Evolution of Duplicated reggieGenes in Zebrafish and Goldfish

Edward Ma laga-Trillo, Ute Laessing, Dirk M. Lang, Axel Meyer, Claudia A.O. Stuermer

Department of Biology, University of Konstanz, 78457 Konstanz, Germany

Received: 24 January 2001 / Accepted: 27 July 2001

Abstract. Invertebrates, tetrapod vertebrates, and fishmight be expected to differ in their number of gene cop-ies, possibly due the occurrence of genome duplicationevents during animal evolution.Reggie(flotillin ) genescode for membrane-associated proteins involved ingrowth signaling in developing and regenerating axons.Until now, there appeared to be only tworeggiegenes infruitflies, mammals, and fish. The aim of this researchwas to search for additional copies ofreggie genes infishes, since a genome duplication might have increasedthe gene copy number in this group. We report the pres-ence of up to four distinctreggiegenes (tworeggie-1andtwo reggie-2 genes) in the genomes of zebrafish andgoldfish. Phylogenetic analyses show that the zebrafishand goldfish sequence pairs are orthologous, and that theadditional copies could have arisen through a genomeduplication in a common ancestor of bony fish. The pres-ence of novelreggiemRNAs in fish embryos indicatesthat the newly discovered gene copies are transcribedand possibly expressed in the developing and regenerat-ing nervous system. The intron/exon boundaries of thenew fish genes characterized here correspond with thoseof human genes, both in location and phase. An evolu-tionary scenario for the evolution ofreggie intron-exonstructure, where loss of introns appears to be a distinctivetrait in invertebratereggiegenes, is presented.

Key words: reggie— flotillin — Gene duplication —Microdomains — Axon growth — Axon regeneration —Carassius auratus— Danio rerio

Introduction

Reggie-1 and reggie-2 (homologous to flotillin-2 and flo-tillin-1, see below) are two closely related 48 kD intra-cellular and membrane-associated proteins whose encod-ing genes have been identified so far in goldfish (Schulteet al. 1997), rats (Lang et al. 1998), mice (Cho et al.1995; Bickel et al. 1997), humans (Schroeder et al.1994), and fruitflies (Galbiati et al. 1998). Reggies areco-expressed in many cell types and define specificplasma membrane microdomains (Lang et al. 1998) Incontrast to what has been suggested by others (Bickel etal. 1997; Volonte et al. 1999), they do not occur in ca-veolae—small plasma membrane microdomains charac-terized by the expression of caveolin—but are expressedclearly outside of these structures (Lang et al. 1998).Moreover, in neurons and T-lymphocytes—which do notpossess caveolae—reggie proteins occur in their ownspecific microdomains (Lang et al. 1998). In goldfishand rats, reggie-1 and -2 are upregulated during axongrowth and regeneration of retinal ganglion cells (RGC)(Schulte et al. 1997; Lang et al. 1998). Although theprecise nature of their role in the central nervous systemis not fully understood, two lines of evidence stronglysuggest that reggie-1 and -2 participate in molecular sig-naling across the plasmamembrane: 1) In cultured neu-rons, reggie-1 and -2 occur all along axons and are as-sociated with GPI-linked CAMs and fyn tyrosine kinase(Stuermer et al. 2001); 2) In activated T-lymphocytes,reggie-1 and -2 occur in micropatches along the plasmamembrane and are recruited into the T-cell receptor sig-naling complex (along with Thy-1 and GM-1) (Langet al. 1998; Stuermer et al. 2001).

It has long been suggested that two rounds of genomeduplication facilitated the appearance of novel mor-phologies early in vertebrate evolution by increasing the

Correspondence to:E. Malaga-Trillo; email: edward.malaga@uni-konstanz.de

J Mol Evol (2002) 54:235–245DOI: 10.1007/s00239-001-0005-1

© Springer-Verlag New York Inc. 2002

levels of genetic complexity (Ohno 1970; Ohno 1999).This hypothesis would explain why vertebrates can haveup to four copies of genes that are present as singlecopies in invertebrates. The recent discovery in the ze-brafish of extra copies of several gene families uncov-ered an additional genome duplication that probably tookplace in the common ancestor of all teleost fish (Amoreset al. 1998; Wittbrodt et al. 1998; Meyer and Schartl1999; Meyer and Ma´laga-Trillo 1999). The generation ofnew gene copies through genome duplications raisesquestions about how much functional redundancy can betolerated in genomes and whether duplicated copies arerandomly deleted, inactivated, or selected to perform anovel function (Malaga-Trillo and Meyer 2001). To an-swer these questions, it is necessary to identify as manyfamilies of duplicated genes as possible, and to examinethe evolution of their functions.

So far only one copy of eachreggie-(flotillin ) genehas been characterized in fruitflies, mammals, and gold-fish, and yet, if the additional fish genome-duplicationhypothesis is correct, then for any given gene family, fishare expected to have up to twice as many copies asmammals and up to eight times as many copies as in-vertebrates, although average gene copy ratios betweenthese groups cannot always be expected to remain wholenumbers because not all duplicated genes are maintainedfor a long evolutionary time. Some duplicated copiesmay be lost as pseudogenes or rendered unrecognizableas “junk DNA.” It is therefore possible that: 1) Only onefunctionalreggie-1or -2 gene was retained in vertebrategenomes, or that, 2) Additionalreggie copies exist (orexisted) in vertebrates but remain undiscovered. Thus,reggiegenes offer a case study for the evolution of ge-netic redundancy and the functional fate of duplicatedgene copies. Given their importance in the developmentof the vertebrate nervous system (Schulte et al. 1997;Lang et al. 1998), it is of interest to know how duplica-tion of these genes affected their function. Until now, theonly fish reggie genes known are those of the goldfishCarassius auratus(Lang et al. 1998), a model systemroutinely employed for the study of retinotectal develop-ment and regeneration of retinal axons (Gaze andSharma 1970). Despite their obvious biomedical interest,reggiegenes have not yet been characterized in the ze-brafish, Danio rerio, the system of choice for the mo-lecular characterization of genes important for diverseaspects of vertebrate development (Haffter et al. 1996).In order to evaluate new evidence for the fish genomeduplication hypothesis, and as a first step towards under-standing the functional evolution of redundant geneproducts in the nervous system, we set out to characterizethe reggiegenes of zebrafish and goldfish.

Materials and Methods

Nucleic acid isolation.Zebrafish and goldfish total genomic DNAswere prepared from fresh muscle tissue of homozygous individuals

using ATL Lysis Buffer (Qiagen, Hilden). We followed standard meth-ods (Sambrook et al. 1989) for the purification steps.

Synthesis of zebrafish 1st strand cDNA for RT-PCR.Total RNAwas isolated from two and three day old zebrafish embryos by theRNAzol B method (WAK Chemie, Bad Soden). First strand cDNA wasgenerated using Oligo-dT primers and the Superscript Reverse Tran-scriptase (Gibco BRL, Eggenstein) or the Ready-to-go T-primed First-strand Synthesis Kit (Amersham Pharmacia), as outlined by the manu-facturer.

Preparation of goldfish cDNA libraries.Two weeks after opticnerve transection, mRNA from goldfish retinae was prepared with theQuick Prep mRNA Purification Kit (Amersham Pharmacia), and re-verse transcribed using Superscript RT (Gibco BRL). After secondstrand synthesis and ligation toEcoRI adaptors (Amersham Pharma-cia), the cDNAs were cloned into Lambda ZapII (Stratagene, Heidel-berg). The titer of the cDNA library was amplified from 2.4 × 106

PFU/ml to 4.5 × 1010 PFU/ml.

Polymerase chain reaction (PCR). Reggie-1and-2 genes fragmentswere amplified with several pairs of degenerate primers located atvarious points along thereggie-1and -2 coding sequence (Table 1).Oneml of template was added to 19ml PCR reaction mix containing0.5 units of Red Taqt Polymerase (Sigma-Aldrich, Taufkirchen), 1×Red Taq Buffer, 800mM dNTPs, and 0.5mM primers. PCR reactionswere performed in GeneAmpt PCR System 9700 thermocyclers (PEApplied Biosystems, Weiterstadt) using the following conditions: Ini-tial denaturation for 2 min at 94°C, followed by 35–40 cycles of:denaturation at 94°C for 1 min, annealing at 50–56°C for 1 min, ex-tension at 72°C for 2 min; and a final extension step at 72°C for 10 min.

Cloning of PCR products.Amplification products were directlyused for cloning, or first excised from 2% agarose gels (Gibco BRL).After purification with the Qiaquick PCR Purification Kit (Qiagen), theDNAs were ligated into the Topo II vector (Invitrogen, Karlsruhe) andthe ligation mixes used to transformE. coli. Transformations wereplated onto LB-agar plates containing the appropriate amounts of am-picillin, X-gal, and IPTG (Sambrook et al. 1989). White recombinantbacteria were grown and tested by PCR for the presence of the correctinsert.

DNA sequencing.PCR products were purified with the aid of theQIAquick PCR purification kit (Qiagen), and plasmid templates fromcloned amplification products were isolated using the QIAprep SpinMiniprep Kit (Qiagen). Sequencing reactions were set up in 10mlvolumes containing up to 7ml of DNA, 1 ml of 10 mM primer and 2ml of Big Dye Terminator Cycle Sequencing Ready Reaction Mix(DNA Sequencing Kit, PE Applied Biosystems). The amount of tem-plate used for cycle sequencing and the thermal cycling profile werechosen according to the kit’s guidelines. Unincorporated dye termina-tors were removed by EtOH/NaOAc precipitation and the purified re-actions were loaded onto 0.2 mm-thick 5% Long Ranger™ gels and ranon an ABI 377 automated DNA sequencer.

DNA sequence analysis and drawing of dendrograms.DNA se-quences were edited and analyzed using PE Applied Biosystems soft-ware (Sequence Analysis and Sequence Navigator) as well as the GCGWisconsin Package Version 9.1 (The Genetics Computer Group 1997).Final sequences have been deposited in the GenBank database underthe accession codes AF315945, AF315946, AF315947, AF315948, andAF315949. Sequence alignments were made with DNA Star’sMegAlign, and phylogenetic analysis was performed using both dis-tance and maximum parsimony algorithms with Paup 4.0b4 (Swofford2000). Previously characterizedreggie sequences were downloaded

236

from Genbank. Accession numbers for theflotillin-1/reggie-2 se-quences of Drosophila, goldfish, mouse, rat, and human are AF044734(genomic: AE003810), U33556, NM_008027, U60976, and NM-005803 (genomic: NT_001520). Accession numbers for theflotillin-2/reggie-1sequences of Drosophila, goldfish, mouse, rat, and human areAF044916 (genomic: AE003497), L36867, NM_008028, AF023303,and NM_004475 (genomic: AC024267). It should be noted that severalreggie 1and 2 cDNA sequences have been reported in the rat. Thisdoes not imply that the rat has morereggiegenes. Based on comparisonwith human genomic sequences (not shown), the additional rat se-quences are rather the result of the isolation of heterogeneous nuclearRNAs and alternative splicing forms. For instance, in the case ofreggie1, a splice variant sequence starts at exon 4 and another one at exon 5.

For distance analysis, genetic distances were estimated with theKimura two-parameter method (Kimura 1980). Dendrograms were pre-pared using the neighbor-joining (NJ) tree constructing method (Saitouand Nei 1987) and reliability of the tree topologies was evaluatedthrough 1000 bootstrap replications (Felsenstein 1985).

Results and Discussion

Homology ofreggieand flotillin Genes

Thereggiegenes of goldfish and rat (Schulte et al. 1997;Lang et al. 1998) were reported independently of the

flotillin genes of fruitflies, humans, and mice (Schroederet al. 1994; Cho et al. 1995; Galbiati et al. 1998). Nev-ertheless, their highly similar DNA sequences and ex-pression patterns strongly suggest that these genes areclosely related and members of the same gene family(above references and unpublished Medline records). Weperformed phylogenetic analysis in order to clarify theirevolutionary relationships. The neighbor-joining tree inFig. 1 defines two major sequence homology groups, onecontaining reggie-1 and flotillin-2 sequences, and theother one includingreggie-2 and flotillin-1 sequences.The same topology was obtained with high bootstrapsupport when the maximum parsimony method was em-ployed (not shown). Therefore, the current nomenclatureis in need of revision: the names “reggie” and “flotillin ”designate homologous genes, but numbering of the loci(1 or 2) is misleading. Here, we will use the names“ reggie-1” and “reggie-2” to refer to reggie-1/flotillin-2and reggie-2/flotillin-1 groups, respectively The pres-ence of pairs of homologousreggie genes in mammalsand arthropods indicates thatreggie-1and -2 arose be-fore the evolutionary split between protostomes and deu-

Fig. 1. Neighbor-joining tree ofpreviously characterizedreggie(Reg) andflotillin (Flo) and DNAsequences. An alignment offull-length coding sequences wasused as a basis for phylogeneticreconstruction. The distance matrixwas generated using the Kimuratwo-parameter method. Thehorizontal length of each branch isproportional to the estimated geneticdistance (number of substitutions)between taxa. Bootstrap values(1000 replications) are indicated oneach branch.

237

terostomes. On average, the intraspecific aminoacid se-quence similarity betweenreggie-1 and -2 genes isapproximately 52.5%, while the similarity between ver-tebrate and invertebratereggiesof the -1 or -2 groups isapproximately 62.8%. The length of thereggie codingregion is 429 aminoacids (aa) and appears to be con-served during evolution, although minor length variationcan be observed at the 38 end of somereggie-2genes. Inall pairwise comparisons, substitutions are randomly dis-tributed along the gene and no mutational hotspots orhighly conserved domains are apparent.

Isolation of NovelreggieSequences From Fish

The goldfish and zebrafish sequences reported here wereobtained by PCR amplification of genomic and cDNAtemplates using several combinations of degenerateprimers. An alignment of all previously availablereg-gie-1and-2 DNA sequences (not shown) was used as thebasis for primer design. In order to increase the speci-ficity of our amplifications towards goldfish/zebrafishsequences, two of the primer pairs (GF-Reg1F/R andGF-Reg2F/R) were constructed specifically for goldfishreggie-1and2 sequences, and used in combination withthe degenerate ones. At the time we started with theseexperiments, no information about the intron-exon orga-nization ofreggiegenes was available. This presented apotential problem for the amplification of genomic tar-gets. Therefore, 20 forward and reverse primers at 13

Table 1. PCR Primers used in this study

Primer ID Sequence

1. Reg-19-43F ACTKKYGGMCCMAAYGARGCVMTSG2. Reg-193-215F GGDGTVSCYWTYWCYGTSACHGG3. Reg-193-215R CCDGTSACRGWRAWRGSBACHCC4. Reg-271-288F SAGMWGTTYCTGGGRAAG5. Reg-271-288R CTTYCCCAGRAACWKCTS6. Reg-360-384F GACKGTGGAGSAGATYTAYMAGGAC7. Reg-271-288R GTCCTKRTARATCTSCTCCACMGTC8. Reg-568-590F GRGATGCNGGGATYMGGGARGC9. Reg-568-590R GCYTCCCKRATCCCNGCATCYC

10. Reg-720-743F SMTGGCCTATCAGCTMCAGGYRGC11. Reg-720-743R GCYRCCTGKAGCTGATAGGCCAKS12. Reg-862-885F GTSMRSMRDCCWGCBGADGCMGAG13. Reg-862-885R CTCKGCHTCVGCWGGHYKSYKSAC14. Reg-1015-1039F GAGGCYGAGCRRATGRSSMWGAAGG15. Reg-1015-1039R CCTTCWKSSYCATYYGCTCRGCCTC16. Reg-1098-1119R GAYCTYYYCWGCMAYCTKGGGC17. GF-Reg1F GARGTIGCIGCICCIGAYGT18. GF-Reg1R GCYTCIGCYTCICCDATYTT19. GF-Reg2F AAYGARGCIATGGTIGTITC20. GF-Reg2R TCYTCDATYTGYTGYTTIGTYTT21. Dare-Reg1aXF ACAGGGATCAGTTTGCCAAG22. Dare-Reg1aXR CTCTCTCTGCTTCAGCCACA23. Dare-Re2aXF CGATGGTGGTGTCAGGTAAG24. Dare-Reg2aXR GGACCCCATGACGTGTGTA25. Dare-Reg2bXF GCTCCTCCTCTCATGATTGC26. Dare-Reg2bXR GCCTTCTTGGTGTTGACCTC

Numbering corresponds to that of Figure 2. Primer orientation is speci-fied by “F” or “R” at the end of the primer’s name.

Fig. 2. PCR strategy for the amplification ofreggie genes in ze-brafish and goldfish. Roman numbers indicate the arbitrary amplifica-tion targets defined by the position of our degenerate primers. Primersare represented by arrows and numbered according to Table 1. Primer

orientation is indicated by the arrows. The final length of the amplifi-cation products obtained is shown for each of the novel genes identifiedin this study.

238

different locations in the least variable segments of thereggie-1and -2 coding region were designed (Table 1).The positions of the PCR primers divide most of thereggie-1and -2 consensus sequence into eight amplifi-cation targets (Fig. 2). All primer combinations encom-passing one, two, or up to three of these segments weretested. The efficiency of the amplifications varied largelydepending on the primer pair and the type of templateused. The strongest and cleanest amplification productsof expected sizes were cloned and sequenced. Primercombinations 271-288F/360-384R, 360-384F/568-590R,and GF-Reg1F/568-590R amplifiedreggie-1 products;primer combinations GF-Reg2F/193-215R, 360-384F/GF-Reg2R, 720-742F/862-885R, and GF-Reg2F/GF-Reg2R amplifiedreggie-2 products. Alignment of theamplified sequences identified four and three differentreggiegenes in goldfish and zebrafish, respectively. Twoof the goldfish sequences had been previously reported(Schulte et al. 1997), but the other two represent novelreggie-1and-2 genes. The three zebrafish sequences arenovel, one of them homologous toreggie-1and the othertwo homologous toreggie-2 (Fig. 4). These new se-quences represent distinct loci and not alleles becausethey were isolated from single homozygous fish. Wedesignated the four fish genes,reggie-1a, -1b, -2a,and-2b,based on their sequence similarity to each other andto reggie-1 genes from other species (Table 2). We

named the newly foundreggie-1from goldfishreggie-1band renamed the previously cloned goldfishreggie-1(Schulte et al. 1997) asreggie-1a.Failure to amplifyreggie-1bfrom the zebrafish could be due to an inde-pendent gene loss in the zebrafish genome or to technicaldifficulties such as inefficient primer annealing duringPCR amplification.

New internal, zebrafish-specific primers were de-signed (Table 1, Fig. 2) and used on zebrafish cDNA incombination with some of the already mentioned degen-erate primers, in order to extend the amount of codingsequence for the transcribed zebrafish genes. As a result,a total of 1188, 681, and 810 bp of coding sequence wereobtained forreggie-1a, reggie-2a,andreggie-2b,respec-tively (Fig. 2). In addition, thereggie-1and-2 introns ofzebrafish and goldfish contribute 243 and 299 bp ofDNA sequence, respectively.

The new zebrafish and goldfish sequences werealigned and compared to otherreggiegenes, in order toestablish their evolutionary history. Percent aminoacidsimilarity between all known representatives of thereg-gie-1and-2 groups are given in Table 2. For both groupsof sequences, it can be observed that “a” or “ b” genes aremore similar between goldfish and zebrafish than to eachother within either one species. For instance,reggie-2agenes of zebrafish and goldfish are 96.6% similar, whilereggie-2aand-2b in zebrafish or goldfish are only 88.3

Fig. 3. Two phylogenetic scenariosfor the tree topology of duplicatedgenes X in hypothetical species 1 and2, depending on the relative timing oftheir duplication.A: A shared geneduplication before speciation results intrans-specific groupings.B:Independent gene duplications (afterspeciation) result in grouping of genesaccording to species.

239

and 83.5% similar. Thus, the origin of “a” and “b” genesappears to predate the divergence of zebrafish and gold-fish.

The Duplication of FishreggieGenes inTeleost Genomes

The presence of additionalreggiegenes in zebrafish andgoldfish, relative to tetrapods, could be explained by agene/genome duplication in the common ancestor of

these fish. Alternatively, the novel genes could havearisen through independent gene/genome duplicationevents in each of the two species. An ancestral duplica-tion is more parsimonious than two independent dupli-cation events in descending lineages. Phylogeneticanalysis of allreggie genes was undertaken to decidebetween these two scenarios, based on the resulting treetopologies. If—as expected—the novelreggie copiesarose in a teleost ancestor, homologousreggiesequencesshould cluster trans-specifically (Fig. 3A); if they arose

Fig. 4. Distance (Kimura’stwo-parameter) and maximum parsimonyconsensus tree ofreggie-1(A) andreggie-2(B) DNA sequences. Fruitfly,mouse, and human sequences have beenrenamed from their original “flotillin ”designations to the “reggie” names used inthis study (see Results). Trees wereconstructed using the neighbor-joiningalgorithm. The horizontal lengths of eachbranch are proportional to the estimatedgenetic distance (number of substitutions)between taxa. Bootstrap values (1000replications) obtained by distance andparsimony methods are indicated on eachbranch, in that order and separated by a“/”.

240

independently in zebrafish and goldfish, then their se-quences would group by species (Fig. 3B). Thus, the treetopology is indicative of the timing of the duplication.Distance and Maximum Parsimony methods using fruit-fly sequences as outgroup, agree on the same tree to-pologies (Fig. 4A and B). Bothreggie-1and-2 trees aresupported with high bootstrap values and clearly dividefish from mammalian sequences. Within the fish groups,sequences do not cluster according to species, but forminstead clusters of “a” or “ b” semiorthologous genes(Holland 1999) across species. Therefore, the existenceof four reggie genes in zebrafish and goldfish is likelythe result of a duplication event in a common ancestor ofmodern fish, after its separation from the lineage leadingto tetrapods. Based only on our data, we cannot decidewhether this event was a gene or an entire genome du-plication. However, a whole-genome duplication in thecommon ancestor of all teleosts is strongly supported bydata onHoxand other gene families (Amores et al. 1998;Postlethwait et al. 1998; Naruse et al. 2000; Ma´laga-Trillo and Meyer 2001; Ma´laga-Trillo et al. unpublisheddata). Although our data are obviously not sufficient toconfirm the fish genome duplication hypothesis, they doadd experimental evidence in its favor.

In the absence of mapping information, our results donot formally rule out the possibility thatreggie genesmay have duplicated as part of a partial duplication in-volving only Hox chromosomes (as opposed to a wholegenome duplication), However, this can be demonstratedon the basis of genomic data from other organisms show-

ing conserved mapping ofreggieandHox genes to dif-ferent syntenic groups. In humans,reggie-2(Flotillin-1 )maps to 6p2l.3, within the Major HistocompatibilityComplex (Mhc) (MHC Sequencing Consortium 1999).The humanMhc is one of four extense paralogous re-gions that map to 6p21.3, 19p13.1, 9q33-q34, and 1q21-q25, independently ofHox clusters (chromosomes 2, 7,12, and 17). The affiliation ofReggie-2within theMhc isan ancient synteny conserved also in the zebrafish (Mi-chalova et al. 2000). The ancestral linkage ofreggie-2genes to theMhcchromosomes, suggests that duplicationevents in vertebrate history involved more than justHoxchromosomes, and allows us to extend this assumption tothe fish specific duplications. HumanReggie-1(Flotillin-2) does map to aHox chromosome (17q12), but verydistant from theHox Bcluster (17q21). This associationis not representative of an ancestral genetic linkage be-tweenReggieand Hox clusters, because in the fruitfly,neitherReggie-1nor -2 (mapping to chromosomes 1 and2, respectively) are linked to the singleHox complex (inchromosome 3). Thus, the presence ofReggie-1in hu-man chromosome 17 is rather likely to be the conse-quence of an independent translocation event.

The Division Teleostei (bony fish) is made up of 38Orders. One of these Orders—the Cypriniformes—includes the Cyprinidae, the largest family of freshwaterfishes to which zebrafish and goldfish also belong (Nel-son 1994). As these two genera represent only a fractionof the genetic and morphological diversity observed inmodern teleosts, it will be necessary to identify ortholo-

Table 2. (A) Percent similarity between knownReggie-1aminoacid sequences

Drme-1 Mumu-1 Rano-1 Hosa-1 Dare-1a Caau-1a Caau-1b

Drme-1 — 59.7 61.3 61.3 63.6 62.2 64Mumu-1 — — 83 80 68 65 74Rano-1 — — — 89 73 81 73Hosa-1 — — — — 74 81 73Dare-1a — — — — — 85 78Caau-1a — — — — — — 88Caau-1b — — — — — — —

Species are designated by the first two letters of the genus and species name: Drme4 Drosophila melanogaster, Rano4 Rattus norvegicus, Hosa4 Homo sapiens, Dare4 Danio rerio, Caau4 Carassius auratus.

(B) Percent similarity between knownReggie-2aminoacid sequences

Drme-2 Mumu-2 Rano-2 Hosa-2 Dare-2a Dare-2b Caau-2a Caau-2b

Drme-2 — 63.3 63.2 63.4 65.7 63.6 62 64.2Mumu-2 — — 99.5 97.9 88.7 78.3 79.2 79.5Rano-2 — — — 97.7 88.7 78.3 79.4 79.5Hosa-2 — — — — 89.1 79 79.4 80.2Dare-2a — — — — — 83 96.1 84.3Dare-2b — — — — — — 82.2 93.4Caau-2a — — — — — — — 81.9Caau-2b — — — — — — — —

Species are designated by the first two letters of the genus and species name: Drme4 Drosophila melanogaster, Mumu4 Mus musculus; Rano4 Rattus norvegicus, Hosa4 Homo sapiens, Dare4 Danio rerio, Caau4 Carassius auratus.

241

Fig

.5.

Hyp

othe

tical

scen

ario

for

the

evol

utio

nof

the

intr

on-e

xon

orga

niza

tion

ofre

gg

iege

nes,

base

don

the

curr

ent

conf

igur

atio

nof

hum

anan

dfr

uitfl

yge

nes.

The

hum

anan

dfr

uitfl

yge

nest

ruct

ures

wer

eob

tain

edby

alig

ning

the

indi

vidu

alreg

gie

-1an

d-2

part

ials

truc

ture

s.E

xons

are

repr

esen

ted

bybl

uebo

xes;

red

boxe

s4in

tron

loss

essh

ared

byhu

man

and

frui

tfly

gene

s;bl

ack

boxe

s4

conc

erte

din

tron

loss

esin

frui

tfly

gene

s;or

ange

boxe

s4fr

uitfl

yre

gg

ie-1

-spe

cific

intr

onlo

sses

;gre

enbo

xes

4fr

uitfl

yre

gg

ie-

2-sp

ecifi

cin

tron

loss

es.T

hefis

hin

tron

sid

entif

ied

inth

epr

esen

tstu

dyar

ein

dica

ted

byve

rtic

alar

row

son

the

corr

espo

ndin

glo

catio

nsin

the

hum

ange

nes.

Exo

n1

isno

n-co

nser

ved;

itis

only

show

nfo

rill

ustr

ativ

epu

rpos

es.

Exo

n2

isde

lete

din

frui

tflyreg

gie

-1(b

oxm

arke

dw

itha

cros

s).

242

Fig

.6.

Am

inoa

cid

alig

nmen

tofh

uman

and

frui

tflyre

gg

ie-1

and

-2co

ding

sequ

ence

s,in

clud

ing

intr

onlo

catio

nsan

dph

ases

.In

thos

eca

ses

whe

rein

tron

phas

eis

“1”

or“2

”,th

enu

cleo

tide

sequ

ence

ofth

ein

terr

upte

dco

dons

are

show

nfla

nkin

gth

ein

tron

.Das

hes

repr

esen

tind

els.

The

star

tcod

onof

reg

gie

gene

sis

loca

ted

inex

on2.

Sin

ceex

on1

isno

ncod

ing

and

high

lyva

riabl

e,th

eex

on1/

intr

on1

and

intr

on1/

exon

2bo

unda

ries

wer

ele

ftou

tof

our

anal

ysis

.E

xon

2is

dele

ted

infr

uitfl

yre

gg

ie-1

(see

Fig

.5)

.

243

gous reggie genes in other teleost families to confirmthat the duplication of these genes occurred in a bonyfish ancestor.

Evolution of the Intron-Exon Organization ofreggieGenes

All goldfish and zebrafish sequences reported here wereamplified from cDNA. In addition, goldfishreggie-1band-2b,as well as zebrafishreggie-2acould be partiallyamplified from genomic DNA, uncovering the existenceof at least five introns in the upstream half of these genes:A 98 and a 145 bp intron inreggie-1;and an 84, a 92,and a 123 bp inreggie-2(genomic sequences from thedownstream half of fishreggiegenes could not be am-plified due to technical limitations). The degree of con-servation in length, and location and phase between thesefish introns and those of other organisms could not beimmediately evaluated because the intron-exon struc-tures of otherreggie genes were not previously de-scribed. Nevertheless, recent data from genome sequenc-ing projects (records unpublished but available at http://www.ncbi.nlm.nih.gov/Genomes/) provide extensiveinformation on the intron-exon structure of human andfly reggie genes. We used this information to find outwhether the partial gene structure of the fish genes ob-tained here corresponds with that of human and fruitflygenes, as well as to learn more about the mechanismsdriving their intron evolution and gene structure.

We compared the cDNA and genomic sequences ofhuman and fruitflyreggie-1and-2 genes (see Materialsand Methods for accession numbers) and inferred thesize and location of introns relative to the reading frame.In contrast to their high degree of DNA sequence con-servation, human and fruitflyreggiegenes differ greatlyin their intron-exon organization, fruitfly genes havingfewer introns than their human counterparts (Fig. 5).Alignment of fruitfly and human coding regions showsgeneral conservation of intron-exon boundaries and in-tron phases (Fig. 6). Except for two minor sliding events(at intron 4 and possibly at intron 10), the locations of allfruitfly introns correspond to those of the human ones.The use of human and fruitfly sequences for the com-parison allowed us to identify shared (ancestral) intronpositions and phases for nine out of the 11 introns ana-lyzed. The difference in the number of introns betweenthe two organisms is therefore likely to be the result ofloss of introns in the lineage leading to fruitflies ratherthan the gain of introns in the lineage leading to humans.The latter scenario is non-parsimonious because it wouldrequire the repeated and independent de novo formationof humanreggie-1 and -2 introns at exactly the samepositions in eight out of 11 cases.

The fish introns reported here are generally shorterthan their human fruitfly counterparts; they correspondto introns 5 and 6 inreggie-1,and to introns 2, 3 and 8

in reggie-2(Fig. 5, indicated by vertical arrows). In allcases, intron locations and phases are conserved betweenhuman and fish introns (Fig. 6). The presence of introns6 and 8 in fishreggie-1and -2, respectively, is sharedwith human but not with fruitfly genes, supporting thenotion that these introns are ancient and that they werelost in the lineage leading to invertebrates.

Using introns as cladistic characters, we attempted toreconstruct the evolutionary patterns of the intron loss inreggiegenes, based on the intron-exon structure of hu-man and fruitfly genes. We assume that the origin ofreggie-1 and -2 predate the split between protostomesand deuterostomes (see phylogeny in Fig. 1), and thatindependent loss of introns is more parsimonious thantheir independent gain.Reggie-1and-2 must have orig-inated in a common ancestor of protostomes and deu-terostomes, from the duplication of aproto-reggie-1/2gene with 13 exons; a subsequent loss of intron 10 oc-curred in thereggie-2gene of this hypothetical ancestor,as this loss is shared by human and fruitflyreggie-2genes. After the split between protostomes and deutero-stomes, the lineage leading to the fruitfly underwent theconcerted loss of introns 7, 8, and 9 in bothreggie-1and-2 genes, as well as the loss of introns 3, 6, and 12 inreggie-1and 4, 5, and 9 inreggie-2(Fig. 5). In addition,fruitfly reggie-1appears to have lost exon 2 (Fig. 6).

Much remains to be learned about the functional fateof the novel genes identified in this study. Our surveyuncovered the presence of duplicatedreggie transcriptsin two- to three-day-old zebrafish embryos. We are cur-rently examining their developmental expression pat-terns in zebrafish, and assessing their role in axongrowth/regeneration and T-cell activation. A particulararea of interest remains the question whether all fourreggiegenes of fish are expressed simultaneously in onemicrodomain, or rather in a combinatorial manner, assubsets that define distinct microdomains with differentsignaling properties, resulting in the development ofnovel regulatory interactions absent in mammals. Thepossible effects of these changes on axon growth/regeneration could hold much awaited answers about thedeveloping nervous system of vertebrates and its inabil-ity to regenerate neurons at the adult state.

Acknowledgments. This work was supported by grants from the Min-isterium fur Wissenschaft und Kunst of Baden-Wu¨rttemberg to C.S.and A.M. and from the DFG to A.M., and is the result of an ongoingcollaboration between the Stuermer and the Meyer laboratories.

References

Amores A, Force A, Yan YL, Joly L, Amemiya C, Fritz A, Ho RK,Langeland J, Prince V, Wang YL, Westerfield M, Ekker M,Postlethwait JH (1998) Zebrafishhox clusters and vertebrate ge-nome evolution. Science 282:1711–1714

Bickel PE, Scherer PE, Schnitzer JE, Oh P, Lisanti MP, Lodish HF(1997) Flotillin and epidermal surface antigen define a new family

244

of caveolae-associated integral membrane proteins. J Biol Chem272:13793–13802

Cho YJ, Chema D, Moskow JJ, Cho M, Schroeder WT, Overbeek P,Buchberg AM, Duvic M (1995) Epidermal surface antigen(MS17S1) is highly conserved between mouse and human. Genom-ics 27:251–258

Felsenstein J (1985) Confidence limits on phylogenies: an approachusing the bootstrap. Evolution 39:783–791

Galbiati F, Volonte D, Goltz JS, Steele Z, Sen J, Jurcsak J, Stein D,Stevens L, Lisanti MP (1998) Identification, sequence and devel-opmental expression of invertebrate flotillins fromDrosophila me-lanogaster.Gene 210:229–237

Gaze RM, Sharma SC (1970) Axial differences in the reinnervation ofthe goldfish optic tectum by regenerating optic nerve fibres. ExpBrain Res 10:171–181

The Genetics Computer Group I (1997) Wisconsin Package 9.1. Madi-son, WI

Haffter P, Granato M, Brand M, Mullins MC, Hammerschmidt M,Kane DA, Odenthal J, van Eeden FJ, Jiang YJ, Heisenberg CP,Kelsh RN, Furutani-Seiki M, Vogelsang E, Beuchle D, Schach U,Fabian C, Nusslein-Volhard C (1996) The identification of geneswith unique and essential functions in the development of the ze-brafish,Danio rerio. Development 123:1–36

Holland PW (1999) The effect of gene duplication on homology. No-vartis Found Symp 222:226–236

Kimura M (1980) A simple method for estimating evolutionary rates ofbase substitutions through comparative studies of nucleotide se-quences. J Mol Evol 16:111–120

Lang DM, Lommel S, Jung M, Ankerhold R, Petrausch B, Laessing U,Wiechers MF, Plattner H, Stuermer CA (1998) Identification ofreggie-1 and reggie-2 as plasmamembrane-associated proteinswhich cocluster with activated GPI-anchored cell adhesion mol-ecules in non-caveolar micropatches in neurons. J Neurobiol 37:502–523

Malaga-Trillo E, Meyer A (2001) Genome duplications and acceleratedevolution of Hox genes and cluster architecture in teleost fishes.Amer Zool, in press

Meyer A, Malaga-Trillo E (1999) Vertebrate genomics: more fishytales aboutHox genes. Curr Biol 9:R210–

Meyer A, Schartl M (1999) Gene and genome duplications in verte-brates: the one-to-four (-to-eight in fish) rule and the evolution ofnovel gene functions. Curr Opin Cell Biol 11:699–704

Michalova V, Murray BW, Sultmann H, Klein J (2000) A contig map

of the Mhc class I genomic region in the zebrafish reveals ancientsynteny. J Immunol 164:5296–5305

Naruse K, Fukamachi S, Mitani H, Kondo M, Matsuoka T, Kondo S,Hanamura N, Morita Y, Hasegawa K, Nishigaki R, Shimada A,Wada H, Kusakabe T, Suzuki N, Kinoshita M, Kanamori A, TeradoT, Kimura H, Nonaka M, Shima A (2000) A detailed linkage mapof medaka,Oryzias latipes.Comparative genomics and genomeevolution. Genetics 154:1773–1784

Nelson JS (1994) Fishes of the world, 3rd ed. J. Wiley, New YorkOhno S (1970) Evolution by gene duplication. Springer-Verlag, Berlin,

New YorkOhno S (1999) The one-to-four rule and paralogues of sex-determining

genes. Cell Mol Life Sci 55:824–830Postlethwait JH, Yan YL, Gates MA, Horne S, Amores A, Brownlie A,

Donovan A, Egan ES, Force A, Gong Z, Goutel C, Fritz A, KelshR, Knapik E, Liao E, Paw B, Ransom D, Singer A, Thomson M,Abduljabbar TS, Yelick P, Beier D, Joly JS, Larhammar D, Rosa F,et al. (1998) Vertebrate genome evolution and the zebrafish genemap. Nat Genet 18:345–349

Saitou N, Nei M (1987) The neighbor-joining method: a new methodfor reconstructing phylogenetic trees. Mol Biol Evol 4:406–425

Sambrook J, Maniatis T, Fritsch EF (1989) Molecular cloning: a labo-ratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, NewYork

Schroeder WT, Stewart-Galetka S, Mandavilli S, Parry DA, GoldsmithL, Duvic M (1994) Cloning and characterization of a novel epider-mal cell surface antigen (ESA). J Biol Chem 269:19983–19991

Schulte T, Paschke KA, Laessing U, Lottspeich F, Stuermer CA (1997)Reggie-1 and reggie-2, two cell surface proteins expressed by reti-nal ganglion cells during axon regeneration. Development 124:577–587

Stuermer CAO, Lang DM, Kirsch F, Deininger S, Wiechers MF,Plattner H (2001) GPI-anchored proteins and fyn kinase assemblein non-caveolar plasmamembrane microdomains defined by reg-gie-1 and reggie-2. Mol Biol Cell (in press)

Swofford DL (2000) PAUP*. Phylogenetic analysis using parsimony(*and other methods). Sinauer Associates, Sunderland, MA

Volonte D, Galbiati F, Li S, Nishiyama K, Okamoto T, Lisanti MP(1999) Flotillins/cavatellins are differentially expressed in cells andtissues and form a hetero-oligomeric complex with caveolins invivo. Characterization and epitope-mapping of a novel flotillin-1monoclonal antibody probe. J Biol Chem 274:12702–12709

Wittbrodt J, Meyer A, Schartl M (1998) More genes in fish? Bioessays20:511–515

245