Abstract
The adaptive immune system (AIS) is characterized by the MHC molecules and the rearranging Ag receptors, and was established in a common ancestor of jawed vertebrates. Fyn, a Src-family tyrosine kinases, is important for normal development and function of T lymphocytes and neuronal cells. Indeed, as the result of an alternative splicing of a distinct exon 7, fyn encodes for two isoforms, FynT in T lymphocytes and FynB in the brain. How this alternative splicing of fyn transcripts has emerged and evolved in relation to the setting of the AIS remains to be established. In this study, we show that exon capture in a vertebrate ancestor by the fynT-like gene has yielded a novel fyn-encoded isoform, fynB. Unexpectedly, the newly established AIS recruited the ancestral Fyn isoform, FynT, whereas the CNS expresses the most recent one, FynB. These results shed new light on the emergence of the AIS.
Many of the genes encoding for key molecules of the adaptive immune system (AIS)3 are thought to exist in jawed vertebrates only. Their functions were likely established in a common ancestor of jawed vertebrates (1, 2). The proper function of the AIS further requires the integrity of a complex network of signaling proteins. Fyn, a member of the Src family of protein tyrosine kinases, is implicated in the development and function of both T lymphocytes and neurons (3, 4). Indeed, as the result of an alternative, mutually exclusive splicing of a distinct exon 7, the fyn gene encodes for two isoforms, FynT in T lymphocytes and FynB in the brain (5). This exon 7 encodes for an intramolecular regulatory linker segment implicated in the regulation of the Src kinases, Src and Hck (6, 7, 8, 9). How the alternative splicing of fyn transcripts has emerged and evolved in relation to the setting of the AIS remains to be established.
Vertebrate Src families share a common genomic organization with similar exon/intron organization and exon size. Based on this conservation, degenerated primers were designed and used to PCR amplify sequences from exon 6 to 8 from the cartilaginous fish Torpedo marmorata, from the bony fish Xiphophorus helleri, and from the amphibian Xenopus laevis. The PCR products were sequenced bidirectionally and Blast searches of these sequences, translated in all reading frames against a protein sequence database, were performed. An exon 7B was retrieved for the three species, whereas exon 7T was only identified in the amphibian and cartilaginous fish species. The best match for these sequences in the GenBank database (National Center for Biotechnology Information, Bethesda, MD) was with the chicken exon 7T peptide. Exon 7T from the cartilaginous fish showed a similarity ranging from 61 to 81% with the other exon 7T found in the database. The essential residues for the interaction of the intramolecular linker segment are conserved (Fig. 1⇓a). Phylogenetic analysis of the two isoforms showed that exon 7T from amphibian and cartilaginous fish species are orthologous to exon 7T of other species (Fig. 1⇓b). The comparison between the human and cartilaginous fish complete fyn protein sequences gave 82% identity, whereas the comparison for exons 7T and 7B gave 61 and 96% identity, respectively. These data suggested that each domain evolved independently, at a different speed, under a different selection pressure. We next verified that the mutually exclusive splicing of alternate exon 7 is conserved in amphibians. RT-PCR analysis was performed on amphibian thymus and brain-derived mRNA (Fig. 1⇓c). Amplification products of 266 and 275 nt were obtained in the thymus and the brain, respectively, and their sequences were verified (not shown). Thus, the alternate splicing of fyn exon 7 and the expression pattern of the two fyn isoforms existed before the amphibian-amniote split.
Phylogenetic relation of different fynT and fynB species and differential expression in amphibians. a, The fyn gene exon 7 sequences from different species coding for FynT and FynB proteins were aligned using Clustal W (Blosum 30 matrix). b, Phylogenetic analysis of sequences aligned in a. The tree was constructed using the NJ algorithm of the Mega program using the Poisson correction distance. Numbers above the branches indicate the BP (based on 500 iterations). Only bootstrap values above 50% are given. c, RT-PCR amplification of fyn transcripts in amphibians using primers matching exon 6 and 8.
To further determine the evolutionary relationship of the alternatively spliced exon 7 of fyn with the remainder of the gene, we aligned the SH1 catalytic domains of src family members found in the genome databases (Table I⇓). Because previous studies have already identified fyn as a member of the srcA subgroup composed of src, yes, fyn, and fgr (10), we conducted our analysis on this group. Rooted at the midpoint, the tree based on the catalytic domain revealed that srcA members formed various phylogenetic groups (Fig. 2⇓a). Fgr and src, yes, fyn, yrk and fyk formed a distinct group (bootstrap proportion (BP) = 98%) itself formed of two subgroups: one subgroup contained src, yes, fyn, yrk, fyk (BP = 89%), itself divided in yes and fyn, yrk, fyk minor subgroups. The second subgroup contained fgr (BP = 100%). We then searched for the presence of exons 7T and 7B for all srcA group members. Genomic sequences were available for src, yes and fgr, but not for fyk and yrk. Hence, except for fyn, all other members had a single exon 7. Next, the phylogenetic analysis was performed on the available exon 7 sequences. As shown in Fig. 2⇓b, exons 7T and 7B formed two monophyletic groups (BP were respectively equal to 79 and 79%). Rooting at the midpoint, the fyn exon 7T clustered with src, yes (BP = 54%), which itself clustered with fgr exon 7 (BP = 67%). By contrast, exon 7B clustered with the yrk and fyk exon 7, with a very significant BP (99%; Fig. 2⇓b). These data demonstrate that src, yes, fgr and fyn duplicated from a common exon 7T-like-containing ancestor and that fyn captured exon 7B after its separation with yes.
FynB exon 7 forms a distinct monophylogenetic group. Phylogenetic analysis of the SH1 catalytic domains (a) and of exon 7 (b) of srcA tyrosine kinases. The sequence alignments were obtained using Clustal X (Blosum 30 matrix), and the phylogenetic trees were constructed on a Poisson correction distance using the NJ algorithm of the Mega program. Numbers represent the percentage of 500 bootstraps supporting the branch.
Evolution of the fyn gene
These data show that not only fynB, but also fynT exon 7, are present in jawed vertebrates, except in bony fish (where fyn exon 7 encoding for the FynT protein could not be PCR amplified). Because the bony fishes clade forms a monophyletic group with tetrapoda, the cartilaginous fishes being a sister group of the bony fishes and tetrapoda, the most parsimonious hypothesis is that bony fishes lost exon 7T during their evolution. Nonetheless, the interspecies conservation of this alternate splicing suggests a biological preponderant role of the protein encoded by these two exons in the cells in which their expression is restricted. The exon 7 encodes the linker region (linker SH2-SH1) connecting the SH2 and SH1 catalytic domains and is implicated in the regulation of Src protein by intra- and intermolecular interactions.
The present report shows that exon 7B was recruited for a brain function while expression of the ancestral fyn gene (fynT) was conserved by the AIS (Fig. 3⇓). Two mechanisms are possible. The different factors essential for the alternative exon 7B expression were only present in the brain and/or fynB and fynT are essential for optimal function in the tissues in which they are selectively expressed. We propose that the “mechanisms” allowing for the alternative splicing of fyn exons 7T and 7B were cocaptured together with exon 7B. The key role played by the linker region encoded by exon 7 of src kinases in folding the enzyme and maintaining it in its inactive conformation (8, 9), together with the oncogenic potential of the activated form of the Fyn protein (11), both support this hypothesis. The role of co-option, in this study, would be to stabilize factors that maintain redundancy. Indeed, changes in expression patterns often, although not always, are associated with many of the cases of co-option that have been proposed (12).
Hypothetical general scheme for the evolution of the fyn gene. The X B-like gene contains the ancestral fyn exon 7B, but its actual representative could not be retrieved from the databases, either because it was lost during evolution or because the corresponding sequences are not yet available in the databases. The ancestral fyn exon 7B was captured by the fynT-like before or during the radiation of the jawed vertebrates.
In a previous report, Hughes (10) proposed that the fyn exon 7T was derived by a recombinational event with another gene (by “exon shuffling”), perhaps one related to fgr . Since this preliminary study, additional sequences became available, including those reported in this study. Together with the presently available databases, our data fail to support Hughes’ hypothesis. The phylogenetic analysis with other src family members shows that exon 7B was captured by the ancestral fynT gene. The common ancestor of the srcA subfamily seems to possess an exon 7 similar to the fyn exon7T. This src-like ancestor duplicated and gave rise to fgr, src, yes and fyn. During evolution, fyn captured an exon 7B (Fig. 3⇑). This explains why only fyn possesses the two exons.
The specialized functions of fynT and fynB clearly remain to be established. In particular, the unique properties of fynB are important to be further studied in light of their potential pathological implications. Indeed, the fynB transcript accumulates in the human T cell leukemia virus type I-infected T cell line (13), and the nonpathogenic isoform of the scrapie prion protein signals through the Fyn kinase in a neuronal differentiation model (14). Also, unraveling the general mechanisms of co-option associated with changes of fyn gene expression patterning should provide important insights toward our understanding of evolutionary processes.
Materials and Methods
Plasmids and tissues
The pXM-fynT and pXM-fynB plasmid constructs were kindly provided by D. Davidson (Institut de Recherches Cliniques de Montréal, Québec, Canada). The X. laevis and T. marmorata tissues were kindly provided by L. Du Pasquier (Basel Institute for Immunology, Basel, Switzerland) and S. O’Regan (Laboratoire de Neurobiologie Cellulaire et Moleculaire, Gif-sur-Yvette, France), respectively.
Genomic DNA extraction
Genomic DNA was extracted from collected tissues using lysis buffer (50 mM Tris-HCl (pH 8.5) containing 100 mM EDTA, 100 mM NaCl, and 1% SDS) in the presence of 25 μg of proteinase K. Digestion was completed within several hours at 55°C under agitation. One volume of phenol/chloroform was added to the lysates followed by agitation for 15 min. After centrifugation at 13,000 rpm for 15 min (Biofuge; Heraeus, Osterode, Denmark), the aqueous phase was collected, and a second phenol/chloroform extraction was performed. Cold 100% ethanol was added to the collected aqueous phase. After centrifugation at 13,000 rpm for 30 min (Biofuge; Heraeus), the pellet was washed using 70% ethanol, resuspended in water overnight, and stored at −20°C.
PCR on genomic DNA
Amplification of fyn exon 7 was performed using the following primer pairs: X. laevis: 5′-gCTgCAggTCTgTgTTgCCg, 3′-ggCTTCAgTgTCTTTATggC; T. marmorata: 5′-gCTgCTggTCTCTgTTgTCg, 3′-CTTgACTgCAACTTTTgTggTTCC; X. helleri: 5′-TTCgTAAgCggACAgCggAgg, 3′-ACTTTggTggTgCCgTTCCACg.
The PCR products were size separated on a 1% agarose gel and visualized by UV light. The specific bands were cut and purified using NucleoSpin columns (NucleoSpin extract; Macherey-Nagel, Düren am Harz, Denmark), followed by sequencing.
RT-PCR
Total RNA was isolated from the spleens, thymuses, and brains of C57BL-6 mice and the thymus and brain of X. laevis, using the Trizol reagent, as recommended by the manufacturer (Life Technologies, Cergy Pontoise, France). Total RNA was isolated from the Trizol lysates by adding chloroform; centrifugation was followed by ethanol/isopropyl alcohol precipitation of the aqueous RNA solution. Total RNA was stored at −20°C. Two micrograms of RNA were reverse transcribed in the presence of 200 U of the Moloney murine leukemia virus reverse transcriptase, 40 U RNasin, 1 μg random primers, 0.5 mM dNTPs, and Moloney murine leukemia virus reverse transcriptase buffer in a total volume of 12 μl at 37°C for 1 h.
The cDNA was amplified using specific primers, and β2-microglobulin primers were used as a control for RNA extraction. The mouse fyn and β2-microglobulin specific primer sequences were as follows: fyn (5′), ATgggCTgTgTgCAATgTAAgg; fyn (3′), TTACAggTTTTCACCAggTTgg; fyn-exon7T (5′), CCCCACAAACTTCTggATTgg; fyn-exon7B (5′), gTCACAAAgggATgCCAAggC. β2-microglobulin: (5′), TggTgCTTgTCTCACTgACC; (3′), ATAgAAAgACCAgTCCTTgC.
The primers used to amplify X. laevis cDNA were as follows: fyn-exon6 (5′), TACgAAAgCTTgATAACgggg, and fyn-exon8 (3′), ggCTTCAgTgTCTTTATggC.
Phylogenetic analysis
Multiple alignments were constructed using Clustal X (15) with the default parameters on selected gene families. The phylogenetic tree was produced from the Clustal W multiple alignment output, using MEGA version 2.0, which is an updated version of MEGA (16). We used the neighbor-joining (NJ) reconstruction on a Poisson correction distance (with no assumption of rate constancy) with pairwise deletion and 500 bootstrap replicates. We rooted the tree on the midpoint.
Database searches
The Online Mendelian Inheritance in Man (http://www.ncbi.nlm.nih.gov/omim), National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/locusLink/), and Ensembl project (http://ensembl.ebi. ac.uk/) databases were used to search for the Src family gene sequences. The Fly base (http://flybase.bio.indiana.edu/) databases were consulted to search for SrcA orthologues in Drosophila. The Blast searches were performed by the Blast program at the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/Blast/Blast.cgi/).
Acknowledgments
We thank N. O’Regan and L. Du Pasquier for providing the T. marmorata DNA and X. laevis tissue materials, respectively, and L. Du Pasquier, Claude Mawas, M. Milinkovitch, and J. Nunes for critical review of the manuscript and helpful discussions.
Footnotes
-
↵1 C.P. is supported by a fellowship from the Agence Nationale de Recherche contre le Sida.
-
↵2 Address correspondence and reprint requests to Dr. Yves Collette, 27 boulevard Leï Roure, 13009 Marseille, France. E-mail address: collette{at}marseille.inserm.fr
-
↵3 Abbreviations used in this paper: AIS, adaptive immune system; BP, bootstrap proportion; NJ, neighbor-joining.
- Received December 18, 2001.
- Accepted January 22, 2002.
- Copyright © 2002 by The American Association of Immunologists