Life, 48: 73±78, 1999 Copyright ° c 1999 IUBMB 1521-6543 /99 $12.00 + .00 IUBMB
Original Research Article Fibroblast Growth Factor Receptor-4 Splice Variants Cause Deletion of a Critical Tyrosine Walter R. A. van Heumen, Christina Claxton, and James O. Pickles Vision, Touch and Hearing Research Centre, Department of Physiology and Pharmacology, University of Queensland, Brisbane, QLD-4072, Australia
Summary We have identi® ed two novel isoforms of ® broblast growth factor receptor-4 (FGFR4). They result from alternative splicing of intron 17. Two transcripts, both slightly larger than the one coding for the known mouse FGFR4, are generated. The shortest (FGFR417a) includes the 31-most 30 -nucleotides of intron 17; the longest (FGFR4-17b) includes all 114 nucleotides of intron 17. Translation of the FGFR4-17a and FGFR4-17b splice variants predicts that both novel putative FGFR4 isoforms have a truncated C-terminal intracellular tail. The ® rst amino acid residue affected by the insertions in both novel isoforms is Tyr-760, a residue that may play a crucial role in intracellular signaling through stimulation of the phosphatidylinositol-biphosphat e pathway. IUBMB Life, 48: 73 ± 78, 1999
tions. This is suggested by evidence that the different extracellular splice variants have different af® nities for the various FGFs (3, 5), by the different downstream pathways stimulated by the intracellular splice variants of FGFR2, and by the tissue-speci® c expression of the different FGFR isoforms. To date, and in contrast, no splice variants for FGFR4 have been described in either the extracellular or intracellular domains. This study demonstrates the expression of 2 novel FGFR4 subtypes in the mouse. These isoforms are the result of alternative splicing of the region corresponding to intron 17 of the human FGFR4 gene (6), to affect expression of a tyrosinecritical for the intracellular signaling pathways. EXPERIMENTAL PROCEDURES
Keywords
Alternative splicing; cochlea; FGFR4 isoform; ® broblast growth factor receptor-4; tyrosine kinase.
INTRODUCTION Fibroblast growth factor receptors (FGFRs)1 have a very high af® nity for members of the ® broblast growth factor (FGF) family, of which at least 9 different members have been described to date (1). The FGFRs form a subgroup within the transmembrane tyrosine kinase family and consist of 4 closely related members, FGFR1± 4, each encoded by its own gene (2). Several isoforms of the FGFR1, 2, and 3 receptors are generated as a result of alternative splicing of the transcript in the extracellular domain, and in the case of FGFR2 also by alternative splicing in the intracellular domain (3± 5). The different splice variants of the FGFRs 1±3 are likely to have different biological func-
Received 1 December 1998; accepted 26 January 1999. Address correspondence to Walter R.A. van Heumen. Fax: +61-733654522 . E-mail:
[email protected] 1 Abbreviations : FGF, ® broblast growth factor; FGFR, FGF receptor; RTPCR, reverse transcription polymerase chain reaction.
Reverse Transcription–Polymerase Chain Reaction (RT-PCR) Quackenbush mice were obtained from the University of Queensland Animal Production Department. Total RNA from spleen, skeletal muscle, kidney, lung, heart, whole brain, liver, and testis obtained from newborn mice was isolated by using Tri-reagent (Molecular Research Centre, Inc.). Poly(A)+ RNA was isolated from 20 l g of each total RNA by extraction with oligo-d(T) 25 magnetic beads (Dynal, Australia). The poly(A)+ RNA in the eluted fraction was further puri® ed by a second round of extraction with oligo-d(T)25 magnetic beads. In addition, poly(A)+ RNA was isolated from 3 whole cochleae and 14 sensory areas of utricles and saccules of newborn mice by the same procedure. All eluates were subsequently subjected to incubation with RNase-free DNase I according to the speci® cations of the supplier (Pharmacia). Poly(A)+ RNA was immediately reverse-transcribed in a 20-l l reaction mix to cDNA with use of Superscript II reverse transcriptase (Gibco BRL). Of the obtained cDNA, 5% was added to 20 l l of PCR reaction-mix (1 £ reaction buffer: 2 mM MgCl2 , 200 l M dNTPs, 0.15 l M of each primer, 1 l l of cDNA, and 0.5 U of Taq polymerase; Perkin-Elmer Cetus, Applied Biosystems, Brisbane, Australia). 73
74
VAN HEUMEN ET AL.
Figure 1. Schematic representation of part of the mouse FGFR4 gene, its transcripts, and the position of the isoform-speci® c PCR primers. Primers P2, P3, and P4 are speci® c for FGFR4, FGFR4-17a, and FGFR4-17b, respectively. Primer P1 is common to all 3 isoforms. Expression of the different FGFR4 isoforms in the above tissues and in the vestibular sensory epithelium was studied by RT-PCR with a downstream (30 -most) primer P1 designed to anneal to a sequence in the 30 -untranslated region common to all 3 FGFR4 isoforms, used in combination with an upstream (50 -most) isoform-speci® c primer (P2, P3, or P4). Primer P4 is complementary to a sequence unique for FGFR4-17b. To obtain FGFR4 and FGFR4-17a, speci® c primers P2 and P3 were designed in such a way that they matched the exon 17/exon 18 boundary (P2) or the exon 17/exon 17a boundary (P3) and therefore could anneal only to FGFR4 or FGFR4-17a cDNAs, respectively (Fig. 1). The following PCR primers were used: reverse primer P1 (all FGFR4 isoforms): 50 -GAGGCTTGTGTGCTGGGGGAT30 ; forward primer P2 (speci® c for FGFR4): 50 -TGTCTCTGAAGAG /TACCTTGACC-30 ; forward primer P3 (speci® c for FGFR4-17a): 50 -GGCTGTCTCTGAAGAGgctgg-30 ; forward primer P4 (speci® c for FGFR4-17b): 50 -atgaacacgaagcatcctgtcccc-30 ; forward primer P5 (in exon 10, all FGFR4 isoforms): 50 -TGGTTCTGCTTGTGCTCCTGC-3 0 ; reverse primer P6 (FGFR4-17a and FGFR4-17b): 50 -GCAAGGAACAGAGCTGGTCAG-3 0 . Lower-case letters in primers P3 and P4 represent intron 17 sequences, the ª /º in P2 indicates the exon 17/exon 18 boundary.
Sequence Analysis Cloned cDNAs were sequenced by using the Dye Deoxy Terminator Cycle Sequencing Kit (Perkin-Elmer) and an automated DNA sequencer (Model 373A, ABI). Sequences were analysed by AssemblyLign and MacVector software (Kodak Scienti® c Imaging Systems) and compared with nucleotide sequences in GenBank by use of BLAST (http://www.ncbi.nlm.nih.gov/ BLAST/).
RESULTS
Alternative Splicing of FGFR4 Transcripts Sequence analysis of products from an FGFR4 screen in inner ear and kidney cDNAs revealed 3 different FGFR4-like PCR products. Two of the products appeared to contain inserts of 31 and 114 bp, respectively, compared with the previously described FGFR4 cDNA sequence (Fig. 2). PCR ampli® cations of cDNA derived from spleen, skeletal muscle, kidney, lung, heart, whole brain, liver, testis, cochleae, and vestibular epithelium, through use of primer combinations designed to speci® cally amplify 1 or more of the identi® ed isoforms, generated products of the sizes obtained in the initial screen (P2/P1: 583-bp FGFR4 fragment; P3/P1: 617-bp FGFR4-17a fragment; and P4/P1: 628-bp FGFR4-17b fragment). Sequence analysis of the cloned PCR products con® rmed that the size and sequences of the PCR products were the same in each tissue. Translation of the FGFR4-17a and 17b transcripts yielded 2 novel putative FGFR4 isoforms, both with truncated cytoplasmic C-terminal tails arising from a premature stop codon. The C-terminal tails of the putative FGFR4-17a and FGFR4-17b are, respectively, 49 and 33 amino acid residues shorter than FGFR4 (Fig. 3). The amino acid sequences of the C-terminal tails of FGFR4-17a and FGFR4-17b have no homology to other known sequences in the protein databases. The ® rst amino acid modi® ed in the truncated isoforms is a tyrosine residue at position 760. This is homologous to Tyr-769, which undergoes similar modi® cation in the FGFR2 K-sam isoform (7). Tissue Distribution of FGFR4s The expression and relative abundance of FGFR4, FGFR417a, and FGFR4-17b were studied by PCR ampli® cation of cDNA derived from spleen, skeletal muscle, kidney, lung, brain,
FGFR4 VARIANTS
75
Figure 2. Nucleotide alignments of parts of the FGFR4, FGFR4-17a, and FGFR4-17b transcripts. The sequences are aligned with sequences from the mouse (GenBank accession number MMFGFR4M). Numbers indicated for the ® rst and last nucleotide of the FGFR4 cDNA fragment correspond to their positions in FGFR4. Introduced gaps are represented by dots. heart, liver, testis,cochleae, and vestibular epithelium of neonate mice with the 3 different primercombinations, 1 speci® c for each isoform (Figs. 4, 5). In general, the most intense bands were obtained with primers P1/P2 amplifying FGFR4; the weakest, with P1/P3 amplifying FGFR4-17a. The intensities of the bands representing the same FGFR4 isoform differed considerably in the various tissues. Analyses of the cDNAs for FGFR4 with constant amounts of starting total RNA revealed that intense bands were generated from kidney, muscle, liver, and cochlear material; bands of weaker intensity were generated from heart, brain, and vestibular epitheliumcDNAs, and a faint band was generated from testiscDNA. Although bands representing FGFR4-17a and FGFR4-17b tended to be most intense in tissues showing the most intense bands for FGFR4, this was not universally the case (e.g., in Fig. 4 compare the intensities of the FGFR4-17b bands generated from muscle and kidney cDNAs and compare those with the intensities of the FGFR4 bands generated from same two cDNAs). Ampli® cation of GAPDH reference gene fragments yielded bands of equal intensity in all lanes, whereas no bands were observed in the RT negative controls (data not shown).
Sequence analysis (Fig. 2) demonstrated the FGFR4-17a and b isoforms contained an insert at exactly the same position as intron 17 in human FGFR4. The question arose whether these isoforms were genuine RT-PCR products or were artefacts generated from minute amounts of contaminating genomic DNA or from immature forms of mRNA in which the introns had not yet been removed. To test this hypothesis, we ampli® ed the above cDNAs by PCR, using primers (P5/P6) that anneal to sequences separated by 4290 nt on the genome and span a region containing 7 introns; ampli® cation from contaminating genomic DNA or immature mRNA should yield 1 product of 4290 bp, whereas ampli® cation from cDNA should result in products of 1184 bp for the FGFR4-17a isoform and 1298 bp for FGFR417b. Gel analysis of the PCR reaction mix revealed the presence of 2 bands, » 1200 and 1300 bp in size (Fig. 6); longer products were not found. Sequence analysis con® rmed that the ampli® ed products did not contain intron sequences other than intron 17 and therefore were indeed coding for fragments of the 2 novel isoforms. We conclude that these 2 PCR products are derived from 2 separate transcripts and code for 2 novel isoforms of FGFR4.
Figure 3. C-terminal tails of FGFR4 isoforms. Comparison of the amino acid sequences of C-terminal tails of mouse FGFR4, FGFR4-17a, and FGFR4-17b. The position of Tyr 760 in FGFR4 is indicated.
76
VAN HEUMEN ET AL.
Figure 4. Agarose gel analysis of PCR products from mouse spleen, skeletal muscle, kidney, lung, brain, heart, liver, and testis cDNAs. Electrophoresis of the products after PCR ampli® cation of the cDNAs with isoform-speci® c primer combinations P2/P1 (top panel), P3/P1 (middle panel), and P4/P1 (lower panel) was as described in the experimental procedures. Arrowheads indicate the positions and expected sizes of the isoform-speci® c PCR products. BPL: molecular mass marker (100-bp DNA ladder; Bio-Rad). DISCUSSION In this study we have identi® ed the expression and tissue distribution of two novel FGFR4-like transcripts, generated by alternative splicing of an intron corresponding to human intron 17. The novel transcripts (FGFR4-17a and FGFR4-17b isoforms) were slightly bigger than the transcript that was expected on the basis of the previously published FGFR4 sequence described by Stark et al. (8). Examination of the nucleotide sequence of intron 17 revealed the presence of a potential RNA splice site that if used would give rise to the FGFR4-17a variant (9). Translation of the cDNAs showed that both transcripts have truncated cytoplasmic C-terminal tails. Compared with the described FGFR4, the C-terminal tails of the putative FGFR4-17a and FGFR4-17b are 49 and 33 amino acid residues shorter, re-
spectively, resulting in C-terminal tails of 5 and 21 residues, respectively (Fig. 5). The amino acid sequences of the C-terminal tails of FGFR4-17a and FGFR4-17b have no homology to other known sequences in the protein databases. The ® rst amino acid to be modi® ed in the 2 novel isoforms is a tyrosine residue at position 760. This Tyr-760 is homologous to Tyr-769 in FGFR2, which undergoes similar modi® cation in the K-sam (7) isoform, and Tyr-766 in FGFR1. Evidence indicates that phosphorylation of Tyr-766 plays an important role in the intracellular signal transduction pathway of FGFR1; phosphorylation of the tyrosine stimulates phosphatidylinositol hydrolysis, whereas DNA synthesis or cellular differentiation appear to be induced by tyrosines situated nearer to the A-terminus (10, 11). Furthermore, phosphorylation of Tyr-766 appears to be important for internalisation of the ligand±receptor complex
FGFR4 VARIANTS
77
Figure 5. Agarose gel analysis of PCR products from mouse cochlea and utricle + saccule sensory epithelium cDNAs. Electrophoresis of the products after PCR ampli® cation of the cDNAs with isoform-speci® c primer combinations for FGFR4, FGFR417a, and FGFR4-17b was as described in the experimental procedures. BPL: molecular mass marker (100-bp DNA ladder; Bio-Rad). (12). Since FGFR1 without Tyr-766 is capable of activating the MAP kinase secondary messenger system, it is possible that these 2 FGFR4-17a and 17b isoforms, neither of which contains a Tyr-766 homologue, play a similar role; i.e., they selectively activate only MAP kinase system and thereby initiate a cellular response that is different from the 1 activated by FGFR4. These different FGFR4 isoforms may therefore play different roles in FGF-mediated activation of cellular processes.
The expression of the 3 isoforms was investigated by RT-PCR in spleen, skeletal muscle, kidney, lung, brain, heart, liver, testis, cochlea, and the sensory epithelium of the utricle, with use of primer pairs that speci® cally amplify only 1 of the 3 isoforms. Gel analysis of the PCR reaction mixes revealed that not all 3 isoforms are expressed in all tissues tested. Furthermore, bands representing the same FGFR4 transcripts in different tissues can differ considerably in intensity. This may re¯ ect tissue-speci® c differences in the abundance of the different FGFR4 isoforms. Differential expression of the various FGFR4 isoforms may be a way for the cells to ® ne-tune their responses to FGFs. ACKNOWLEDGEMENTS We thank Victor Nurcombe for his encouraging comments. This work was supported by grants from the Garnett-Passe and Rodney Williams Memorial Foundation. The nucleotide sequence data reported appear in the GenBank Nucleotide Sequence Database under accession number AF127140 (FGFR417b) and AF127141 (FGFR4-17a). REFERENCES
Figure 6. Agarose gel analysis of FGFR4-17a and FGFR417b PCR products. PCR products from mouse kidney cDNA were generated by primer combination P5/P6 as described in the experimental procedures. The FGFR4-17a-speci® c PCR product (1184 bp) and the FGFR4-17b-speci® c PCR product (1298 bp) are indicated. BPL: molecular mass marker (100-bp DNA ladder; Bio-Rad).
1. Hughes, S. E. (1997) Differential expression of the ® broblast growth factor receptor (FGFR) multigene family in normal human adult tissues. J. Histochem. Cytochem. 45, 1005 ±1019. 2. Johnson, D. E., and Williams, L. T. (1993) Structural and functional diversity in the FGF receptor multigene family. Adv. Cancer Res. 60, 1±41. 3. Champion Arnaud, P., Ronsin, C., Gilbert, E., Gesnel, M. C., Houssaint, E., and Breathnach, R. (1991) Multiple mRNAs code for proteins related to the BEK ® broblast growth factor receptor. Oncogene 6, 979 ±987. 4. Dionne, C. A., Crumley, G., Bellot, F., Kaplow, J. M., Searfoss, G., Ruta, M., Burgess, W. H., Jaye, M., and Schlessinger, J. (1990) Cloning and
78
5.
6. 7.
8.
VAN HEUMEN ET AL. expression of two distinc t high-af® nity receptors cross-reacting with acidic and basic ® broblast growth factors. EMBO J. 9, 2685 ±2692. Ornitz, D. M., and Leder, P. (1992) Ligand speci® city and heparin dependenc e of ® broblast growth factor receptors 1 and 3. J. Biol. Chem. 267, 16305 ±16311. Kostrzewa, M., and Muller, U. (1998) Genomic structure and complete sequence of the human FGFR4 gene. Mamm. Genome 9, 131 ±135. Hattori, Y., Odagiri, H., Nakatani, H., Miyagawa, K., Naito, K., Sakamoto, H., Katoh, O., Yoshida, T., Sugimura, T., and Terada, M. (1990) K-sam, an ampli® ed gene in stomach cancer, is a member of the heparin-bindin g growth factor receptor genes. Proc. Natl. Acad. Sci. U.S.A. 87, 5983 ± 5987. Stark, K. L., McMahon, J. A., and McMahon, A. P. (1991) FGFR-4, a new member of the ® broblast growth factor receptor family, expressed in the de® nitive endoderm and skeletal muscle lineages of the mouse. Development 113, 641±651.
9. Breathnach, R., and Chambon, P. (1981) Organization and expression of eucaryotic split genes coding for proteins. Ann. Rev. Biochem. 50, 349± 383. 10. Burgess, W. H., Dionne, C. A., Kaplow, J., Mudd, R., Friesel, R., Zilberstein, A., Schlessinger, J., and Jaye, M. (1990) Characterization and cDNAcloning of phospholipas e C-gamma, a major substrate for heparin-bindin g growth factor 1 (acidic ® broblast growth factor)-activated tyrosine kinase. Mol. Cell. Biol. 10, 4770 ±4777. 11. Mohammadi , M., Honegger, A. M., Rotin, D., Fischer, R., Bellot, F., Li, W., Dionne, C. A., Jaye, M., Rubinstein, M., and Schlessinger, J. (1991) A tyrosine-phosphorylate d carboxy-termina l peptide of the ® broblast growth factor receptor (Flg) is a binding site for the SH2 domain of phospholipas e C-gamma 1. Mol. Cell. Biol. 11, 5068 ±5078. 12. Sorokin, A., Mohammadi, M., Huang, J., and Schlessinger, J. (1994) Internalization of ® broblast growth factor receptor is inhibited by a point mutation at tyrosine 766. J. Biol. Chem. 269, 17056 ±17061.
