The two classes of sequences for recognition and splicing of pre-mRNA in eukaryotes, GT-AG and AT-AC, are characterized by the nearly invariant dinucleotides present at the extreme 5' (donor) and 3' (acceptor) ends of the intron. Amongst GT-AG introns, which comprise the vast majority, the more extended consensus sequence at the 5' splice site is ACAG/GTAGAGT (where / indicates the exon-intron boundary). This sequence is complementary to part of the U1 snRNA and is important in intron recognition. We have determined the genomic structure of the mouse fibroblast growth factor receptor 2 gene (Fgfr2) and identified a divergent 5' splice site (ACA/GAAAGT), conserved in FGFR1, -2 and -3 from humans, mice and Xenopus that is used for alternative splicing of a hexanucleotide sequence, encoding Val-Thr, at the end of exon 10. This is the only example known of the use of /GA in vertebrate splicing. Similarities to a splice site in the Antennapedia gene of Drosophila suggest that this variant motif is involved in alternative splicing of short sequences at the 5' splice site. Inclusion or exclusion of the Val-Thr dipeptide may play an important role in controlling FGFR signalling through the Ras/MAPK pathway.
The four vertebrate fibroblast growth factor receptors (FGFRs) are paralogous members of the transmembrane receptor tyrosine kinase family. FGFRs are responsible for mediating cellular responses to fibroblast growth factor (FGF) signalling, which include mitogenesis, mesoderm induction, angiogenesis, chemotaxis and neuronal survival (1). Their importance is underlined by the wide variety of developmental abnormalities that can arise through mutations of FGFR1, -2 and -3 (2,3). Transduction of signals by FGFRs depends in part on the Ras/mitogen-activated protein kinase (MAPK) signalling pathway (reviewed in 4).
An important mechanism for increasing structural diversity and selective responsiveness amongst FGFRs is alternative splicing (5). For example, FGF binding specificity is altered in FGFR1, -2 and -3 by differential splicing of exons which make up the third immunoglobulin-like domain, whereby the IIIa exon (exon 7) is joined either to the IIIb (exon 8) or the IIIc exon (exon 9) (6-11). Amongst the many other FGFR splice forms are an FGFR1 variant with two amino acids (Arg-Met) deleted at the end of exon 4, which involves alternative use of two conventional (/GT) 5' splice sites (6,12,13), and variants of FGFR2 and FGFR3 with exon 7 spliced directly to exon 11, which lack the transmembrane region and may be targeted to the nucleus (14,15). A further splice variant of FGFR1 and -2 is generated by the use of alternative 5' splice sites located 6 bp apart at the 3'-end of exon 10, whereby splicing from the downstream site leads to inclusion of two additional amino acids, Val-Thr, in the intracellular juxtamembrane region (13). We refer to mRNA or protein containing and excluding this motif as VT+ and VT- respectively. It has been demonstrated in Xenopus (16) that the Thr residue in the FGFR1 VT+ isoform is phosphorylated by protein kinase C (PKC). VT+ forms, which predominate in cDNAs isolated from FGFR1 and FGFR2, have been identified for both paralogues from human (7,13,14,17,18), mouse (8,19-22), rat (23,24), chicken (25,26) and Xenopus (16,27). The minority VT- form has been observed in cDNA sequences from human (13), rat (23,28), Xenopus (16) and Notophthalmus (29) FGFR1 and human (14,30,31), rat (24) and mouse (S.R.F.Twigg and A.O.M.Wilkie, unpublished data) FGFR2. Sequences corresponding to the VT+ form have not previously been identified in FGFR3 or -4 in any species.
In this paper we describe the genomic structure of the mouse Fgfr2 gene and show that the VT+ isoform is generated by an extremely unusual 5' splice site. The same sequence is present in FGFR1 and FGFR3 and led us to predict the existence of a VT+ form of FGFR3, which we subsequently demonstrated. Phosphorylation of extracellular signal-regulated kinase (ERK2), a component of the Ras/MAPK signalling pathway, in the presence of the VT+ but not the VT- form of FGFR2 provides further evidence for the biological importance of this alternative splicing mechanism.
The Fgfr2 gene (GenBank accession nos Y16152-Y16167) contains 19 exons and extends over ~120 kb (Fig. 1), making it considerably larger than the previously defined human FGFR1 (6,32) and FGFR3 (33) and mouse Fgfr3 (34) genes. The number of exons and position of introns are, however, conserved, the size difference being due to the increased size of the Fgfr2 introns. Intron size has previously been shown to correlate with local GC content, whereby GC-poor isochores contain genes with on average three times longer intervening sequences (35). The GC4 value (GC content at the third position of four-fold redundant codons) of human FGFR2 (58%) is much lower than that of the other human FGFR genes (71-86%). In mice the trend is similar but less marked, the GC4 value of Fgfr2 (61%) being only slightly lower than the other three receptors (63-67%). There appears to be a relative size conservation of introns 7 and 8 between FGFRs which could reflect positional constraints associated with the alternative splicing of exons 8 and 9.
We compared the sequence of the mouse Fgfr2 exon 10-intron 10 boundary with the previously described genomic sequences for human (6) and Xenopus (16) FGFR1, human (33) and mouse (34) FGFR3 and human (38) FGFR4. Surprisingly, all except FGFR4 exhibit a 100% homologous 18 nt stretch, including the /GA dinucleotide, in this region (Fig. 2a). This suggests that the conserved sequence is required for alternative splicing and predicts that a VT+ form of FGFR3 should be synthesized. To test the latter hypothesis we isolated Fgfr3 cDNA by RT-PCR from mouse fibroblast- and brain-derived cell lines and hybridized the products to 32P-labelled oligonucleotides specific for the VT+ and VT- forms. A VT+ product was clearly detectable (data not shown) and a cDNA clone corresponding to this sequence was isolated (Fig. 2b). Examination of the 297 bp intron 10 sequence of the human FGFR3 gene (33) excluded the formal possibility that the GTAACA hexanucleotide could be encoded by a separate exon.
To determine whether the VT dipeptide of FGFR2 plays a functional role in signalling we analysed ERK2 phosphorylation in cells expressing either the VT- or VT+ form of FGFR2. Flow cytometric profiles indicate that the BaF3/FGFR2 VT- and BaF3/FGFR2 VT+ cell populations express comparable levels of FGFR2 on the cell surface; BaF3 cells transfected with vector alone display no absorbance shift (Fig. 4a). Figure 4b shows ERK2 phosphorylation in BaF3/FGFR2 VT+ cells (seen as a shift in mobility) 5 min after FGF2/heparin stimulation. In contrast, there is no detectable level of ERK2 phosphorylation in BaF3/FGFR2 VT- cells 15 min after stimulation. The difference in signalling potential between the VT+ and VT- cells is consistent with the observation that BaF3/FGFR2 VT+ cells show a weak proliferative response to FGF2/heparin, whereas BaF3/FGFR2 VT- cells exhibit no detectable response (data not shown). These findings indicate that the alternative splicing of the VT motif influences ligand-mediated signalling functions of FGFR2.
We have extended the report of Gillespie et al. (16), who described the /GA splice site in Xenopus FGFR1, to encompass three paralogous genes (FGFR1, -2 and -3) from three different species (Xenopus, mouse and human). Use of this atypical donor splice site may be a conserved feature of the biology of vertebrate FGFRs (except FGFR4). In the case of FGFR1 and -2 the atypical /GA site (corresponding to the VT+ form) seems to be utilized preferentially in alternative splicing (see Introduction). The 5' site motif ACA/GAAAGT involved in VT+ splicing is 100% conserved in the five available FGFR1, -2 and -3 genomic sequences (Fig. 2a) and appears to be the most divergent of any GT-AG intron yet examined. In a survey of 3294 5' splice sites no case was observed of a site exhibiting simultaneous mismatches to consensus at the -2, -1 and +2 positions (41). Although in a minority of cases a /GC dinucleotide is used at the start of the intron, these sequences otherwise conform better than average to the 5' consensus (42). Unlike /GA, the /GC dinucleotide supports a low level of correct completed splicing in vitro and in vivo (43,44). Pathological /GT -> /GA mutations have been described in many human genes, including CFTR, DMD, F9, HBB, PAH and SPTB (45). To our knowledge, only one other example of a 5' splice site that uses /GA has been documented, in the Antennapedia gene of Drosophila (46-48). In that case the overall sequence (AAG/GAAAGT) otherwise conforms perfectly to the consensus. It is interesting, however, that this splice site occurs in a very similar context: it is 12 nt downstream of a conventional 5' site and is involved in alternative splicing of a four amino acid motif. Although it is conceivable that these rare /GA splice sites could be modified in the pre-mRNA to /GU, no such RNA editing activity has yet been described (49).
The very rare occurrence of (+2)A at the 5' splice site raises the question of why (+2)U is normally required at this position in the pre-mRNA. Splicing involves two transesterification reactions (formation of lariat followed by exon joining) and proceeds through a number of intermediate complexes (50). In the first stage of U2-type splicesome assembly (E complex) U1 snRNA base pairs with the 5' splice site (-2 to +6) and (+2)U maximizes U1 complementarity (Fig. 3). However, U1 binding is dispensable in the presence of excess SR (serine/arginine-rich) proteins (51,52). In the subsequent B complex U1 is replaced at the 5' site by the U4/U5/U6 small nuclear ribonucleoprotein particle (snRNP), with binding of the U5 and U6 snRNAs to the exon (-3 to -1) and intron (+5,+6) respectively (40). It is unlikely that the (+2)U -> A substitution blocks these stages, because in vitro and in vivo studies indicate that it is compatible with formation of the lariat intermediate but exon joining is prevented (43). The U5 accessory protein p220 has been shown to contact (+2)U and the (+2)U -> A substitution reduces binding of a 5' splice site RNA oligonucleotide to U4/U5/U6 snRNP by >90% (53). U5 is involved both in lariat formation and exon joining and is the only snRNA shared by both U2-type and U12-type splicesomes. Conservation of a U at the +2 position of pre-mRNA is also seen in U12-type introns and raises the possibility that p220 also binds there (54). Given this potentially extreme functional conservation, the finding of introns with (+2)A is very unexpected.
How may alternative splicing be achieved at the VT- and VT+ sites, which are separated by only 6 nt? Splicing is inhibited when two alternative 5' splice sites are in close proximity, especially when both are consensus sites (55). In the case of FGFR1, -2 and -3, however, the alternative sites have a very different sequence structure (Fig. 3). The upstream VT- site predicts much stronger U1 binding; conversely, U6 binding will be favoured at the downstream VT+ site. We propose that relative splicing at the two sites represents a balance between a conventional U1-dependent process (VT-) and a U1-independent, U6/SR protein-dependent (51,56) process (VT+). This accords both with a model for alternative 5' splice site selection that requires low affinity for U1 at the proximal (downstream) site for splicing (which could be hnRNP-dependent; 57) to occur at a distal site (58) and with evidence that high concentrations of SR proteins favour use of proximal splice sites (59). In addition, proximity of the atypical VT+ site to the canonical VT- site may facilitate recognition of the VT+ site, which must be spliced to a 3' site that varies in separation from 297 bp (human FGFR3; 33) to 15 kb (mouse Fgfr2; this paper). The question of how exon joining is achieved at the VT+ site in the presence of /GA remains. Given that VT+/VT- splicing is a conserved feature of the biology of FGFR1, -2 and -3 and may be functionally important in the Ras/MAPK signalling pathway, we speculate that the ratio of these isoforms is controlled by a specific mechanism for recognition of the atypical 5' splice site, involving novel RNA and/or protein interactions. The very similar context in which /GA occurs in the Antennapedia gene of Drosophila raises the possibility that this reflects an evolutionarily conserved component of a specific pathway for alternative splicing of short sequences at the 5' site.
A mouse BEK cDNA probe containing exons 1-11 (21) was used to screen a [lambda]FIXII phage library derived from D3 mouse embryonic stem cell genomic DNA (Stratagene). The same probe was used to isolate a mouse yeast artificial chromosome (YAC) (ICRFy902C0595, C3H background; 60) containing the Fgfr2 gene.
Sizes for introns 7 and 8 have been published (37; see Table 1). The sizes of 13 of the 18 Fgfr2 introns were estimated by PCR using either the phage (for introns 1, 3, 5, 9 and 10) or YAC (for introns 11-18) as template and primers designed to the published mouse BEK cDNA sequence (22). (A complete list of oligonucleotides can be obtained from A.O.M.Wilkie). PCR was carried out in a volume of 25 µl, with 20 ng DNA, 350 µM dNTPs, 10 pmol each primer, 50 mM Tris, pH 9.2, 16 mM (NH4)2SO4, 2.25 mM MgCl2, 2.5 U Amplitaq (Perkin Elmer) and 0.25 U Pfu DNA polymerase (Stratagene). Thermocycling was performed on a Hybaid OmniGene Temperature Cycler and consisted of a 2 min denaturing step followed by 30 cycles of 94°C for 30 s, 50°C for 30 s and 68°C for 8 min, with a 20 s increment in the elongation step every cycle. Product sizes were estimated by comparison with [lambda] HindIII markers. The remaining three introns (2, 4 and 6) did not amplify by PCR. Their sizes were estimated by restriction mapping of the Fgfr2 YAC. Eight micrograms of YAC DNA was digested with 20 U EcoRI, BamHI or BglII in single or double digests, electrophoresed and blotted and hybridized to probes specific to exons either side of the intron. The sizes of all introns except 4 and 6 were confirmed from two independent sources.
For all boundaries except those surrounding introns 2, 4 and 6 PCR products obtained above were gel purified (QIAquick gel extraction kit; Qiagen) and the nucleotide sequence determined using a Thermo Sequenase cycle sequencing kit (Amersham) and one of the primers used in the PCR. To obtain the remaining boundary sequences two strategies were adopted. (i) Different phages containing exons 2-7 were digested with MboI and the DNA fragments cloned into BamHI-cut, alkaline phosphatase-treated pUC18 (Pharmacia). The libraries were screened with corresponding exon-specific probes. Positive clones were sequenced with the relevant exon-specific primer as above. (ii) HaeIII-digested Fgfr2 YAC DNA was ligated to blunt-ended Vectorette II units (Genosys) (according to the manufacturer's instructions) and the products used as templates for PCR amplification of boundaries with a Vectorette-specific primer and an exon primer. Vectorette PCR was carried out in a total volume of 25 µl, with 4/50 µl ligated DNA, 350 µM dNTPs, 10 pmol each primer, 50 mM Tris, pH 9.2, 16 mM (NH4)2SO4, 2.25 mM MgCl2, 0.1 U Amplitaq (Perkin Elmer). Thermocycling consisted of a 2 min denaturing step followed by 30 cycles of 94°C for 30 s, 55°C for 30 s and 72°C for 30 s. Products were sequenced with the exon primer. The boundaries were determined by aligning the resulting sequences with the Fgfr2 cDNA sequence.
RNA was extracted (61) from embryonic fibroblasts and brain-derived cell lines (CBA * C57/BL). Ten micrograms of total RNA were used as template for first strand synthesis in 40 µl using 340 pmol random hexamer primers, 32 U RNase inhibitor (Boehringer Mannheim) and 400 U MMLV reverse transcriptase (Gibco BRL). Four microlitres of the product were used for PCR in a total volume of 25 µl using the same buffer and conditions as described above for the Vectorette PCR, except that the annealing temperature was 58°C. Oligonucleotide pairs used to amplify the VT+/VT- region of Fgfr2 (22) and Fgfr3 (62) respectively were designed to exons 10 (forward) and 11 (reverse): MR2TM-F, 5'-CGCCTGTGAGAGAGAAGGAG-3'; H9R, 5'-GGCGTGTTGTTATCCTCACCAG-3'; MR3E10F, 5'-TCCTCAGCTACGGGGTG-3'; MR3E11R, 5'-CCTTCTCCAAGAGGCTTAC-3'.
RT-PCR products were gel purified and ligated to pT7 (Novagene). The original products and EcoRI/XbaI-digested clones were electrophoresed on a 6% polyacrylamide denaturing gel. The gel was blotted onto Hybond N+ (Amersham) and hybridized either to a random primed 32P-labelled probe (using the exon 10-exon 11 RT-PCR product as template) or to oligonucleotides (labelled with [[alpha]-32P]dCTP using terminal transferase) specific for the VT+ and VT- products: VT+, 5'-GCGACAGGTAACAGTGTCC-3'; VT-, 5'-AAGCGACAGGTGTCCTTGG-3').
Human FGFR2 (VT-) cDNA (TK14 clone; 30) was inserted into the 6P-IresNeo-[beta]S vector (63; kindly provided by A.Smith) as an SacI-XbaI fragment. FGFR2 VT+ in 6P-IresNeo-[beta]S was constructed using overlap PCR (64) to introduce the VT-encoding hexanucleotide at the end of exon 10 (oligonucleotides: FGFR25', 5'-TTTCCGAGCTCATGGTCAGCTGGGGTCGT-3'; HB9, 5'-GCTGGACTCAGCCGAAACTGTTACTGTTCCGCAGGG- GG-3'; HB10, 5'-TTCGGCTGAGTCCAGCTCC-3'; FGFR2 3', 5'-GACAGTCTAGATTCATGTTTTAACACTGCC-3'). Twenty micrograms of plasmid DNA was transfected by electroporation (200 V, 960 µF) into BaF3 mouse proB cells in cold phosphate-buffered saline (PBS). Electroporated cells were incubated in RPMI 1640 growth medium (Gibco BRL) supplemented with 10% fetal calf serum (FCS; Labtech) and 1 ng/ml recombinant mouse IL3 (R&D Systems) for 24 h. Stably transfected cells were selected by adding 600 µg/ml geneticin (Gibco BRL) to the growth medium. FGFR2 expression on the cell surface was confirmed by flow cytometry analysis. A polyclonal sheep antibody against the human FGFR2 extracellular domain (Alta Bioscience) was incubated at 1/1000 with 1 * 106 geneticin-resistant cells (in 100 µl PBS, 10% FCS) at 4°C for 90 min. Cells were then incubated with a fluorescein isothiocyanate-conjugated Dk anti-Sh antibody (diluted 1/200; Jackson ImmunoResearch Laboratories) for 30 min at 4°C and analysed in 1 ml 1% paraformaldehyde by flow cytometry on a Becton Dickinson FACS 440. BaF3/FGFR2 cells (2.5 * 106) which had been serum starved in UltraCHO (Bio Whittaker) in the absence of IL3 for 5 h were incubated with FGF2 (100 ng/ml) and heparin (2 µg/ml). The cells were lysed on ice for 30 min in 100 µl lysis buffer [20 mM Tris, pH 8, 5 mM MgCl2, 10 mM EGTA, 1% Triton X-100, 1 mM Na3VO4, 50 mM NaF and 1 mini tablet/10 ml protease complete (Boehringer Mannheim)]. Cell lysates were separated on 15% SDS-polyacrylamide gels, transferred onto polyvinylidene difluoride membrane and incubated with anti-ERK2 antibody 122 (1:7500) (kindly provided by C.Marshall), which was detected using peroxidase-conjugated anti-rabbit antibody (Amersham) and enhanced chemiluminescence (Pierce).
We are grateful to R.Cox, L.Kearney, S.Horsley and R.Regan for help in isolating and characterizing the YAC, F.Gleig and L.Mahadevan for advice on ERK2 assays, Z.Larin for discussions and G.Screaton for comments on the manuscript. This work was funded by the Wellcome Trust (J.K.H. and A.O.M.W.).
Human Molecular Genetics
Pages
Introduction
Results
Genomic structure of mouse Fgfr2
Conservation of VT+/VT- sequences and splicing in FGFR1, -2 and-3
Functional comparison of VT+/VT- splice variants
Discussion
Materials And Methods
Isolation of mouse genomic Fgfr2 DNA
Fgfr2 intron size
Sequence of Fgfr2 intron-exon boundaries
RNA isolation and reverse transcriptase PCR (RT-PCR)
VT+/VT- analysis
ERK2 phosphorylation assay
Acknowledgements
References
REFERENCES
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 14 Mar 1998
Copyright© Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. C. Barnard, A. J. Williams, B. Rabier, O. Chassande, J. Samarut, S.-y. Cheng, J. H. D. Bassett, and G. R. Williams Thyroid Hormones Regulate Fibroblast Growth Factor Receptor Signaling during Chondrogenesis Endocrinology, December 1, 2005; 146(12): 5568 - 5580. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K J Bradley, B M Cavaco, M R Bowl, B Harding, A Young, and R V Thakker Utilisation of a cryptic non-canonical donor splice site of the gene encoding PARAFIBROMIN is associated with familial isolated primary hyperparathyroidism J. Med. Genet., August 1, 2005; 42(8): e51 - e51. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X. Li, V. G. Brunton, H. R. Burgar, L. M. Wheldon, and J. K. Heath FRS2-dependent SRC activation is required for fibroblast growth factor receptor-induced phosphorylation of Sprouty and suppression of ERK activity J. Cell Sci., December 1, 2004; 117(25): 6007 - 6017. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. L. Kirby, L. Yang, J. C. Labus, and B. T. Hinton Characterization of Fibroblast Growth Factor Receptors Expressed in Principal Cells in the Initial Segment of the Rat Epididymis Biol Reprod, June 1, 2003; 68(6): 2314 - 2321. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. D. Lalioti and J. K. Heath A new method for generating point mutations in bacterial artificial chromosomes by homologous recombination in Escherichia coli Nucleic Acids Res., February 1, 2001; 29(3): e14 - e14. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Arman, R. Haffner-Krausz, M. Gorivodsky, and P. Lonai Fgfr2 is required for limb outgrowth and lung-branching morphogenesis PNAS, October 12, 1999; 96(21): 11895 - 11899. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Popovici, B. Zhang, M.-J. Gregoire, P. Jonveaux, M. Lafage-Pochitaloff, D. Birnbaum, and M.-J. Pebusque The t(6;8)(q27;p11) Translocation in a Stem Cell Myeloproliferative Disorder Fuses a Novel Gene, FOP, to Fibroblast Growth Factor Receptor 1 Blood, February 15, 1999; 93(4): 1381 - 1389. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. R. Burgar, H. D. Burns, J. L. Elsden, M. D. Lalioti, and J. K. Heath Association of the Signaling Adaptor FRS2 with Fibroblast Growth Factor Receptor 1 (Fgfr1) Is Mediated by Alternative Splicing of the Juxtamembrane Domain J. Biol. Chem., February 1, 2002; 277(6): 4018 - 4023. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||