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

J. Biol. Chem., Vol. 277, Issue 18, 15241-15251, May 3, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/18/15241    most recent M109046200v1 Alert me when this article is cited 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 Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Pollard, A. J. Articles by Europe-Finner, G. N. Search for Related Content PubMed PubMed Citation Articles by Pollard, A. J. Articles by Europe-Finner, G. N. Social Bookmarking                      What's this?

Alternative Splicing of the Adenylyl Cyclase Stimulatory G-protein G_s Is Regulated by SF2/ASF and Heterogeneous Nuclear Ribonucleoprotein A1 (hnRNPA1) and Involves the Use of an Unusual TG 3'-Splice Site^*

Alison J. Pollard, Adrian R. Krainer^§, Stephen C. Robson, and G. Nicholas Europe-Finner

From the Department of Obstetrics and Gynaecology, University of Newcastle upon Tyne, Royal Victoria Infirmary, Richardson Road, Newcastle upon Tyne, NE1 4LP, United Kingdom

Received for publication, September 19, 2001, and in revised form, January 30, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The factors involved in regulating alternative splicing of the human adenylyl cyclase stimulatory G-protein G_s in different cell types remain undefined. We have designed a G_s minigene that retains the signals required for G_s alternative splicing in vivo. Employing transient transfection of human myometrial smooth muscle cells and HeLa cells, as well as in vitro splicing assays, we have provided evidence that the antagonistic splicing factors SF2/ASF and hnRNPA1 act as potent regulators of G_s isoform expression in these cells. Both SF2/ASF and hnRNPA1 control the selection of competing 5'-splice sites and respectively promote inclusion or skipping of the small cassette-type exon 3 of G_s transcripts, resulting in the generation of G_s-long and G_s-short mRNA isoforms. We have also provided evidence that SF2/ASF and hnRNPA1 play a role in 3'-splice site selection involving the use of a non-canonical TG 3'-splice site preceding exon 4. Using a score-matrix analysis to identify putative exonic enhancer sequences (ESEs), we found multiple high score ESE motifs for SF2/ASF, SC35, and SRp40 in exon 3 of G_s. These results suggest that tissue-specific expression of SF2/ASF and hnRNPA1 governs the expression of alternative isoforms of G_s in these different cells types.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The GTP-binding protein G_s is the primary stimulatory component of adenylyl cyclase in most cell types and has also been associated with activation of dihydropyridine-sensitive voltage-gated calcium channels (1) in skeletal muscle as well as inactivation of cardiac sodium channels (2). Two ubiquitously expressed forms of G_s have been identified via ADP-ribosylation with cholera toxin or by Western blotting (3, 4). These proteins migrate in SDS-PAGE with apparent molecular masses of 52 and 45 kDa, depending on the experimental conditions, and have been designated the long and short isoforms of G_s, respectively (3-5). Both isoforms of G_s are generated by alternative splicing of a single precursor mRNA transcript.

In humans, a single copy of the G_s gene is found on chromosome 20 and is composed of 13 exons separated by 12 introns, altogether spanning a 20-kb region of genomic DNA (6). Cloning of the human G_s gene and G_s complementary DNAs showed that the short form of G_s is distinguished from the long form by the exclusion of the 45-bp exon 3, which encodes 15 amino acids. Similar studies have pointed to the existence of two additional G_s mRNA species. These isoforms may be produced by the use of an unusual alternative TG 3'-splice site, instead of the consensus AG 3'-splice site upstream of exon 4, resulting in the inclusion of an extra triplet coding for a serine residue after amino acids 87 and 72 of the long and short forms of G_s, respectively (6) (Fig. 1). It has been suggested that inclusion of this extra serine residue into G_s proteins confers additional consensus sequence sites for phosphoregulation by protein kinases C and A (7, 8). Tissue-dependent alternative splicing of the G_s precursor transcript may therefore result in the expression of two long and two short forms of G_s, depending upon the presence or absence of the extra serine residue. Initially, it was thought that these isoforms were the only splicing products of the G_s gene; however, several other variants (reviewed in Ref. 9) have been observed involving the novel exons A and/or B upstream from exon 2, which are generated either by alternative splicing or by use of an alternative promoter. Remarkably, the use of an alternative TG 3'-splice site in splicing of human G_s pre-mRNA appears to occur also in Drosophila melanogaster (10). In this organism, alternative splicing of intron 7 of the G_s gene, involving the use of either a non-consensus TG or a consensus AG 3'-splice site, results in transcripts that code for either a long or short form of G_s. These protein species differ by inclusion or exclusion of three amino acids and substitution of serine for a glycine residue.

View larger version (25K): [in this window] [in a new window]   Fig. 1.   Exon/intron organization map of exons 2-4 of the human Gs gene. Constitutively spliced exons (2 and 4) are represented by open boxes and the alternatively spliced exon 3 by a hatched box. In addition to the consensus 3'-splice site AG preceding exon 4, Gs has an non-canonical 3'-splice site TG. Use of the TG 3'-splice site incorporates an additional CAG triplet into the spliced mRNA, resulting in an extra serine residue (shaded box). The four individual protein isoforms generated from inclusion and skipping of exon 3 with or without the extra serine residue are shown.

There is evidence to suggest that the various isoforms of G_s have different regulatory functions. Sternweis et al. (11) used partially purified preparations of long and short isoforms of G_s from rabbit liver and found that the larger species has a greater ability to support hormone-stimulated adenylyl cyclase activity. Yagami (12) studied the interaction of G_s species with -adrenergic receptors in liver plasma membranes and found that the receptors were preferentially coupled to the long G_s isoform. More recently, Unson et al. (13) showed that the long and short forms of G_s interact differently with the glucagon receptor, such that specific coupling of the receptor to the long G_s isoform results in 10-fold higher binding affinity for glucagon. Furthermore, there is substantial evidence that expression of alternatively spliced isoforms of G_s differs in various tissues. For instance, the long form of G_s predominates in cerebellum, cortex, kidney, adrenal medulla, and placenta, whereas the short form is predominant in platelets, liver, neostriatum, and heart (9). Dramatic changes in steady state mRNA and protein levels of G_s isoforms have also been observed during ontogenetic development, aging, cellular differentiation, and pathophysiological states such as obesity, hypertension, diabetes, and alcoholism (9). Further evidence for variable isoform expression was obtained in the human myometrium, where short and long G_s isoforms were found to be significantly increased during gestation and subsequently down-regulated during labor, associated with a concomitant increase and decrease in adenylyl cyclase activity (5, 14).

The above observations support the premise that expression of alternative spliced isoforms of G_s is regulated in a tissue-specific manner, depending on the activity requirements of the individual tissue or cell type and, as such, is controlled at the pre-mRNA processing level by the splicing machinery. However, the regulatory trans-acting factors and cis-elements involved in modulating expression of the four potential spliced isoforms of G_s have not been elucidated.

In general, alternative precursor mRNA splicing is regulated by trans-acting splicing factors, which include the ubiquitously expressed small nuclear ribonucleoproteins, the serine/arginine-rich (SR)¹ family of nuclear phosphoproteins, and the heterogeneous nuclear ribonucleoproteins (hnRNPs) (15, 16). Variations in concentrations and ratios of specific splicing factors present in different cell types and tissues have also been shown to be a determinant in controlling pre-mRNA processing (17-19), and it has been suggested that individual splicing factors are specific for a particular pre-mRNA or the tissue type in which it is expressed. SF2/ASF, an SR protein family member, has been well characterized as a key factor involved in 5'-splice site determination in alternative splicing and has also been shown to affect 3'-splice site selection in vitro (20-22). Several studies have shown that variations in the concentration of SF2/ASF can influence which 5'-splice site is selected (21-23). hnRNPA1 has also been comprehensively studied in combination with SF2/ASF (24-26), such that subtle changes in the concentrations of these two proteins can define the formation of different spliced mRNA isoforms derived from a number of precursor mRNA species. SF2/ASF and hnRNPA1 have an antagonistic relationship and can counteract each other in a concentration-dependent manner. Increased expression of SF2/ASF generally activates proximal (downstream) 5'-splice sites, whereas increased expression of hnRNPA1 usually activates distal (upstream) 5'-splice sites (23, 24, 27). This switching of splice site usage also occurs after overexpression of these proteins in vivo; when tested with a range of alternatively spliced reporter genes, the proteins promote the expected changes in exon skipping and inclusion (23, 26, 28). In this context, changes in the tissue levels of SF2/ASF and hnRNPA1 have recently been described by Pollard et al. (29), who demonstrated that expression of these specific splicing factors is both spatially and temporally regulated in the human myometrium during fetal maturation. This finding further supports the premise that concentration ratios of SF2/ASF to hnRNPA1 in vivo may therefore be critical in defining the expression of specific protein isoforms in different tissues. Several studies have also indicated that cis-acting elements within exons can function as splicing enhancers and contribute to the accuracy and efficiency of alternative splicing when bound by specific trans-acting factors (30-32). Purine-rich splicing enhancer sequences (ESEs) have been identified in exons from a number of different tissue-specific or developmentally regulated precursor mRNA species (33, 34). These and other degenerate short RNA sequences, which are recognized by one or more SR proteins, are capable of influencing 5'- and 3'-splice site selection by affecting the activity of weak splice sites in adjacent introns (32, 35, 36).

Consequently, to determine the trans-acting factors involved in regulating expression of alternatively spliced isoforms of G_s in different cell types, we have designed a human G_s minigene construct that incorporates intronic and exonic regulatory cis-elements associated with G_s alternative splicing in vivo. Transfection of human myometrial smooth muscle cells and HeLa cells in culture with this G_s minigene and plasmids coding for SF2/ASF and hnRNPA1, in conjunction with in vitro splicing assays, strongly implicate these antagonistic splicing factors in regulating G_s expression in human tissues and cells. Overexpression of SF2/ASF and hnRNPA1 promoted inclusion and skipping of exon 3 of G_s, respectively, which not only involved the use of the consensus AG 3'-splice site preceding exon 4 but also of an unusual TG 3'-splice site immediately upstream. Moreover, motif-prediction analysis based on randomization and functional selection of sequences with enhancer activity (37) enabled us to identify putative elements recognized by other members of the SR protein family in antagonizing hnRNPA1 to control long and short G_s isoform expression.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Plasmid Construction-- The G_s minigene splicing construct (pcDNA3.1G_s) was generated from human genomic DNA using specific G_s PCR primers, containing specific restriction sites, for the individual exon/intron fragments of G_s to facilitate the sequential insertion of each fragment into the pcDNA3.1 expression vector (Invitrogen), as shown in Fig. 2A. These primers amplify G_s exons 2-4, together with truncated versions of their adjacent intronic sequences. The DNA sequence for the human G_s gene has been described previously (6). Briefly, the first fragment, consisting of the 73-bp exon 2 and the first 87 bp (of 3258 bp) of intron 2, was amplified by PCR using Pfu polymerase with primers A and B (see Fig. 2B for full details of all primers used). The second fragment, consisting of the last 132 bp of intron 2, the 45-bp exon 3, and the first 85 bp (of 4547 bp) of intron 3, was then generated using primers C and D. The third fragment, consisting of the last 76 bp of intron 3 and the 55-bp exon 4, was amplified using primers E and F. PCR products were individually cloned into the TOPO®-TA cloning vector (Invitrogen) and sequenced to confirm exon/intron DNA sequences and ensure that the necessary cis-acting regulatory signals were present. Each G_s fragment was then subcloned sequentially into pcDNA3.1 by repeated restriction digestion and ligation using T4 DNA ligase (Promega) as shown in Fig. 2A to generate the complete pcDNA3.1G_s minigene splicing construct, which is under transcriptional control of the cytomegalovirus promoter (Fig. 2A, P1) and contains the bGH polyadenylation signal. Plasmids encoding splicing factors SF2/ASF (pCG-SF2) and hnRNPA1 (pCG-A1) have been described (23).

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Schematic diagram showing subcloning procedure to generate Gs minigene splicing construct consisting of exons 2, 3, and 4, together with their flanking intronic sequences. A, individual exon/intron fragments were amplified from human genomic DNA using Gs-specific primers that contain specific restriction sites as shown. The plasmid pcDNA3.1 was initially linearized with NheI and XhoI, and the first Gs fragment of exon 2 and 87 bp of intron 2 (with NheI and XhoI restriction sites at the 5'- and 3'-ends, respectively) inserted. The subcloning procedure was then repeated using XhoI and EcoRV to facilitate insertion of the second Gs fragment (132 bp of intron 2, exon 3, and 85 bp of intron 3), and the third fragment (76 bp of intron 3 and exon 4) was finally subcloned using EcoRV and Pme I. B, DNA sequences for the Gs-specific primers (A-F) used throughout this study in RT-PCR and to amplify individual Gs exon/intron fragments by PCR.

Tissues and Cell Culture-- Both primary myometrial smooth muscle cell cultures and an immortalized myometrial cell line (PHM1-41) were used in this study. The PHM1-41 smooth muscle cells (provided by Professor B. M. Sanborn, University of Texas) were originally isolated from the myometrium of a pregnant woman and immortalized using a replication-defective adenovirus vector expressing E6 and E7 proteins from human papillomavirus HPV16 (38). Samples of myometrial tissue were obtained from pregnant women undergoing elective caesarean section at term. Written consent was obtained from all women, and ethical approval was granted by the Newcastle and North Tyneside Health Authority Ethics Committee. Myometrial samples were taken from the upper corpus region (termed upper segment) and from close to the cervix (termed lower uterine segment). Primary myocyte isolation and culture were essentially as described by Phaneuf et al. (39). Primary myocytes were cultured with complete D-valine medium (Invitrogen), whereas the immortalized pregnant human myometrial cells (PHM1-41) and HeLa cells were cultured in Dulbecco's Glutamax II modified Eagle's medium (Sigma), supplemented with 10% fetal calf serum, penicillin (1 unit/ml), and streptomycin (1 ng/ml), and cultured in T75 flasks under standard conditions at 37 °C with 5% CO2. The culture medium for PHM1-41 cells contained additional supplements of 4.5 g/ml glucose and 50 µg/ml G418 (Invitrogen).

Transfection-- All transfection experiments on primary myometrial cells were undertaken on subculture passages 2-3. Myometrial and HeLa cells were co-transfected, at 60-70% confluency (in the absence of antibiotics) using Mirus LT-1 (Cambridge Bioscience) or TransFast^TM (Promega) cationic-lipid transfection reagents with 3 µg of pcDNA3.1G_s and 3 µg of pCG-SF2, pCG-A1, or pCG control vector. Transfection efficiencies were in the range of 25-30% for all experiments, as determined by transfection with a -galactosidase-encoding plasmid, pcDNA3.1LacZ (Invitrogen) (data not shown). Cells were harvested either 48 h (HeLa cells) or 72 h (myometrial cells) after transfection, using trypsin (0.5 g/liter)-EDTA (0.2 g/liter). Confirmation that the pcDNA3.1G_s minigene plasmid contained the necessary cis-acting regulatory signals for efficient pre-mRNA splicing was obtained by runoff transcription using a mMESSAGE-mMACHINE capping/transcription kit (described below) and RT-PCR of the G_s spliced products from total RNA extracted from PHM1-41 cells post-transfection.

Western Blotting-- Transfection efficiencies were further confirmed by Western immunoblotting using monoclonal antibodies to SF2/ASF (mAb96) and hnRNPA1 (4B10), both of which have been described (18, 40). Endogenous levels of SF2/ASF and hnRNPA1 in myometrial cells and HeLa cells were also calculated by this procedure. Briefly, SDS-PAGE was performed under standard conditions using equal amounts of cell homogenate solubilized in 0.5 M Tris, 5 M urea, 2.5% SDS, and 3% -mercaptoethanol and resolved on 12% polyacrylamide gels. Immunoreactive bands were detected by enhanced chemiluminescence (Amersham Biosciences).

G_s mRNA Splicing Analysis by RT-PCR and Restriction Digestion-- G_s mRNA splice variants generated from the minigene in transfected cells were analyzed by RT-PCR. Total RNA was isolated from individual transfection experiments, fresh pregnant myometrial tissue, or cultured cells, using SV total RNA isolation kits, as recommended by the manufacturer (Promega) and first strand cDNA synthesized from 1-3 µg of RNA using 20 units of Superscript II reverse transcriptase (Invitrogen) with 100 ng of oligo(dT)₁₆ as primer. PCR amplification was carried out with 2-4 µl of cDNA template with the G_s sense primer for exon 2 (primer A) and reverse primer for exon 4 (primer F) that have 8 and 14 terminal nucleotides, respectively, derived from the pcDNA3.1G_s minigene construct and amplifying all four mRNA isoforms generated from pcDNA3.1G_s, although not endogenous G_s mRNA transcripts (see Fig. 3A, lanes 2 and 3). PCR was performed under standard conditions with an initial hot start cycle at 94 °C (4 min), 50 °C (30 s), and 72 °C (1 min), followed by 18-24 cycles at 94 °C (1 min), 50 °C (30 s), and 72 °C (1 min). Reactions were terminated in the exponential phase in order to determine the relative ratios of individual G_s mRNA species. PCR products were analyzed by 2% agarose gel electrophoresis. Note that the long G_s isoform results in a PCR product of 195 bp, which includes the 8 extra nucleotides on primer A and 14 extra nucleotides on primer F plus 173 bp of exons 2, 3, and 4, whereas the short G_s isoform results in a product of 150 bp because of exclusion of the 45 bp of exon 3. Long or short G_s isoforms that included or lacked the extra CAG triplet at the 3'-end of intron 3 could not be resolved under these electrophoretic conditions. Therefore, to further analyze the relative abundance of each spliced variant of G_s, the sense G_s primer A was end-labeled with [-³²P]ATP by polynucleotide kinase (Promega) prior to PCR amplification. The use of the alternative TG 3'-splice site upstream of exon 4 (Fig. 1) generates the G_s-short variant with three additional CAG nucleotides, creating a unique BsmA1 endonuclease restriction site (see Fig. 5A) as described previously (41). Thus, the two G_s-short mRNA variants with and without the CAG triplet can be distinguished from each other by restriction analysis. A 10-µl aliquot of each radiolabeled PCR product was digested with BsmA1 at 55 °C, analyzed by PAGE on 6 or 8% acrylamide, 7 M urea gels, and visualized by autoradiography. The different banding patterns were quantified by scanning densitometry using a UMAX PS2400 scanner at 700 dots per inch coupled to the Intelligent Quantifier software from BioImage (Ann Arbor, MI).

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Controls for transient transfection experiments using the minigene splicing construct pcDNA3.1Gs and plasmids encoding SF2/ASF and hnRNPA1. A, lane 1, precursor RNA synthesized via runoff transcription from pcDNA3.1. Gs; lane 2, RT-PCR using Gs A and F primers (see Fig. 2B) and total RNA extracted from untransfected PHM1-41 cells; lane 3, RT-PCR using Gs A and F primers and total RNA extracted from PHM1-41cells transfected with the pcDNA3.1Gs minigene; note that the two PCR products represent the Gs spliced isoforms with and without exon 3. B, Western blotting using the specific monoclonal antibody to SF2/ASF (mAb96), demonstrating overexpression of SF2/ASF after transfection in HeLa cells with pCG-SF2. Lane 1, transfection with pCG control vector; lane 2, transfection with pCG-SF2. C, Western blotting using the specific monoclonal antibody to hnRNPA1 (4B10), demonstrating overexpression of hnRNPA1 after transfection in HeLa cells with pCG-A1. Lane 1, transfection with pCG control vector; lane 2, transfection with pCG-A1.

Endogenous G_s mRNA variants were detected using G_s-specific oligonucleotide primers CAACGAGGAGAAGGCGCAGCGTGA and CTCTGAGGTTCTCGGGGTTGG. GAPDH-specific oligonucleotide exon primers (spanning a short 104-bp intron) were also designed as control primers for use in RT-PCR with each cDNA sample. DNA sequences for the GAPDH primers were CTGCCGTCTAGAAAAACC and CCACC- TTCGTTGTCATACC.

In Vitro Splicing Assays-- To generate transcripts for in vitro splicing, the pcDNA3.1G_s minigene plasmid, which also harbors a T7 promoter (Fig. 2A, P2) was linearized with PmeI. GpppG-capped and [-³²P]UTP-labeled pre-mRNA substrate (Fig. 3A, lane 1) was synthesized by runoff transcription using a mMESSAGE-mMACHINE capping/transcription kit with T7 RNA polymerase (Ambion, Inc.). In vitro splicing assays were undertaken essentially as described (42). Briefly, 25-30 fmol of radiolabeled precursor mRNA was incubated for 4 h at 30 °C with 10 µl of HeLa nuclear extract supplemented with 15-25 pmol of SF2/ASF, SC35, or hnRNPA1 recombinant proteins, which were expressed and purified as described (27, 37). The MgCl2 concentration was 3.5 mM. The RNA products generated from splicing in vitro were then fractionated by denaturing polyacrylamide electrophoresis followed by autoradiography.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Effect of SF2/ASF and hnRNPA1 on G_s Isoform Expression-- Transient co-transfection experiments were performed on primary myometrial cell cultures, immortalized myometrial cells (PHM1-41), and HeLa cells to evaluate the effect of increased levels of SF2/ASF and hnRNPA1 on the expression of spliced variants of G_s in these different cell types. The nuclear abundance of SF2/ASF and hnRNPA1 in HeLa cells has been calculated previously at 3 × 10⁷ copies/cell for SF2/ASF and 6-7 × 10⁷ copies/cell for hnRNPA1 (18, 45).

Western immunoblotting using equal amounts of protein lysate (200/400 µg) from myometrial/PHM1-41 smooth muscle cells and HeLa cell cultures was undertaken to determine the relative abundance of SF2/ASF and hnRNPA1 mRNA in myometrial cells compared with HeLa cells (Fig. 4, A and B). SF2/ASF levels in both primary myometrial cell cultures and PHM1-41 smooth muscle cells were calculated to be 42% of that found in HeLa cells with a calculated estimate of 1.3 × 10⁷ copies/cell, whereas the abundance of hnRNPA1 in myometrial cells/PHM1-41 smooth muscle cells was 59% of that found in HeLa cells with a calculated estimate of 3.8 × 10⁷ copies/cell. The ratio of SF2/ASF to hnRNPA1 in both myometrial cells and HeLa cells is thus ~1:3. However, employing RT-PCR with the G_s-specific primers to monitor endogenous splicing in these cell types (Fig. 4C) indicates that in the two myometrial cell types the long form of G_s is predominant, whereas the short form is predominant in HeLa cells. The latter is as predicted from the SF2/ASF:hnRNPA1 ratio of 1:3 in these cells. The discrepancy in myometrial cells is discussed.

View larger version (30K): [in this window] [in a new window]   Fig. 4.   Abundance of SF2/ASF and hnRNPA1 and expression of Gs isoforms in human myometrial and HeLa cells. Western immunodetection and subsequent densitometric analysis of SF2/ASF protein expression (A) and hnRNPA1 protein expression (B). Lane 1, primary myometrial cells cultures; lane 2, immortalized myometrial cells (PHM1-41); lane 3, HeLa cells. SF2/ASF levels in myometrial cells were calculated to be 42% of that found in HeLa cells with a calculated estimate of 1.3 × 107 copies/cell. hnRNPA1 levels in myometrial cells were calculated to be 59% of that found in HeLa cells with a calculated estimate of 3.8 × 107 copies/cell. C, identification of endogenous Gs splice variants by RT-PCR. Lane 1, primary myometrial cell cultures; lane 2, PHM1-41 cells; lane 3, HeLa cells. The two PCR products of 344 and 299 bp represent the long and short Gs spliced variants.

Alternative splicing of human G_s mRNA transcripts involves the inclusion or exclusion of a small internal cassette-type exon of 45 bp (exon 3) as well as the inclusion/exclusion of an extra triplet. The G_s-spliced mRNA variants generated from the pcDNA3.1G_s plasmid, co-transfected with pCG-SF2 or pCG-A1, were analyzed by RT-PCR using a sense primer for exon 2 (primer A) and an antisense primer for exon 4 (primer F), resulting in amplification of all four exogenous G_s mRNA isoforms. The data presented are based on each transfection experiment being performed in triplicate. Representative RT-PCR splicing analyses of transfection experiments in primary myometrial, PHM1-41, and HeLa cells are shown in Fig. 5. The two PCR products (195 and 150 bp) represent the long and short G_s mRNA isoforms (plus or minus the extra 3 bp) generated by the presence or absence of exon 3. Note that G_s-spliced variants containing or excluding the three nucleotides (CAG) at the 5'-end of exon 4 cannot be resolved under these experimental conditions.

View larger version (43K): [in this window] [in a new window]   Fig. 5.   Effect of SF2/ASF and hnRNPA1 on Gs isoform expression. RT-PCR with Gs A and F primers from transient co-transfection of pcDNA3.1Gs, together with plasmids encoding SF2/ASF (pCGSF2) and hnRNPA1 (pCGA1) in human primary myometrial cell cultures (A), immortalized myometrial cells (PHM1-41) (B), and HeLa cells (C). Lane 1, cells transfected with pCGSF2; lane 2, cells transfected with pCGA1; lane 3, cells transfected with the pCG control vector; lane 4, no template control. The quantitations are based on each transfection experiment being performed in triplicate. GAPDH controls for each cell type are shown below.

Co-transfection of pcDNA3.1G_s and pCG-SF2 plasmids in primary myometrial and PHM1-41 cells resulted in increased expression of the long G_s mRNA isoform, which includes exon 3, as reflected by the intensity of the 195-bp band compared with the 150-bp band (Fig. 5, A and B, lane 1). In contrast, co-transfection with pCG-A1 in both of these cell types resulted in a significant increase in the expression of the smaller 150-bp G_s isoform, which has exon 3 spliced out (Fig. 5, A and B, lane 2). The observed switch in splicing of G_s pre-mRNA upon overexpression of SF2/ASF or hnRNPA1, resulting in the inclusion or exclusion of exon 3, is consistent with the hypothesis that the ratio of SF2/ASF and hnRNPA1 is important in determining splice site selection of G_s pre-mRNA in these cells. Note that in myometrial cells the basal splicing pattern for pcDNA3.1G_s in the presence of the control pCG plasmid (Fig. 5, A and B, lane 3) is similar to endogenous splicing of G_s (Fig. 4C), in that the predominant isoform expressed in these cells is G_s-long, as indicated by an increase in the intensity of the large 195-bp band compared with the small 150-bp band.

In HeLa cells, the basal splicing pattern for pcDNA3.1G_s was different from that found in myometrial cells. RT-PCR analysis of HeLa cells transfected with pcDNA3.1G_s and the control pCG plasmid (Fig. 5C, lane 3) indicated that the principal isoform expressed in these cells is G_s-short, as indicated by a high intensity band of 150 bp and a low intensity band of 195 bp. This is in agreement with endogenous splicing of G_s in these cells (Fig. 4C). However, a switch in the splicing pattern of the G_s minigene as a consequence of overexpression of SF2/ASF or hnRNPA1 was also observed in HeLa cells, consistent with the results obtained in myometrial cells. Overexpression of SF2/ASF resulted in a slight increase in expression of G_s-long, which includes exon 3 (Fig. 5C, lane 1). In contrast, overexpression of hnRNPA1 promoted skipping of exon 3, increasing the expression of the G_s-short isoform (Fig. 5C, lane 2).

Effect of SF2/ASF and hnRNPA1 on Use of a Non-canonical TG 3'-Splice Site-- Inclusion or skipping of exon 3 of G_s may not only involve the use of the consensus AG 3'-splice site at the 5'-end of exon 4 but also of a non-canonical TG 3'-splice site located three nucleotides upstream (Fig. 1). The use of this TG 3'-splice site generates G_s variants (with and without exon 3) that have an additional triplet (CAG) that results in an extra serine residue at the 5'-end of exon 4. This CAG base insertion creates a unique BsmA1 restriction site in G_s-short, which was utilized to distinguish G_s isoforms generated from the use of the atypical TG 3'-splice site from those generated via the consensus AG 3'-splice site (Fig. 6A).

View larger version (41K): [in this window] [in a new window]   Fig. 6.   Restriction analysis of the short Gs isoform generated from pcDNA3.1Gs, showing the use of the non-canonical TG 3'-splice site by SF2/ASF and hnRNPA1. A, the short Gs isoform resulting from the use of the TG 3'-splice site preceding exon 4 has a unique BsmA1 restriction site. B, RT-PCR of extracted RNA from transfected primary and PHM1-41 myometrial cells was carried out using Gs primers A (end-labeled with [-32P]ATP) and F, with subsequent BsmA1 restriction analysis of the small Gs product. Lane 1, primary cells transfected with the pCG control vector; lane 2, primary cells transfected with pCGSF2; lane 3, primary cells transfected with pCGA1; lane 4, PHM1-41 cells transfected with the pCG control vector; lane 5, PHM1-41 cells transfected with pCGSF2; lane 6, PHM1-41 cells transfected with pCGA1. C, GAPDH control for each of the lanes described in B. D, BsmA1 restriction analysis of the small Gs product in HeLa cells. Lane 1, cells transfected with pCGSF2; lane 2, cells transfected with pCGA1; lane 3, cells transfected with the pCG control vector. E, GAPDH control for each lane described in (D).

To ascertain whether SF2/ASF and hnRNPA1 influence 3'-splice site selection in G_s pre-mRNA, the PCR products generated in the above transfection experiments were digested with BsmA1. Representative analyses are shown in Fig. 6. In Fig. 6B, lanes 1-6, the larger uncleaved PCR bands of 150 bp represent the G_s-short mRNA variant generated from the use of the consensus AG 3'-splice site, whereas the smaller cleaved PCR bands of 76 bp generated after digestion with BsmA1 represent the G_s-short mRNA variant derived from use of the alternative TG 3'-splice site.

The predominant exogenous G_s-short isoform expressed in both primary myocytes and PHM1-41 cells under all transfection regimes resulted from the use of the alternative TG 3'-acceptor site, as indicated by the intensity of the BsmA1-cleaved band of 76 bp. Co-transfection with hnRNPA1 resulted in a decrease in the level of the 76-bp band and a corresponding increase in the level of the 150-bp band (Fig. 6B, lanes 3 and 6, asterisk) indicating that overexpression of hnRNPA1 switched splice site selection in myometrial cells from the TG 3'-splice site to the consensus AG 3'-splice. In contrast, overexpression of SF2/ASF in these cells resulted in a slight increase in the expression of the G_s spliced isoform derived from the use of the TG 3'-splice site (Fig. 6B, lanes 2 and 5). The observation that SF2/ASF promoted the use of the alternative TG 3'-splice site was confirmed upon overexpression of this regulatory factor in HeLa cells (Fig. 6D, lane 1). Note that the predominant band after co-transfection with pCG-SF2 was that of the BsmA1-digested PCR product of 76 bp, compared with the uncleaved PCR product of 150 bp (Fig. 6D, lane 1). Co-transfection with pCG-A1 in HeLa cells resulted in a shift in 3'-splice site usage consistent with that observed in myometrial cells, in that overexpression of hnRNPA1 reduced the use of the TG 3'-splice site and favored selection of the consensus AG 3'-splice site (Fig. 6D, lane 2, asterisk).

Effect of SF2/ASF, hnRNPA1, and SC35 on G_s Isoform Expression by in Vitro Splicing Assays-- The role of SF2/ASF and hnRNPA1 in regulating alternative splicing of G_s pre-mRNA was further examined by in vitro splicing. Transcripts from the pcDNA3.1G_s minigene construct were spliced in HeLa nuclear extract supplemented with recombinant SF2/ASF and hnRNPA1 proteins (Fig. 7A). The G_s pre-mRNA was spliced under these experimental conditions, generating mRNA products in which introns 2 and 3 were removed, with and without skipping of exon 3. Addition of hnRNPA1 induced skipping of exon 3, generating the short G_s variants consisting of exons 2 and 4, (Fig. 7A, lane 2), whereas addition of SF2/ASF promoted inclusion of exon 3, thus generating the long G_s isoform containing exons 2, 3, and 4 (Fig. 7A, lane 1). The spliced mRNA product representing the long G_s isoform was also the principal band generated when the SR protein family member SC35 was added to the splicing reactions (Fig. 7A, lane 3). SC35 has been shown previously to have similar splicing activities to SF2/ASF in modulating exon inclusion in vitro (43). Note that the use of the TG 3'-splice site cannot be determined under these experimental conditions.

View larger version (20K): [in this window] [in a new window]   Fig. 7.   A, in vitro splicing of pcDNA3.1Gs pre-mRNA in HeLa cell nuclear extracts supplemented with recombinant SF2/ASF, SC35, or hnRNPA1. Precursor RNA was synthesized by runoff transcription with subsequent in vitro splicing assays as described under "Experimental Procedures." Lane 1, nuclear extract supplemented with SF2/ASF; lane 2, nuclear extract supplemented with hnRNPA1; lane 3, nuclear extract supplemented with SC35; lane 4, unsupplemented nuclear extract control. B, high score SR protein motifs in alternative exon 3 of Gs. The 45-nucleotide sequence was searched with four nucleotide-frequency matrices derived from pools of functional enhancer sequences selected in vitro (22). Motif scores reflect the extent of matching to a degenerate consensus, and only the scores above the threshold for each SR protein are shown. High score motifs are shown in black for SF2/ASF, dark gray for SC35, and light gray for SRp40. The width of each bar represents the length of the motif (7 or 8 nucleotides), the placement of each bar indicates the position of the motif along the exon 3 sequence (shown below the x axis), and the height of each bar corresponds to the numerical score on the y axis. No high score motifs for SRp55 were found in this exon.

Identification of Putative Exonic Splicing Enhancer Sequences in Exon 3 of G_s-- Cis-acting nucleotide sequences other than the consensus splice sites at intron:exon junctions can influence pre-mRNA splicing by activating splice sites in nearby introns (28, 30, 31). These ESEs consist of short RNA sequences that fit one of several degenerate consensus motifs and have been identified in both alternatively and constitutively spliced exons (27, 32, 34, 35). Specific ESEs interact functionally with individual trans-acting factors, which include SF2/ASF and SC35, as well as other members of the SR protein family (27, 32, 36, 44). To investigate whether G_s contains any potential ESE sequences that may be involved in regulating the selection of alternative 5'- and 3'-splice sites, the alternative exon 3 sequence of G_s was analyzed using statistical scoring matrices for four SR proteins, previously derived on the basis of functional selection from random sequence pools (35, 37). Briefly, this procedure identifies putative SR protein-specific enhancer sequences by a motif-search algorithm that calculates numerical scores greater than a previously established threshold. The results from this analysis demonstrate that ESE motifs for three SR proteins, SF2/ASF, SC35, and SRp40, are present within the G_s exon 3 sequence (Fig. 7B). Two heptameric high score motifs (AAGAGGA and CGCAGGC) for SF2/ASF were identified, which overlap with two octameric SC35 motifs (GACCCGCA and GGCTGCAA). Furthermore, two ESE motifs for SRp40 were also found (CCGCAGG and CAACAGC). No high score motifs for SRp55 were found. The significance of these observations is discussed below.

Spatio-temporal Expression of SF2/ASF, hnRNPA1, and G_s Isoforms within the Myometrium during Pregnancy-- Both SF2/ASF and hnRNPA1 were found to be spatially and temporally expressed within the myometrium during gestation. Notably, SF2/ASF levels increased significantly in the lower uterine region (Fig. 8A), and this was associated with a parallel decrease in levels of hnRNPA1 (Fig. 8B). Conversely, the reverse pattern was observed for the upper uterine region where hnRNPA1 was significantly up-regulated (Fig. 8B), which was associated with a concurrent decrease in SF2/ASF levels (Fig. 8A). RT-PCR using RNA extracted from the upper and lower regions of the myometrium was undertaken to ascertain whether endogenous G_s splice variants were also spatially expressed within these different regions during fetal maturation. The splicing pattern of G_s splice variants is shown in Fig. 8C; the two PCR products (344 and 299 bp) represent the long and short mRNA variants (plus and minus the extra 3 bp). The data indicate an increase in expression of G_s-long in the lower uterine region, as reflected by the intensity of the 344-bp band compared with the 299-bp band (Fig. 8C), whereas an increase in expression of G_s-short is observed within the upper uterine region, as reflected by the increase in the intensity of the 299-bp band compared with the 344-bp band (Fig. 8C). The significance of the differential expression of SF2/ASF, hnRNPA1, and G_s spliced variants is discussed.

View larger version (18K): [in this window] [in a new window]   Fig. 8.   Spatial expression of myometrial SF2/ASF, hnRNPA1, and Gs isoforms within the upper and lower uterine regions during human gestation. This is described by Pollard et al. (29). Western immunodetection and subsequent densitometric analysis of SF2/ASF protein expression in the upper (U) and lower (L) myometrium (A). B, hnRNPA1 protein expression in the upper and lower myometrium. C, identification of endogenous Gs splice variants by RT-PCR using total RNA extracted from the upper and lower myometrium during pregnancy. The two PCR products of 344 and 299 bp represent the long and short Gs spliced variants. Data indicate an increase in expression of Gs-long in the lower uterine region, whereas an increase in expression of Gs-short is observed within the upper uterine region.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The mechanism by which alternative splicing of the adenylyl cyclase stimulatory GTP-binding protein G_s is controlled has remained undefined. We now provide evidence that G_s isoform expression in human myometrial and HeLa cells is regulated by the trans-acting splicing factors SF2/ASF and hnRNPA1. Our data from transfection of a G_s minigene in human myometrial and HeLa cells are consistent with previous findings with other splicing reporters, in that high levels of hnRNPA1 favored the distal alternative 5'-splice site, such that exon 2 of G_s was spliced to exon 4. In contrast, increased levels of SF2/ASF selected the proximal alternative 5'-splice site, thus promoting mRNA transcripts in which exon 3 was spliced to exon 4.

Our studies also indicate that the basal splicing pattern of G_s transcripts appears to be different in myometrial cells compared with HeLa cells. In both primary/PHM1-41 myometrial cells the principal G_s isoform expressed was the long spliced variant that included exon 3, whereas in HeLa cells the short form of G_s resulting from the skipping of exon 3 was predominant. This pattern was also observed for endogenous splicing of G_s in these cells. The predominant expression of G_s-short in HeLa cells under endogenous conditions conforms with the abundance of hnRNPA1 6-7 × 10⁷ copies/cell (18, 45) in these cells, which promotes exon skipping by favoring the use of the distal 5'-splice site. However, in myometrial cells, although hnRNPA1 is more abundant than SF2/ASF, G_s-long is the principal isoform expressed. This may indicate that in these smooth muscle cell cultures endogenous levels of the SR proteins SC35 and SRp40 may play a role in regulating alternative splicing of G_s, because we have observed two high score motifs within exon 3 for these factors (see below). The expression profile of these factors remains to be determined in the cells. However, it must be recognized that these cells are used as an in vitro model and that the splicing of G_s in vivo in the upper and lower myometrial segments during pregnancy does conform to the ratio of SF2/ASF to hnRNPA1 found in these regions as detailed later.

The strengths of different splice sites relate to their binding affinity for trans-acting factors, including the extent of their sequence complementarity to appropriate units of snRNAs, and there is a correlation between splice site strength and the match to the splice site consensus sequence (46). The 5'- and 3'-splice sites of G_s introns 2 and 3 conform to the GT-AG dinucleotide consensus. However, we report that an alternative, non-canonical TG 3'-splice site just upstream of the 5'-end of exon 4, as described by Kozasa et al. (6), is also selected, to an extent that appears to be governed by the expression levels of SF2/ASF and hnRNPA1. A database of canonical and non-canonical mammalian splice sites has recently been compiled, based on the analysis of 22,489 verified splice site junctions (47). Of relevance to our study, 98.71% of these splice sites have the consensus GT-AG splice site pairs, 0.56% have non-canonical GC-AG junctions, and of the remainder (0.73%) no GT-TG splice site pairs (as found in the human G_s gene sequence) were found.

Usage of the alternative TG 3'-splice site generates G_s mRNA transcripts (with and without exon 3) that have an additional CAG triplet coding for an extra serine residue after amino acids 87 and 72 of the long and short isoforms of G_s, respectively. The predominant short G_s isoform expressed in myometrial and HeLa cells resulted from the use of the TG 3'-splice site and not the consensus AG 3'-splice site. However, both SF2/ASF and hnRNPA1 could influence the use of this atypical TG 3'-splice site, whereby an increase in the level of hnRNPA1 switched splice site selection from the TG 3'-splice site to that of the consensus AG 3'-splice site. Conversely, and in keeping with its antagonistic relationship with hnRNPA1, SF2/ASF promoted the use of the alternative TG 3'-splice site. These data further demonstrate that the short variant of G_s is primarily generated from the use of the TG 3'-splice site and incorporates an additional serine residue. Similarly, since SF2/ASF promoted the use of the non-canonical TG 3'-splice site and the inclusion of exon 3 into the long isoform of G_s, it is reasonable to predict that the long G_s variant is also primarily generated from the use of the alternate TG 3'-splice site and not the consensus AG 3'-splice site. Consequently, expression of the long and short G_s isoforms in different tissues and cell types may occur from the predominant selection of the atypical TG 3'-splice site, resulting in protein species containing an extra serine residue.

The use of the alternative TG 3'-splice site in producing G_s spliced isoforms may be conserved in other mammalian species, such as the rat, in which the homology with the human G_s gene is high and in which proteins of 52 and 45 kDa have been observed (41). This peculiar feature may also be present in lower organisms, including D. melanogaster, in which a TG 3'-splice site or a consensus AG 3'-splice site results in transcripts that code for either a long or short form of G_s (10). Given the unusual nature of the non-consensus TG 3'-splice site, it is remarkable that the Drosophila G_s gene also contains such a splice site, yet it corresponds to intron 7, which interrupts a different region of the coding sequence (10). In this case, there is also an alternative AG 3'-splice site, but it is located 9 bp downstream. Neither unusual intron in human or Drosophila G_s has a direct counterpart in the orthologous gene.

Two other examples of a TG 3'-splice site have been reported in the human DRD2 and DLG4 genes, which encode the dopamine receptor D2 and the postsynaptic density 95 protein (48, 49). In the DRD2 pre-mRNA, the UG 3'-splice site is located six nucleotides upstream of the conventional AG 3'-splice site of intron 5, and its use generates the D2_Longer isoform, which is expressed at low levels in human brain (48). In the DLG4 pre-mRNA, the UG dinucleotide was originally reported as the sole 3'-splice site for intron 5, and the splice junction was confirmed by cDNA sequencing (49). However, the Ensemble database (www.ensembl.org) reports an AG nine nucleotides downstream as the 3'-splice site for this intron, as supported by EST sequences (gene ID ENSG00000132535). Thus, in all the known cases of TG 3'-splice site dinucleotides, an alternative AG 3'-splice site dinucleotide is located 1, 2, or 3 codons downstream.

ESEs are sequences that have been shown to influence the selection of both 5'- and 3'-splice sites (50, 51). Some of these cis-acting sequences are purine-rich, although non-purine enhancer elements have also been identified (30, 32, 35, 37). ESEs regulate alternative splicing by acting in combination with other cis-acting signals at intron/exon junctions. The specific recognition and binding of ESEs by protein factors, including members of the SR protein family, is well documented, and characterization of their mechanism of action is an active area of research. In this context, we investigated whether the alternative exon 3 of G_s has any such ESE motifs that may be relevant to regulation of G_s post-transcriptional gene expression. Exon 3 is rich in purines, particularly G residues, and contains only two T residues (Fig. 7B). ESE motifs for several SR proteins were identified in exon 3 using scoring matrices derived from a functional SELEX procedure that identifies motifs with the potential to interact functionally with specific SR proteins to promote splicing (35, 37). Notably, two high score ESE motifs for SF2/ASF are present within the exon 3 sequence, which overlap with two high score SC35 motifs. In addition, two high score motifs for another SR protein member, SRp40, are present in exon 3, but no high score SRp55 motifs are present in this exon. SRp40 has been tested for its role in alternative splicing and appears to function in a substrate-specific manner in selecting either distal or proximal 5'-splice sites depending on the pre-mRNA substrate (17). SC35 has also been shown to have similar splicing activities as SF2/ASF in favoring proximal sites in 5'- and 3'-splice site selection and in antagonizing the effect of hnRNPA1 in alternative 5'-splice site selection (43). The specificity of SR proteins in regulating the efficiency or pattern of alternative splicing of different genes has been attributed in part to the specific recognition of different ESEs. Consequently, it is of interest that the G_s exon 3 sequence contains three different types of SR protein-specific ESE motifs, which raises the question of whether different sets of SR proteins are involved in regulating G_s isoform expression in vivo. Interestingly, it was recently reported that previously uncharacterized SR-like proteins also antagonize the roles of SF2/ASF and SC35 in regulating alternative pre-mRNA splicing in a activity that is similar to that of hnRNPA1 (53).

Variations in the concentration ratios of SF2/ASF and hnRNPA1 in different cell types and tissues may be important in controlling processing of G_s pre-mRNA. In this context, previous studies have demonstrated that SF2/ASF and hnRNPA1 levels vary considerably in different tissues from rat and mouse (18, 55). Moreover, it has also been reported that hnRNPA1 expression in the mouse fluctuates considerably during different stages of differentiation (55). Recently, Pollard et al. (29) have described the differential expression of SF2/ASF and hnRNPA1 in the myometrium of the human uterus during pregnancy and parturition. Essentially the data, as summarized in Fig. 8, indicate that SF2/ASF and hnRNPA1 are spatio-temporally regulated within the lower and upper uterine regions during fetal maturation (29). In this context, we have also demonstrated that G_s splice variants are differentially expressed within these regions and that the pattern of splicing for G_s is consistent with the known roles for SF2/ASF and hnRNPA1.

In conclusion, our studies provide strong evidence that alternative splicing of human G_s gene transcripts is regulated in part by the antagonistic relationship of the trans-acting splicing factors SF2/ASF and hnRNPA1 and involves the use of a non-canonical, alternative TG 3'-splice site. The differential expression of SF2/ASF and hnRNPA1 in various tissues/cell types may thus under normal physiological and pathophysiological conditions define the degree of formation of the long and short protein isoforms of G_s, with their subsequent effects on cell function.

	ACKNOWLEDGEMENTS

We thank Professor B. M. Sanborn (University of Texas) for kindly providing the PHM1-41 cells. Anti-hnRNPA1/A1B 4B10 monoclonal antibody was a gift from Dr. Gideon Dreyfuss (University of Pennsylvania, Philadelphia).

	FOOTNOTES

^* This study was funded by a grant from Action Research (S/P/3232).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.

To whom correspondence should be addressed. Tel.: 44-191-282-4107; Fax: 44-191-222-5066; E-mail: A.J.Pollard@ncl.ac.uk.s.

^§ Funded in part by NCI, National Institutes of Health Grant CA13106. Present address: Cold Spring Harbor Laboratory, Box 100, 1 Bungtown Rd., Cold Spring Harbor, NY 11724.

Published, JBC Papers in Press, February 1, 2002, DOI 10.1074/jbc.M109046200

	ABBREVIATIONS

The abbreviations used are: SR, serine/arginine-rich; ESE, exonic splicing enhancer sequence; GADPH, glyceraldehyde-3-phosphate dehydrogenase; hnRNP, heterogeneous nuclear ribonucleoprotein..

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES