functionally relevant example of alternative splicing involves the mutually exclusive splicing of exons IIIb and IIIc of fibroblast growth factor receptor 2 (FGFR2). Inclusion of the 148-nt IIIb exon or the 145-nt IIIc exon in the region encoding the second half of the third extracellular ligand binding domain results in either FGFR2-IIIb or FGFR2-IIIc, respectively (Fig. 1). For a given cell type the splicing pattern is exclusive and only one of the isoforms is expressed (34, 35). Although both receptors bind FGF-1 with equivalent affinity, they otherwise display distinct and exclusive differences in binding to most other FGFs thus far characterized (36). Appropriate tissue-specific expression of FGFR2-IIIb or FGFR2-IIIc is crucial for maintenance of normal tissue homeostasis and dysregulated splicing of this transcript has been proposed to be one event that can lead to progression of cancers of epithelial origin, including prostate and bladder cancers (34, 37-39). Several studies have utilized rat and human FGFR2 minigenes to identify cis-elements in the introns and exons of these transcripts that influence splicing of exons IIIb and IIIc and are shown schematically in Fig. 1 (26, 40-44) Within the intron separating exons IIIb and IIIc, the ISAR (intronic splicing activator and repressor) element in rat transcripts was shown to play a dual role in splicing regulation; activation of the upstream exon IIIb and repression of the downstream exon IIIc in cells that express FGFR2-IIIb (40). A nearly identical sequence in human transcripts termed IAS3 displays the same splicing regulatory activities (43, 45). The function of IAS3 was proposed to involve formation of an RNA secondary structure with another element upstream in the same intron termed IAS2 (43). A rat sequence that is identical to IAS2 is likewise predicted to form a secondary structure with ISAR and evidence has been put forth that suggests this secondary structure also is involved in splicing regulation of rat transcripts (46). Although this rat element has previously been referred to as rIAS-2, we will henceforth refer to the element as ISE-2 to conform with a convention using "ISE" to refer to intronic sequences that positively regulate splicing. The ability of ISE-2 and/or ISAR to regulate FGFR2 splicing has thus far only been demonstrated in cells that express FGFR2-IIIb. It is therefore of great interest to identify proteins that interact with these elements as they may include proteins that are exclusively expressed in cell types that splice exon IIIb. Given data that suggest that these elements function through formation of an RNA secondary structure, it is important to perform a systematic analysis of the precise sequence and structural requirements of these elements that are required for their ability to regulate splicing.

In this study, we make use of cell lines, DT3 and AT3, that have previously been used to study FGFR2 splicing regulation (26, 40, 46, 47). DT3 cells exclusively express FGFR2-IIIb and AT3 cells express FGFR2-IIIc and transfected minigenes that recapitulate the splicing pattern of the endogenous FGFR2 gene transcript have been described (40). We performed a comparative sequence analysis of elements that are related to ISE-2 and ISAR. We found that these intronic sequences are highly conserved among vertebrates. Furthermore, we found related sequences in the corresponding intron of fibroblast growth factor receptor 1 (FGFR1). The predicted secondary structures by these related sequences form 17-21-nt base-paired stems containing short internal loops or bulges. In the case of rat FGFR2 ISE-2 and ISAR elements, we found that deletion of sequences that form a bulge in a predicted 18-nt stem does not impair the function of these elements in FGFR2 splicing regulation. We then determined the effect of replacing each element with unrelated sequences that would also form a complementary stem. Surprisingly, we found that these unrelated sequences were equally capable of mediating cell-type-specific splicing regulation of exons IIIb and IIIc. Thus, we conclude that the sequences of ISE-2 and ISAR themselves are not required for the function of these elements. Rather, the function of these elements can be recapitulated by any RNA sequences that form related RNA secondary structures. Thus, a non-sequence-specific RNA secondary structure constitutes the functional element that mediates splicing regulation by ISE-2 and ISAR.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Plasmid Construction-- All plasmids and minigene constructs employed standard cloning techniques as previously described (40). The pI-11-FS and pI-11-NC-W/W37 (the latter previously denoted as pI-11-FS-NC-SAR-20) splicing constructs were obtained as previously described (40). pI-11-pPANC-W/W37 was created as a modification of pI-11-NC-W/W37 by inserting a PacI restriction site directly 5' to the identified ISE-2 region (5'- ACACCCGTAAAAAGGTACAA-3') and AscI and BspEI sites were inserted immediately 3' of this ISE-2 region. By convention all minigenes derived from pI-11-PANC are described with a suffix in which sequences in the position of ISE-2 precede sequences in the position of ISAR, separated by a hash mark. Sequences inserted in these positions are as summarized in Fig. 3A. Thus, minigenes in which ISAR is modified (pI-11-PANC: W/M2, W/M3, W/B37, W/G19, W/B19, W/W19, and W/WS) were achieved by deleting the 37-nucleotide sequence containing ISAR (W37 or SAR-20) from pI-11-pPANC-W/W37 with NotI and ClaI and inserting annealed oligonucleotides with complementary NotI and ClaI sites, as represented by the following oligonucleotide pairs: M2: Mut2-F, 5'-GGCCGCCAAAGAGAACGGACTCTGTGGGCTGAAAGATCCATGTAT-3'and Mut2-R, 3'-CGATACATGGATCTTTCAGCCCACAGAGTCCGTTCTCTTTGGC-3'; M3: Mut3-F, 5'-GGCCGCCAAAGAGAACGGACTCTGTGGGCTGATTTTTCACGCTAT-3' and Mut3-R, 5'-CGATAGCGTGAAAAATCAGCCCACAGAGTCCGTTCTCTTTTGGC-3'; B37: Blu-F, 5'-GGCCGCAACTGGTGGCCTAACTACGGCTACACTAGAAGGACACAT-3'and Blu-R, 5'-CGATGTGTCCTTCTAGTGTAGCCGTAGTTAGGCCACCACTTGC-3'; G19: Globin-F, 5'-GGCCGCACCTATGTCTCCACATGCCAT-3' and Globin-R, 5'-CGATGGCATGTGGAGACATAGGTGC-3'; B19: Blue19-F, 5'-GGCCGCCCGTATCGTAGTTATCTACAT-3'and Blue19-R, 5'-CGATGTAGATAACTACGATACGGGC-3'; W19: SAR-38-F, 5'-GGCCGCTGTGGGCATTTTTCCATGTAT-3'and SAR-38-R, 5'-CGATACATGGAAAAATGCCCACAGC-3'; WS-F, 5'-GGCCGCCCATGGAAAAATGCCCACAAAT-3'and W'-R, 5'-CGATGTGTGGGCATTTTTCCATGGGC-3'. The ISE-2 mutant constructs (pI-11-pPANC: M2'/W37, M3'/W37, B37'/W37, G19'/W37, B19'/W37, and W37S/W37) were created by deleting ISE-2 from pI-11-pPANC-W/W37 with PacI and AscI and inserting annealed oligonucleotides with complementary PacI and AscI sites, as represented by the following oligonucleotide pairs: M2': M2'-F, 5'-TAACCATGGATCTTTGCCCACAAGG-3' and M2'-R, 5'-CGCGCCTTGTTGGCAAAGATCCATGGTTAAT-3'; M3': M3'-F, 5'-TAACGCGTGAAAAATGCCCACAAGG-3'and M3'-R, 5'-CGCGCCTTGTGGGCATTTTTCACGCGTTAAT-3'; B37': IS-BLU-F, 5'-TAAGTGTCCTTCTAGTGTAGCCGTAGTTAGGCCACCACTTGG-3' and IS-BLU-R, 5'-CGCGCCAAGTGGTGGCCTAACTACGGCTACACTAGAAGGACACTTAAT-3'; G19': Globin'-F, 5'-TAAGGCATGTGGAGACATAGGTGGG-3'and Globin'-R, 5'-CGCGCCCACCTATGTCTCCACATGCCTTAAT-3'; B19': BLU19'-F, 5'-TAATGTAGATAACTACGATACGGGG-3'and BLU19'-R, 5'-CGCGCCCCGTATCGTAGTTATCTACATTAAT-3'; W37S: W37'-F, 5'-TAACTGTGGGCTGATTTTTCCATGTGG-3'and W37'-R, 5'-CGCGCCACATGGAAAAATCAGCCCACAGTTAAT-3'. Plasmids pI-11-pPANC: M2'/M2, M3'/M3, B37'/B37, G19'/G19, B19'/B19, and W37S/WS were generated by deleting wild type ISE-2 (W) from pI-11-pPANC: W/M3, W/M4, W/B37, W/G19, W/B19, and W/WS with PacI and AscI and inserting, respectively, annealed oligonucleotides corresponding to M2', M3', B37', G19',B19', and W37S as described above. Minigene pI-11-PANC-del-AN (13) was generated by digesting pI-11-PANC-W/W37 with AscI and NotI to delete sequences between these sites in intron 8, followed by gel purification, blunting of the digested ends with Pfu polymerase (Stratagene), and religation using methods previously described (40). Minigene pI-11-PANC-del-AN (17) was generated by digesting pI-11-PANC-W/W37 with AscI and NotI to delete sequences between these sites in intron 8 and replacing with the following annealed oligos: FGFR2-AN-F, 5'-CGCGCCGC-3' and FGFR2-AN-R, 5'-GGCCGCGG-3'. All plasmid minigene constructs were prepared with Qiagen plasmid maxi kits. Sequences of all the minigenes were confirmed by sequence analysis using the University of Pennsylvania sequencing facility.

Genomic Sequence Data Acquisition-- Complete human and mouse genomic sequence data corresponding to the intron between exons IIIb and IIIc of FGFR1 and FGFR2 was obtained from public sequencing projects through the National Center for Biotechnology Information. The same sequences from rat and frog (Xenopus laevis) were obtained using long distance PCR with primers corresponding to exons IIIb and IIIc from each species, cloning into PCR-Blunt-Zero (Invitrogen), and sequencing. Identification of conserved intronic sequences was carried out using multiple alignment comparisons (Megalign; DNASTAR) and by visual inspection.

Cell Culture and Transfection-- AT3 and DT3 cells were maintained in Dulbecco's Modified Eagle Medium (Invitrogen) supplemented with 10% fetal bovine serum (LTI). Transfections were carried out using LipofectAMINE 2000 (Invitrogen) according to the supplier's recommendations. Selection was performed using Geneticin (Invitrogen) at an active concentration of 400-500 µg/ml until isolated colonies were obtained, and no cells remained from a control transfection with a plasmid lacking neomycin resistance. Pooled colonies were then harvested for RNA preparation.

RNA Purification and RT-PCR Analysis-- Total cellular RNA and RT-PCR was performed as previously described with several exceptions described below (40). Primers PI-11(-H3)-F (5'-GCTGGAATTCGAGCTCACTCTCTTC-3') and PIP11-R (5'-CCCGGGACTAGTAAGCTTAGGCTCTTGGCGTT-3') were used for analysis of the transfected minigenes and were complementary to sequences in the upstream and downstream adenoviral exons contained in all minigenes. PCR amplifications were performed using conditions previously described but with cycles consisting of initial denaturation at 94 °C for 5 min, followed by 25-30 cycles of denaturation at 94 °C for 30 s, annealing at 55° for 30 s, and extension at 72 °C for 1 min. After completion of the final cycle, there was a final extension at 72 °C for 7 min. In all PCR reactions, a water and RT control were included and were negative in all presented data. The PCR products were digested with either AvaI or HincII (New England Biolabs) restriction endonucleases. Aliquots representing equal amounts of each digested or undigested PCR reaction mixtures were directly loaded on non-denaturing 5% polyacrylamide gels and electrophoresed at 110 V for 3-4 h, followed by drying and autoradiography (Amersham Biosciences Hyperfilm-MP). Quantification of data was performed with a Molecular Dynamics PhosphorImager. Because equal amounts of AvaI- and HincII-digested PCR products were loaded onto each gel, quantification of cDNAs containing exon IIIB (UBD or UCD, respectively, where U and D are the 5' and 3' exons of pI-11) was obtained by using the quantification of the band at 286 bp, which remained following HincII digestion as the numerator. The denominator consisted of the sum of both bands (286 and 283) that remained following AvaI and HincII digestion (UBD + UCD). When these results were also expressed with the contribution of products with IIIB and IIIC skipped, the average value of the 138-bp band was also used in the sum of the denominator (UBD + UCD + UD), corrected for molar equivalents.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Regulatory FGFR2 cis-Elements Predicted to Form a Secondary Structure Are Highly Conserved at a Sequence and Structural Level-- Previous studies of rat and human FGFR2 splicing regulation have implicated several RNA cis-elements that are required for appropriate cell type-specific splicing of exons IIIb and IIIc (Fig. 1). Within the intron between these two alternative exons, intron 8, the previously described ISAR in rat transcripts or IAS3 in human transcripts, is required for exon IIIb inclusion in cells that express FGFR2-IIIb (40, 43). We currently use ISAR to refer to a 20-nucleotide element located from 945 to 964-nt downstream of exon IIIb, which corresponds to the approximate location of IAS3 in human FGFR2. The human IAS2 is located over 700-nt upstream of IAS3 (43). The rat ISE-2 sequence is located in the same approximate location as IAS2 in intron 8 (191-208-nt downstream of exon IIIb) (40, 46). The predicted secondary structure formed by ISE-2 and ISAR as well as the related structure from the homologous human elements is shown in Fig. 2B. It is noteworthy that the ISE-2 and ISAR are separated by 736 nt suggesting that a loop of this size separates a putative stem formed between these elements (775 nt in human transcripts). ISE-1 and ISE-3 are also required for efficient exon IIIb splicing in DT3 cells. In order to further investigate the hypothesis that ISE-2 and ISAR function through formation of an RNA stem structure, we also obtained sequence from the same FGFR2 intron from mouse (Mus musculus) and frog (X. laevis) and identified highly similar sequences that correspond to these two elements. These sequences are aligned in Fig. 2A and it can be seen that these intronic elements are highly conserved at the sequence level. For each organism the ability of the ISE-2 and ISAR element to form related secondary structures was determined using the Mfold version 3.1 program (48). As shown in Fig. 2B, the human, mouse, and rat sequences predict a highly similar stem structure with slightly variant internal loops or bulges. The frog sequences displayed the greatest difference with the rat sequences of the ISE-2 and ISAR elements, yet when secondary structure was determined, they nonetheless also formed a similar secondary structure that, like the other sequences, consisted of two interrupted stems. Of note, the most highly similar structures were those of rat and mouse, which consisted of 18 Watson-Crick base-paired bases with a 2-nucleotide bulge. The only two bases that differed between rat and mouse ISE-2 or ISAR were the two nucleotides that comprised the non-base-paired bulge in each element. This observation, together with the more divergent internal loops contained within the human and frog secondary structures suggested that sequences and/or structures of these non-base-paired residues are not critical for their function.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Schematic representation of alternative splicing variants of FGFR2. At the top is a protein domain map with the region encoded by mutually exclusive exons IIIb or IIIc indicated by a thick line. A dark box indicates an acidic domain and the hatched box indicates the transmembrane domain (TM). Ig, immunoglobulin-like domain. TK, tyrosine kinase domain. At the bottom is a map of the pre-mRNA in the region containing exons IIIb and IIIc. Hatched boxes represent intronic splicing elements shown to influence splicing of exon IIIb or IIIc. ISE, intronic splicing enhancer. ISS, intronic splicing silencer. ESS, exonic splicing silencer. The human elements analogous to the rat ISE-1, ISE-2, and ISAR elements are shown in parentheses. The horizontal arrows and dashed line underneath the pre-mRNA represent the FGFR2 genomic sequence that was amplified by PCR and used to generate minigene pI-11-FS.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Sequences similar to ISE-2 and ISAR are highly conserved and predict similar stem structures in FGFR1 and FGFR2 transcripts. A, alignment of FGFR1 and FGFR2 sequences from the intron between exon IIIb and IIIc from human, mouse, rat, and frog that are highly similar to ISE-2 and ISAR. Sequences designated as ISE-2 are shown at left and those designated as ISAR are shown at right. Stars at the top of each alignment represent nucleotides that are identical in all four species. B, secondary structures predicted to form between ISE-2 and ISAR from FGFR2 transcripts. C, secondary structures predicted to form between ISE-2 and ISAR from FGFR1 transcripts. The loops drawn at the top of each stem structure are not at scale as they represent 500 or more nucleotides between the elements in each transcript.

Sequences Homologous to ISE-2 and ISAR Are Present between Alternatively Spliced Exons IIIb and IIIc of FGFR1-- Alternative splicing of exons that represent homologues of exons IIIb and IIIc has been observed for FGFR1, FGFR2, and FGFR3. The two members of the group with the highest level of sequence and genomic structural similarity are FGFR1 and FGFR2 (49). In addition, the exon and intron sizes in the genomic region that encodes exons IIIb, IIIc, and the constitutive upstream and downstream exons are highly similar. Alternative splicing of FGF-R1 IIIb and IIIc exons is highly regulated and cell types with exclusive expression of FGF-R1 (IIIb) have been described (50). FGFR1 and FGFR2 most likely arose from a gene duplication event of a single common ancestral gene that followed a previous exon duplication that gave rise to alternative exons IIIb and IIIc. We thus speculated that sequence elements that direct splicing regulation of FGFR1 exons IIIb and IIIc may be conserved between these gene paralogues. Sequence analysis of the intron between FGFR1 exons IIIb and IIIc from human, mouse, rat, and frog DNA revealed the presence of sequence elements highly similar to ISE-2 and ISAR in the same approximate locations within the intron (Fig. 2A). Thus, for example, the rat intron separating exons IIIb and IIIc of FGFR1 is 1004 nt (compared with 1200 for FGFR2) and homologues to ISE-2 and ISAR are positioned between 126-143 and 696-717 downstream of exon IIIb, respectively. We used Mfold to determine whether these elements could also form similar secondary structures. As shown in Fig. 2C, secondary structures that were similar to those formed with the FGFR2 elements were predicted. Thus, for example, the rat element contained a base-paired region at the base of the stem that was identical to that of FGFR2, whereas sequences that were not identical to the FGFR2 ISE-2 and ISAR elements nonetheless predicted a different base-paired region with an intervening small internal loop. In the case of the human structure, two additional small bulges were predicted in the top of the secondary structure that would be predicted to disrupt stem formation in this region, yet the extent of base pairing could be further extended to yield a total of 21 base-paired residues compared with 19 for the frog structure and 17 for the mouse and rat structures. Thus, in general, these phylogenetic comparisons were consistent with a hypothesis that ISE-2 and ISAR elements regulate splicing through formation of a secondary structure that arises through base pairing interactions between these elements. Furthermore, these results suggest that the sequences and/or structures that comprise these elements have been maintained through a long evolutionary history.

Minigene pI-11-PANC-W/W37 Recapitulates the Splicing Pattern of the Endogenous Gene in AT3 and DT3 Cells-- In order to further investigate the sequences and/or structures of ISE-2 and ISAR that are required for splicing regulation we further modified a minigene, pI-11-FS, that was previously shown to recapitulate the splicing pattern of endogenous FGFR2 when stably transfected into AT3 and DT3 cells (40). These minigenes contain a 1804 nt region of FGFR2 that includes exons IIIb and IIIc as well as flanking intron sequences upstream of exon IIIb, downstream of exon IIIc, and the entire intron (intron 8) that separates the exons. This genomic region was positioned in the intron of an adenoviral expression cassette that consists of two constitutively spliced adenoviral exons (Fig. 3A). Pools of stably transfected AT3 and DT3 cells were used to harvest RNA and used with a previously validated RT-PCR assay that accurately distinguishes between spliced RNAs that contain exon IIIb or exon IIIc. Using minigene-specific PCR primers a predominant product of 286 or 283 bp is observed when spliced products contain exon IIIb or exon IIIc, respectively. In order to distinguish whether these products contain exon IIIb or IIIc, they are digested with AvaI and HincII, which specifically digest only those products that contain exon IIIb or IIIc, respectively. As shown previously, transfection of pI-11-FS into DT3 and AT3 cells yields products that nearly exclusively contain exon IIIb or IIIc, respectively (Fig. 3B, lanes 1-3, and C, lanes 1-3). Of note, two additional PCR products are seen in these transfections. One such product contains both exon IIIb and IIIc (431 bp) and another consists products that represent mRNAs in which the adenoviral exons have been directly ligated and both exon IIIb and IIIc skipped (138 bp). As will be described, one consequence of loss of splicing regulation in DT3 cells is a switch to exon IIIc inclusion, but also an increase in products in which both exon IIIb and IIIc are skipped. Therefore, using phosphorimager analysis we quantify the percentage of exon IIIb containing products relative to that of exon IIIc, but also relative to that of exon IIIc together with skipped products.

View larger version (38K): [in this window] [in a new window]   Fig. 3.   Splicing regulation is maintained following insertion of restriction sites immediately flanking ISE-2 and ISAR in rat FGFR2 minigenes. A, schematic demonstrating the construction of pI-11-PANC-W/W37. At top is shown minigene pI-11-FS. Boxes represent exons and lines indicate introns. CMV, cytomegalovirus promoter; pA, polyadenylation signal. Exons IIIb and IIIc as well as the adenoviral exons (Adeno) are indicated. In the middle is a representation of FGFR2 intron 8 showing the ISE-2 and ISAR elements after the indicated restriction sites have been introduced flanking each element. (Note, boxes in this region indicated the intron modifications, not exons.) At bottom the sequences positioned in the respective locations of derived minigenes are shown. Sequences shown in bold represent mutations of either element or unrelated sequences used to replace ISE-2 or ISAR. B, minigene pI-11-PANC-W/W37 maintains exon IIIb splicing when stably transfected into DT3 cells, but with slightly reduced efficiency compared with pI-11-FS and pI-11-NC-W/W37. Results for each minigene represent RT-PCR products that are undigested (U), digested with AvaI (A), or digested with HincII (H) as shown above each lane in this figure. All subsequent results use the same order of undigested, AvaI-digested, and HincII-digested products. M, pBR/MspI molecular weight markers. At right spliced products represented by each band are shown. U and D indicate upstream and downstream adenoviral exons, respectively. Quantified results are shown below each set of results using percentages of exon IIIb inclusion as described under "Experimental Procedures." C, exon IIIc, and not exon IIIb, is included in spliced products from these minigenes in AT3 cells.

In order to further characterize the ISE-2 and ISAR regulatory sequences of the ISE-2 element we created a new minigene pI-11-PANC-W/W37. We had previously generated a minigene, pI-11-FS-NC: SAR-20 in which a 57-nt region from intron 8 that contained ISAR was deleted and replaced with a shorter 37-nucleotide region (W37) that contained all functional ISAR sequences, but flanked by NotI and ClaI restriction sites at the 5'- and 3'-end of the element. This minigene, designated here as pNC-11-W/W37, also recapitulated cell-specific splicing of exons IIIb and IIIc (Fig. 3B, lanes 4-6 and C, lanes 4-6). The minigene pPANC-W/W37 was generated by further modifying pNC-W/W37 by engineering PacI and AscI restriction sites immediately 5' and 3' of a 20-nt sequence containing ISE-2. Therefore, this minigene contains both ISE-2 and ISAR, but with restriction sites flanking each element. Because the wild-type ISE-2 and ISAR elements are present in this minigene we use a convention W/W37 in which W refers to wild type sequences and the hash mark separates the sequences contained, in order, in the position of ISE-2 and ISAR. This minigene and other derived minigenes are displayed schematically in Fig. 3A. When pI-11-PANC-W/W37 was transfected into AT3 cells we again observed nearly exclusive use of exon IIIc (Fig. 3C, lanes 7-9). However, in DT3 cells we noted that a slightly increased proportion of products contained exon IIIc and also an increase in skipped products when compared with pI-11-FS or pI-11-NC-W/W37 (Fig. 3B, lanes 7-9). Thus, it is evident that introduction of either the PacI or AscI site slightly impairs the function of ISE-2 and ISAR to mediate exon IIIb inclusion in DT3 cells. Nonetheless, this construct still clearly recapitulates cell specific regulation and allows ISE-2 and ISAR to be directly manipulated to further investigate the role of these sequences on splicing regulation. This new construct was thus subsequently used as a positive control for further investigation of the role ISE-2 and ISAR play in splicing regulation. Because all minigenes to be described in subsequent sections all yielded exclusively exon IIIc containing spliced products in transfected AT3 cells, we will henceforth only show results from transfections in DT3 cells.

Mutations in ISAR Predicted to Disrupt a Proposed Secondary Structure Cause Loss of Splicing Regulation and Complementary Mutations in ISE-2 Restore Splicing Regulation-- We sought to further test whether formation of a secondary structure between ISE-2 and ISAR is involved in regulation of rat FGFR2 splicing as previously proposed for the related cis-elements in human FGFR2 transcripts (43). We first determined whether mutations in ISAR that result in loss of splicing regulation in DT3 cells can be "rescued" by making corresponding complementary mutations in ISE-2. A similar approach using human minigenes showed that complementary mutations increased exon IIIb splicing, although different pairs of complementary mutations predicted to restore base pairing between these elements did not all restore exon IIIb inclusion to the levels obtained with the wild-type sequences (43). Two previously described mutations in ISAR, Mut 2 and Mut 3, were previously shown to result in a significant loss of exon IIIb inclusion (40). These mutations are shown next to the wild type ISAR sequence in the context of the putative secondary structure in Fig. 4A and it can be seen that either mutation would result in disruption of a stem formed at the base of this structure. These mutated sequences (denoted M2 and M3) were placed in pI-11-PANC in place of wild-type ISAR (W37), and complementary mutations (M2' and M3') were used to replace the wild-type ISE-2 (W). Introducing any of these ISE-2 or ISAR mutations led to a reduction in exon IIIb inclusion in stably transfected DT3 cells when compared with pI-11-PANC-W/W37 (Fig. 4, B and C, lanes 4-9). In Fig. 4 and subsequent figures, a representative experiment is shown at the top and the average percentage of exon IIIb inclusion from four independently performed transfections is shown graphically at the bottom. It should be noted that some reductions in exon IIIb splicing in these experiments appear small, but it is worth noting that in addition to a proportional increase in products that include exon IIIc, the proportion of skipped products (containing neither exon IIIb nor exon IIIc) is greatly increased by these mutations. However, when the ISAR mutations (M2 and M3) were tested together with complementary ISE-2 mutations (M2' and M3') exon IIIb splicing was restored (Fig. 4, B and C, lanes 10-12).

View larger version (47K): [in this window] [in a new window]   Fig. 4.   Complementary mutations in ISE-2 or ISAR that restore a predicted RNA stem structure maintain exon IIIb inclusion in DT3 cells. A, schematic of the predicted secondary structure between rat FGFR2 ISE-2 and ISAR with the indicated mutations positioned adjacent to the sequences they replace. B, mutations M2 or M2' result in loss of splicing regulation, but when both mutations are present together, splicing regulation is restored. C, mutations M3 or M3' result in loss of splicing regulation, but when both mutations are present together, splicing regulation is restored. D, switching the positions of ISE-2 and ISAR maintains splicing regulation. WS indicates wild-type ISE-2 in the position normally occupied by ISAR. W37S indicated wild-type ISAR in the position normally occupied by ISE-2. Abbreviations and lane designations are as described in the legend to Fig. 3. In this and subsequent figures, data from a representative experiment is presented at the top. Bar graphs represent the results from four independently performed sets of transfections. Numbers in the bars represent the mean % IIIb inclusion from all experiments, and error bars indicate S.D.

We also hypothesized that if formation of a secondary structure determines the ability of ISE-2 and ISAR to function in splicing regulation, the positions of these elements should be able to be switched without impairing the ability of the elements to function together to regulate exon IIIb inclusion. We therefore generated constructs in which ISAR (W37S) was used to replace ISE-2 and vice versa to generate minigene construct pI-PANC-W37S/WS. As controls, either element was present in both positions (pI-11-PANC-W/WS and pI-11-PANC-W37S/W37; note the designation "S" indicates the same sequence, but whose position has been switched in the minigenes). As shown in Fig. 4D, switching the positions of these elements indeed preserved splicing regulation whereas duplication of either element was associated with loss of splicing regulation. In fact, the level of exon IIIb inclusion when these positions were reversed was higher than that of pI-11-PANC W/W37.

ISAR Sequences Predicted to Base Pair with ISE-2 Are Sufficient to Mediate Function of This cis-Element-- As shown in Fig. 2B, the putative secondary structure between rat ISAR and ISE-2 consists of an 18-nt base-paired stem with an internal 2-base bulge. The 37-nt ISAR sequence previously described (W37) encodes additional nucleotides that normally flank the sequences that are contained within the secondary structure. We replaced W37 with W19 in which the two bases that should form a bulge have been deleted as have all nonbase-paired ISAR sequences except a 3'-terminal U. Thus an uninterrupted 18-nucleotide stem is predicted to result if a secondary structure forms from ISE-2(W) and W19. The predicted secondary structure that would result from replacing W37 with W19 is shown in Fig. 5A. Transfection of the resulting minigene, pI-PANC-W/W19, in DT3 cells indicates that deletion of nonbase-paired sequences, including those encoding a putative bulge in the stem, does not impair splicing regulation and in fact exon IIIb inclusion was reproducibly higher (Fig. 5B, lanes 1-3, compare with lanes 1-3 from Fig. 4, B, C, or D).

View larger version (31K): [in this window] [in a new window]   Fig. 5.   An 18-19-nucleotide base-paired stem is sufficient to mediate splicing regulation in DT3 cells. A, schematic presentation of secondary structures predicted to form in pI-11-PANC-W/W37, pI-11-PANC-W/W19, pI-11-PANC-G19'/G19, and pI-11-PANC-B19'/B19. The bulge containing the UG dinucleotide is boxed. B, splicing regulation is maintained when the bulged nucleotides are removed and when ISE-2 and ISAR are replaced by unrelated sequences that are predicted to form a perfect 19-nucleotide stem structure. Abbreviations and lane designations are as described in the legend to Fig. 3.

The Splicing Regulatory Functions of ISE-2 and ISAR Can Be Recapitulated When They Are Replaced by Randomly Selected Sequences That Also Form a Stem Structure-- The predicted FGFR2 stems derived from ISE-2 and ISAR shown in Fig. 2B consist of interrupted stems that contain between 17 and 20 base-paired nucleotides. The base-paired nucleotides most distal to the loop (the bottom of the structures as shown) display the highest degree of sequence conservation both between these species as well as with putative FGFR1 elements. Sequences located toward the top of the structures display less sequence conservation, but in general maintain the potential to form base pairing interactions that maintain stems of similar overall length. Given the high degree of sequence conservation of portions of each element we suspected that while an 18-20-nucleotide stem structure may indeed be required for function of ISE-2 and ISAR, it was still possible that specific sequences within this stem may nevertheless be required in order to regulate splicing. In order to determine whether any specific sequences of ISE-2 and ISAR are required for the function of this element we selected two unrelated 19 nucleotide sequences, one from the second intron of human -globin (G19) and the other (B19) from the pBluescript (Stratagene) plasmid sequence. These sequences were used to replace ISAR in pI-11-PANC-W/W37. We then replaced ISE-2 in the resulting plasmids with sequences, G19' and B19', that are predicted to form a perfect 19-nt stem with the G19 or B19 sequences, respectively. These sequences and the predicted RNA stem structures are shown in Fig. 5A. As controls, each of these sequences was also cloned into pI-11-FS-PANC in all combinations that would not form a 19 nt stem (e.g. G19 replacing ISAR and B19' replacing ISE-2). The resulting plasmids were stably transfected into DT3 cells and exon IIIb inclusion determined by RT-PCR. The results were then compared with the plasmid in which a bulgeless stem comprised of ISE-2 and ISAR sequences was predicted to form (pI-11-PANC-W/W19). Surprisingly, both random sequences (G19 or B19), when present together with their complementary sequence (G19' or B19') were able to restore exon IIIb splicing in place of the ISE-2 and ISAR sequences (Fig. 5B, lanes 10-12 and 19-21). Replacing any of these sequences in combinations that did not predict a 19-nt stem gave results that were equivalent to that seen in which either ISE-2 or ISAR was deleted. Therefore, it is evident that the functions of ISE-2 and ISAR can be accounted for solely by their ability to form an RNA stem structure and there is no specific requirement for any of the sequence content of either element for FGFR2 splicing regulation.

Sequences in a Putative Loop between ISE-2 and ISAR Are Not Required for Splicing Regulation-- A number of RNA stem structures have been shown to function by presenting loop sequences in a single-stranded conformation that facilitates recognition of the loop sequences by RNA-binding proteins. For example, the U1A protein binds to a conserved loop sequence in stem-loop II of the U1 snRNA (51). While such loop sequences are generally present in much shorter loops than the ones proposed here, we sought to determine whether any of the sequences between ISE-2 and ISAR are required for cell specific activation of exon IIIb splicing. The ability to introduce unrelated restriction sites downstream of ISE-2 and upstream of ISAR suggests that FGFR2 loop sequences do not need to be immediately adjacent to the proposed stem. It was previously shown that a deletion of 611 of the 736 nucleotides between ISE-2 and ISAR can be deleted without loss of splicing regulation (Fig. 1, non-regulatory RNA) (40). We made a deletion from the AscI site to the NotI site of pI-11-PANC W/W37 that removes essentially all FGFR2 intron 8 sequences between ISE-2 and ISAR. The resulting minigene, pI-11-PANC-del-AN (13), contains 13 nucleotides between the base-paired regions of ISE-2 and ISAR of which 12 are derived from the AscI and NotI restriction sites. Transfection of this minigene results in cell-type-specific inclusion of exon IIIb in DT3 cells (Fig. 6, lanes 4-6). We also inserted 4 nucleotides that increased the size of the loop from 13 to 17 nucleotides in minigene pI-11-PANC-del-AN (17) and still observed maintenance of splicing regulation (Fig. 6A, lanes 7-9). Although a single nonbase-paired A residue at the end of ISE-2 is present in the loops predicted in these constructs, we conclude that the length and sequence of FGFR2 sequences that comprise the loop between ISE-2 and ISAR are not critical in cell-type specific activation of exon IIIb splicing.

View larger version (40K): [in this window] [in a new window]   Fig. 6.   Deleting all FGFR2 intron 8 sequences between ISE-2 and ISAR does not impair splicing regulation. The numbers in parentheses in the deletion constructs (13 and 17) indicates the size of the resulting loop between ISE-2 and ISAR compared with a 742-nucleotide loop in pI-11-PANC-W/W37. Abbreviations and lane designations are as described in the legend to Fig. 3.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

In this report we have investigated in detail the sequence and structural components of two intronic cis-elements, ISE-2 and ISAR, which are required for splicing regulation of rat FGFR2 transcripts. Deletion or mutation of either ISE-2 or ISAR results in a significant decrease in exon IIIb inclusion and an increase in exon IIIc inclusion in DT3 cells, suggesting that each element is involved in cell-type-specific activation of exon IIIb splicing and repression of exon IIIc splicing (40, 46). On the other hand deletion of either element has no effect on cell-type-specific exon IIIc inclusion in AT3 cells or other cells that express FGFR2-IIIc. Thus, it has become evident that these elements are required for splicing regulation only in cells that express FGFR2-IIIb. It has previously been proposed that two equivalent human elements (IAS2 and IAS3) function through formation of an RNA secondary structure consisting of 17-18-base-paired nucleotides from each element (43, 46). We have used phylogenetic comparisons of similar sequences from different species as well as the FGFR1 paralogue of FGFR2 and such analysis further supports the hypothesis that splicing regulation by these elements involves formation of a stem structure. Previous analysis of mutations that would disrupt stems formed by the human IAS2 and IAS3 elements showed that such mutations impaired splicing regulation and complementary mutations that would restore the stems likewise restored splicing regulation (43). However, not all of the complementary mutations restored splicing regulation to the levels achieved with the wild type sequences, leading these authors to suggest that some specific sequences within these elements may also be required for efficient splicing regulation. To systematically determine the degree to which a secondary structure as well as the primary sequence content of each element is required for splicing regulation, we determined the effects of complementary mutations in the rat elements, removing the internal bulged nucleotides, and of replacing both elements with unrelated sequences that would maintain stem formation. We found that removal of the bulged nucleotides from ISAR did not impair splicing regulation. Then, to our initial surprise, we also found that replacing ISE-2 and ISAR with completely unrelated sequences that form a stem of similar length maintained splicing regulation that was equivalent to that seen with the wild-type sequences. In studies of the related human RNA structure formed by IAS2 and IAS3 it was hypothesized that some sequence elements within the elements as well as the RNA structure were required for efficient splicing regulation (43). In contrast, we conclude that the primary sequences of ISE-2 and ISAR are not required for splicing regulation, but rather sequences positioned in their location within the intron influence FGFR2 splicing solely through formation of a secondary structure consisting of 17-20 base-paired nucleotides.

Consideration of possible mechanisms through which ISE-2 and ISAR influence splice site selection needs to be considered in the context of other nonsplice-site cis-elements that affect splicing of exons IIIb and IIIc (summarized in Fig. 1). An ESS in exon IIIb as well as an ISS upstream of exon IIIb also reduce the efficiency of exon IIIb splicing (26, 41, 42, 52, 53). It has been shown that hnRNPA1 and polypyrimidine tract-binding protein (PTB) play a role in the splicing repression mediated by the ESS and ISS, respectively. However, the repressive effect of these elements on exon IIIb splicing are seen in cell types that express FGFR2-IIIb or FGFR2-IIIc, and therefore they alone cannot account for cell-type specific differences in splicing. In AT3 cells, it appears that this repression is sufficient to preclude exon IIIb inclusion, whereas in DT3 cells factors that can overcome this repression are able to enforce exon IIIb splicing. This activity in DT3 cells (or other cells that include exon IIIb) involves the participation of four different intronic sequences (ISEs) that are collectively able to overcome exon IIIb splicing repression. These include ISE-1 (IAS1), ISE-2 (IAS2), ISAR (IAS3) and ISE-3. In human FGFR2, IAS1 is an ~20 nucleotide U-rich sequence located almost immediately downstream of the exon IIIb 5'-splice site and the protein TIA-1 has been proposed to increase exon IIIb splicing through binding to this sequence (29). However, TIA-1 is not tissue specific and, depending on its sequence context, IAS-1 can activate splicing of an upstream exon in cell types that express either FGFR2 isoform. A rat sequence we call ISE-1 is a much longer 45 nucleotide U-rich sequence and has also been shown to activate exon IIIb splicing.² In addition to ISE-2 and ISAR, we also implicate a role of a fourth additional element, ISE-3, located downstream of ISAR that also functions in activating exon IIIb splicing (40). Breathnach and co-workers (29) have also shown that human sequences downstream of the IAS3 sequences that base pair with IAS2 are also involved in regulation, but include these sequences in the element they refer to as IAS3. In rat FGFR2 transcripts, we have shown that distinct sequences downstream, but not contiguous with ISAR are required for efficient exon IIIb inclusion.² Therefore, we refer to such sequences as a separate ISE-3 element. In addition to activating exon IIIb splicing, ISE-2 and ISAR have also been shown to repress splicing of exon IIIc (40, 46). The ability to repress exon IIIc splicing in DT3 cells may be assisted by ESS elements in exon IIIc that are also bound by PTB (45). There is currently no evidence that ISE-2, ISAR, or ISE-3 influence splicing in AT3 (or HeLa) cells regardless of context and thus it is possible that in DT3 cells a cell-type specific factor that interacts with one of these elements may "tip the balance" toward exon IIIb inclusion and exon IIIc skipping (29).

The finding that specific sequences in ISE-2 and ISAR are not required for function other than to generate a double-stranded RNA stem structure has implications for the mechanism through which this secondary structure influences splicing. Numerous studies have proposed models through which RNA secondary structures may be involved in constitutive and alternative splicing. Models have been proposed in which RNA secondary structures that encompass 3'- or 5'-splice sites block recognition of these sites for splicing (54-66). It has also been suggested that secondary structures can influence splice site selection by approximating splice signals across introns (64, 67, 68). In these models the secondary structures themselves are not necessarily bound by proteins that influence splicing, but may simply determine the efficiency with which the constitutive splicing apparatus is able to recognize splice sites in the vicinity of the structure. RNA secondary structures can also influence splicing by facilitating recognition of single stranded RNA sequences in loops predicted to form within the structure (69, 70). Thus far, however, the functions of splicing regulatory proteins have not been shown to involve specific recognition of double-stranded RNA sequences contained within such structures. However, the possibility that RNA-binding proteins can recognize specific bases within a stem was shown for the Drosophila B52 protein. Involvement of B52 in regulated splicing is suggested by several studies, although the in vivo targets have not been elucidated (71, 72). However, using Systematic Evolution of Ligands by Systematic Enrichment (SELEX), high affinity targets of B52 were identified that were predicted to form short conserved stem-loop structures (71). In this case primary sequences in the loop as well as specific sequences in a predicted stem were both required for efficient binding by B52. Thus, there is a precedent for the possibility that binding of target RNAs by RNA-binding proteins can involve recognition of base-paired sequences within secondary structures.

The secondary structure formed by ISE-2 and ISAR is distinct from other structures suggested to influence splicing in several ways. The most obvious difference is the large distance (736 nt) that separates the ISE-2 and ISAR elements that form the stem. The ability of an artificial stem structure to sequester and block splicing of an exon located between them was determined (73, 74). In these experiments the complementary sequences were separated by 285 nucleotides and repression of splicing to an exon located between the elements in vivo was only seen when the predicted stem structures were at least 50 nucleotides in length. The results suggested that stems separated by this distance either do not form with shorter stems or are not sufficiently stable to influence splice site selection. Experiments using shorter artificial stems to block recognition of a 5'-splice site suggested that RNA structures do not generally form if the loop between them is significantly greater than 50 nt (57). A possible explanation for such a limit in loop size was that following transcription there is a limited "window" during which transcribed RNA exists in a naked form before being bound by hnRNP proteins that would preclude formation of such RNA stem structures (57). The ability of ISE-2 and ISAR to form a secondary structure despite containing 18 base-paired nucleotides located far apart from one another in the intron could be accounted for by the function of proteins that facilitate and stabilize the interaction of these elements. Thus, one possible mechanism suggested that proteins that first bind independently to ISE-2 (IAS2) and ISAR (IAS3) are able to interact with each other and then promote formation of the secondary structure (43). However, based on the findings presented here showing that the specific sequences of ISE-2 and ISAR are not required for splicing regulation, this appears less likely since sequence-specific recognition of ISE-2 or ISAR would be central to this hypothesis. Therefore, as an alternative, proteins bound either upstream of ISE-2 and/or downstream of ISAR may facilitate the approximation of these elements. Likely candidate proteins to perform such a function would include members of the hnRNP family. Packaged mRNP is a dynamic structure and the possibility exists that protein-protein interactions between hnRNP proteins positioned at a distance along a transcript may themselves approximate different regions of pre-mRNAs. Thus, although hnRNP may indeed often prevent RNA structures from forming (57), in selected cases they may facilitate the formation of certain RNA structures.

How does a stem structure formed between ISE-2 and ISAR function to regulate splicing? One possibility is that the stem merely functions to approximate other elements; for example by bringing ISE-3 closer to ISE-1 and exon IIIb. However, when such approximation was achieved by deleting intron sequences from the 5' boundary of IAS3 to a position immediately upstream of and including IAS2, exon IIIb splicing activation was still impaired (43). Similarly removing most of the sequence between rat ISE-2 and ISAR as well as ISAR does not maintain splicing regulation (40). Another mechanism through which the structure might regulate splicing would be through direct binding of a regulatory protein to the double-stranded stems formed between ISE-2 and ISAR. An expanding number of RNA-binding proteins that interact specifically with double-stranded RNA (dsRNA) continue to be described (75). These RNA-binding proteins contain a conserved core of 65-75 amino acids that comprise a conserved double-stranded RNA binding domain (dsRBD) (75). Thus far, the interactions of RNA-binding proteins that contain the dsRBD with target RNAs appear to be sequence-independent and only require a dsRNA of sufficient length for binding (76-80). A crystal structure of the second dsRBD of X. laevis X1rbpa has provided information regarding the components of the domain and its RNA target that promote high affinity binding (80). This study determined that a majority of interactions of the dsRBD with RNA involve the phosphodiester backbone and 2'-OH groups, accounting for sequence-independent binding. Furthermore a 16-base pair dsRNA appears to be sufficient to promote binding. Thus, if in fact the role of the secondary structure formed by ISE-2 and ISAR is, in fact, mediated through binding of a dsRNA-binding protein it is perhaps not so surprising that the specific sequences comprising the stem are not important. The length of the ISE-2/ISAR stem would also be consistent with binding by a dsRBD-containing protein. A number of dsRBD containing proteins have been shown to interact with specific RNAs in vivo. For example the Drosophila Staufen protein interacts specifically with 3'-UTR sequences of bicoid and oskar mRNAs and directs their localization (81). However, an unanswered question is how dsRNA-binding proteins can recognize specific transcripts given the sequence-independent binding of the dsRBD. If a specific dsRNA-binding protein is involved in FGFR2 splicing regulation through binding to ISE-2 and ISAR, it is possible that other domains of the protein recognize other intron 8 sequences or structures. Alternatively, an RNA-binding protein (or proteins) bound to specific sequences adjacent to the structure (for example ISE-3) could promote cooperative binding of such a protein to an RNA stem structure through specific protein-protein interactions.

Removing the nucleotides from ISAR that form a bulge in the secondary structure does not impair splicing regulation, and the nonsequence-specific elements tested were also uninterrupted stem structures that were essentially equivalent to the wild-type structure in their ability to maintain splicing regulation. However, all of the predicted wild type structures presented (Fig. 2, B and C) contain either a central bulge or internal loop. Thus, it might be asked why these conserved elements do not also predict uninterrupted stems. One potential answer to this question can be provided by using an analogy to the interaction of dsRNA with the dsRBD of X1rbpa. In this structure, the central nucleotides of the dsRNA binding site are not contacted by the protein (80). Furthermore, natural substrates of dsRBDs have been described in which duplex regions are separated by non-Watson Crick base-paired nucleotides (82). Therefore, if in fact the ISE-2/ISAR structure is bound by a dsRBD-containing protein, the bulges may not be necessary for function, but there also may not have been selective pressures to maintain base-pairing in the central region of the stem.

In conclusion, these studies provide further support of the hypothesis that the rat ISE-2 and ISAR elements, like their IAS2 and IAS3 human counterparts, regulate splicing through formation of base pairing interactions between these elements. We have also found that the minimal components of these elements involve only the 18 nucleotides that form base pairing interactions with one another. However, splicing regulation is also maintained when either element is replaced by a completely unrelated sequence provided that a complementary sequence is provided in place of the other element. To our knowledge, this is the first study to demonstrate the function of a naturally occurring RNA cis-element involved in regulated splicing that involves an RNA structure independent of any specific nucleotide sequences. Taken together these findings suggest a model in which one trans-acting protein component required for FGFR2 splicing regulation is a double-stranded RNA-binding protein. Although numerous RNA-binding proteins have been shown to be involved in regulated splicing, to date a role for a specific double-stranded RNA-binding protein in this process has not been shown. Further identification of proteins that interact with the structure formed by these elements will be required in order to further determine the molecular mechanism through which cell-type specific splicing of FGFR2 is achieved.

	FOOTNOTES

^* This work was supported by start-up funds from the University of Pennsylvania School of Medicine, Department of Defense Grant PC991539, and United States Public Health Services Grant K08 CA72560 from the NCI, National Institutes of Health.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBank^TM/EBI Data Bank with accession number(s) AYI61008 and AYI61009.

To whom correspondence should be addressed: University of Pennsylvania School of Medicine, 700 Clinical Research Bldg., 415 Curie Blvd., Philadelphia, PA 19104-6144. Tel.: 215-573-1838; Fax: 215-898-0189; E-mail: russcars@mail.med.upenn.edu.

Published, JBC Papers in Press, October 21, 2002, DOI 10.1074/jbc.M207409200

² R. Hovhannisyan and R. P. Carstens, unpublished results.

	ABBREVIATIONS

The abbreviations used are: ESE, exonic splicing enhancer; FGF, fibroblast growth factor; FGFR, fibroblast growth factor receptor; ESS, exonic splicing silencer, ISE, intronic splicing enhancer; ISS, intronic splicing silencer; ISAR, intronic splicing activator and repressor; hnRNP, heterogeneous nuclear ribonucleoprotein; PTB, polypyrimidine tract-binding protein; UTR, untranslated region; nt, nucleotide.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				G. Ohno, M. Hagiwara, and H. Kuroyanagi STAR family RNA-binding protein ASD-2 regulates developmental switching of mutually exclusive alternative splicing in vivo Genes & Dev., February 1, 2008; 22(3): 360 - 374. [Abstract] [Full Text] [PDF]

				R. H. Hovhannisyan and R. P. Carstens Heterogeneous Ribonucleoprotein M Is a Splicing Regulatory Protein That Can Enhance or Silence Splicing of Alternatively Spliced Exons J. Biol. Chem., December 14, 2007; 282(50): 36265 - 36274. [Abstract] [Full Text] [PDF]

				N. N. Singh, R. N. Singh, and E. J. Androphy Modulating role of RNA structure in alternative splicing of a critical exon in the spinal muscular atrophy genes Nucleic Acids Res., January 28, 2007; 35(2): 371 - 389. [Abstract] [Full Text] [PDF]

				J. Kraunus, D. Zychlinski, T. Heise, M. Galla, J. Bohne, and C. Baum Murine Leukemia Virus Regulates Alternative Splicing through Sequences Upstream of the 5{middle dot} Splice Site J. Biol. Chem., December 8, 2006; 281(49): 37381 - 37390. [Abstract] [Full Text] [PDF]

				V. I. Bonano, S. Oltean, R. M. Brazas, and M. A. Garcia-Blanco Imaging the alternative silencing of FGFR2 exon IIIb in vivo RNA, December 1, 2006; 12(12): 2073 - 2079. [Abstract] [Full Text] [PDF]

				E. A. Newman, S. J. Muh, R. H. Hovhannisyan, C. C. Warzecha, R. B. Jones, W. L. McKeehan, and R. P. Carstens Identification of RNA-binding proteins that regulate FGFR2 splicing through the use of sensitive and specific dual color fluorescence minigene assays RNA, June 1, 2006; 12(6): 1129 - 1141. [Abstract] [Full Text] [PDF]

				A. P. Baraniak, J. R. Chen, and M. A. Garcia-Blanco Fox-2 Mediates Epithelial Cell-Specific Fibroblast Growth Factor Receptor 2 Exon Choice Mol. Cell. Biol., February 15, 2006; 26(4): 1209 - 1222. [Abstract] [Full Text] [PDF]

				R. H. Hovhannisyan, C. C. Warzecha, and R. P. Carstens Characterization of sequences and mechanisms through which ISE/ISS-3 regulates FGFR2 splicing Nucleic Acids Res., January 12, 2006; 34(1): 373 - 385. [Abstract] [Full Text] [PDF]

				E. A. Glazov, M. Pheasant, E. A. McGraw, G. Bejerano, and J. S. Mattick Ultraconserved elements in insect genomes: A highly conserved intronic sequence implicated in the control of homothorax mRNA splicing Genome Res., June 1, 2005; 15(6): 800 - 808. [Abstract] [Full Text] [PDF]

				R. Canete-Soler, K. S. Reddy, D. R. Tolan, and J. Zhai Aldolases A and C Are Ribonucleolytic Components of a Neuronal Complex That Regulates the Stability of the Light-Neurofilament mRNA J. Neurosci., April 27, 2005; 25(17): 4353 - 4364. [Abstract] [Full Text] [PDF]

				E. J. Wagner, A. P. Baraniak, O. M. Sessions, D. Mauger, E. Moskowitz, and M. A. Garcia-Blanco Characterization of the Intronic Splicing Silencers Flanking FGFR2 Exon IIIb J. Biol. Chem., April 8, 2005; 280(14): 14017 - 14027. [Abstract] [Full Text] [PDF]

				R. H. Hovhannisyan and R. P. Carstens A Novel Intronic cis Element, ISE/ISS-3, Regulates Rat Fibroblast Growth Factor Receptor 2 Splicing through Activation of an Upstream Exon and Repression of a Downstream Exon Containing a Noncanonical Branch Point Sequence Mol. Cell. Biol., January 1, 2005; 25(1): 250 - 263. [Abstract] [Full Text] [PDF]

				E. Buratti and F. E. Baralle Influence of RNA Secondary Structure on the Pre-mRNA Splicing Process Mol. Cell. Biol., December 15, 2004; 24(24): 10505 - 10514. [Full Text] [PDF]

				E. Buratti, A. F. Muro, M. Giombi, D. Gherbassi, A. Iaconcig, and F. E. Baralle RNA Folding Affects the Recruitment of SR Proteins by Mouse and Human Polypurinic Enhancer Elements in the Fibronectin EDA Exon Mol. Cell. Biol., February 1, 2004; 24(3): 1387 - 1400. [Abstract] [Full Text] [PDF]

				A. P. Baraniak, E. L. Lasda, E. J. Wagner, and M. A. Garcia-Blanco A Stem Structure in Fibroblast Growth Factor Receptor 2 Transcripts Mediates Cell-Type-Specific Splicing by Approximating Intronic Control Elements Mol. Cell. Biol., December 15, 2003; 23(24): 9327 - 9337. [Abstract] [Full Text] [PDF]

				F Gualandi, C Trabanelli, P Rimessi, E Calzolari, L Toffolatti, T Patarnello, G Kunz, F Muntoni, and A Ferlini Multiple exon skipping and RNA circularisation contribute to the severe phenotypic expression of exon 5 dystrophin deletion J. Med. Genet., August 1, 2003; 40(8): e100 - 100. [Full Text] [PDF]

This Article Abstract Full Text (PDF) All Versions of this Article: 277/51/50143    most recent M207409200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Muh, S. J. Articles by Carstens, R. P. Search for Related Content PubMed PubMed Citation Articles by Muh, S. J. Articles by Carstens, R. P. Social Bookmarking                      What's this?