Evolution of Chordate Actin Genes: Evidence from Genomic Organization and Amino Acid Sequences

The origin and evolutionary relationship of actin isoforms was investigated in chordates by isolating and characterizing two new ascidian cytoplasmic and muscle actin genes. The exon–intron organization and sequences of these genes were compared with those of other invertebrate and vertebrate actin genes. The gene HrCA1 encodes a cytoplasmic (nonmuscle)-type actin, whereas the MocuMA2 gene encodes an adult muscle-type actin. Our analysis of these genes showed that intron positions are conserved among the deuterostome actin genes. This suggests that actin gene families evolved from a single actin gene in the ancestral deuterostome. Sequence comparisons and molecular phylogenetic analyses also suggested a close relationship between the ascidian and vertebrate actin isoforms. It was also found that there are two distinct lineages of muscle actin isoforms in ascidians: the larval muscle and adult body-wall isoforms. The four muscle isoforms in vertebrates show a closer relationship to each other than to the ascidian muscle isoforms. Similarly, the two cytoplasmic isoforms in vertebrates show a closer relationship to each other than to the ascidian and echinoderm cytoplasmic isoforms. In contrast, the two types of ascidian muscle actin diverge from each other. The close relationship between the ascidian larval muscle actin and the vertebrate muscle isoforms was supported by both neighbor-joining and maximum parsimony analyses. These results suggest that the chordate ancestor had at least two muscle actin isoforms and that the vertebrate actin isoforms evolved after the separation of the vertebrates and urochordates. Received: 20 June 1996 / Accepted: 16 October 1996     

Structural Comparisons of Muscle and Nonmuscle Actins Give Insights Into the Evolution of Their Functional Differences

Actin is a highly conserved protein although many isoforms exist. In vertebrates and insects the different actin isoforms can be grouped by their amino acid sequence and tissue-specific gene expression into muscle and nonmuscle actins, suggesting that the different actins may have a functional significance. We ask here whether atomic models for G- and F-actins may help to explain this functional diversity. Using a molecular graphics program we have mapped the few amino acids that differ between isoactins. A small number of residues specific for muscle actins are buried in internal positions and some present a remarkable organization. Within the molecule, the replacements observed between muscle and nonmuscle actins are often accompanied by compensatory changes. The others are dispersed on the protein surface, except for a cluster located at the N-terminus which protrudes outward. Only a few of these residues specific for muscle actins are present in known ligand binding sites except the N-terminus, which has a sequence specific for each isoactin and is directly implicated in the binding to myosin. When we simulated the replacements of side chains of residues specific for muscle actins to those specific for nonmuscle actins, the N-terminus appears to be less compact and more flexible in nonmuscle actins. This would represent the first conformational grounds for proposing that muscle and nonmuscle actins may be functionally distinguishable. The rest of the molecule is very similar or identical in all the actins, except for a possible higher internal flexibility in muscle actins. We propose that muscle actin genes have evolved from genes of nonmuscle actins by substitutions leading to some conformational changes in the protruding N-terminus and the internal dynamics of the main body of the protein. Received: 15 March 1996 / Accepted: 14 July 1996     

Isolation and Characterization of Five Actin cDNAs from the Cestode Diphyllobothrium dendriticum: A Phylogenetic Study of the Multigene Family

Five cDNAs (pDidact2–pDidact6), representing different actin genes, were isolated from a Diphyllobothrium dendriticum cDNA library, and the DNA as well as the putative amino acid sequences were determined. The corresponding Didact2 and Didact4 genes code for peptides 376 amino acids long, with molecular weights 41,772 and 41,744 Da, respectively, while the deduced Didact3 protein is 377 amino acids long and weighs 41,912 Da. The pDidact5 and -6 cDNAs lack nucleotides corresponding to three to six amino acids at the amino-terminus. Two of the five cDNAs contain the conventional AATAAA as the putative polyadenylation signal, one has the common variant ATTAAA, whereas the hexanucleotide AATAGA is found 15 and 18 nucleotides, respectively, upstream of the poly(A) site in two of the cDNAs. Phylogenetic studies including 102 actin protein sequences revealed that there are at least four different types of cestode actins. In this study three of these types were found to be expressed in the adult D. dendriticum tapeworm. Structurally the cestode actin groupings differ from each other to an extent seen only among the metazoan actins between the vertebrate muscle and cytoplasmic isoforms. In the phylogenetic trees constructed, cestode actins were seen to map to two different regions, one on the border of the metazoan actins and the other within this group. It is, however, difficult to say whether the cestode actins branched off early in the metazoan evolution or if this position in the phylogenetic tree only reflects upon differences in evolutionary rate. Received: 19 June 1996 / Accepted: 20 August 1996     

Evolution of cis-Acting Elements in 5′ Flanking Regions of Vertebrate Actin Genes

Regulation of the vertebrate actin multigene family involves the recognition of various regulatory sequences (cis-acting elements) that specify the distinct tissue type and developmental program of expression for each actin paralogue, which implies that the distribution of cis-acting elements may be unique for each paralogue gene. To elucidate the evolution of these unique distribution patterns, we improved a method to scan for cis-acting elements in the 5′ flanking regulatory region of genes and used it to analyze five cis-acting elements (SRE, MyoD binding site, Elk-1 binding site, positive and negative YY1 binding sites) of six actin paralogue genes (β and γ cytoplasmic actins, α and γ smooth muscle actins, and α skeletal and α cardiac actins) among various vertebrates. It was shown that although an element(s) may exist in all paralogue genes of the same species, its numbers, compositions, and distribution patterns or even sequences vary remarkably among paralogues, which contributes to their different tissue- and developmental-specific expression. However, each pair of coexpressed paralogues has some certain similarity in distribution patterns. Furthermore, among various orthologues of actin genes derived from diverse vertebrates, the sequences, numbers, and distribution patterns of these cis-acting elements are highly conserved or even identical in the long run of phylogeny of vertebrates. Taken together, the results described above strongly indicate that not only the structures of actins but also their expression patterns are essential in both the phylogeny and the physiology of vertebrates. The distribution patterns of cis-acting elements of various actin genes can be regarded as indicators of both horizontal (paralogous) and vertical (orthologous) evolution of actins. Received: 1 March 1999 / Accepted: 6 August 1999     

Evidence for Gene Conversion Between Tandemly Duplicated Cytoplasmic Actin Genes of Helicoverpa armigera (Lepidoptera:Noctuidae)

Tandemly duplicated actin genes have been isolated from a Helicoverpa armigera genomic library. Sequence comparisons with actin genes from other species suggest they encode cytoplasmic actins, being most closely related to the Bombyx mori A3 actin gene. The duplicated H. armigera actin genes, termed A3a and A3b, share 98.3% nucleotide sequence identity over their entire putative coding region. Analysis of the distribution of nucleotide differences shows the first 763 bp are identical between the two coding regions, with the 18 nucleotide changes occurring in the remaining 366 bp. This observation suggests a gene conversion event has taken place between the duplicated H. armigera A3a and A3b actin genes. Translation of the open-reading frames indicates the products of these genes are identical, apart from a single amino acid difference at codon 273. Polymerase chain reaction and northern blot analysis have shown both H. armigera A3a and A3b genes are expressed during pupal development and in the brain of newly eclosed adults. A region 5′ of the H. armigera A3a actin gene start codon has been identified which contains regulatory sequences commonly found in the promoter region of actin genes, including TATA, CAAT, and CArG motifs. Received: 10 January 1996 / Accepted: 12 March 1996     

Gene Conversions May Obscure Actin Gene Family Relationships

Phylogenetic hypotheses of muscle actin evolution are significantly different when a sea urchin is used as a representative echinoderm than when a sea star is used. While sea urchin muscle actins support an echinoderm–chordate sister relationship, sea star sequences suggest that echinoderm muscle actins are convergent with chordate muscle actins. Our results suggest that gene conversion in the sea star muscle actin may be responsible for these discordant results. Received: 19 July 1999 / Accepted: 1 October 1999     

The Peperomia Mitochondrial coxI Group I Intron: Timing of Horizontal Transfer and Subsequent Evolution of the Intron

The Peperomia polybotrya coxI gene intron is the only currently reported group I intron in a vascular plant mitochondrial genome and it likely originated by horizontal transfer from a fungal donor. We provide a clearer picture of the horizontal transfer and a portrayal of the evolution of the group I intron since it was gained by the Peperomia mitochondrial genome. The intron was transferred recently in terms of plant evolution, being restricted to the single genus Peperomia among the order Piperales. Additional support is presented for the suggestion that a recombination/repair mechanism was used by the intron for integration into the Peperomia mitochondrial genome, as a perfect 1:1 correspondence exists between the intron's presence in a species and the presence of divergent nucleotide markers flanking the intron insertion site. Sequencing of coxI introns from additional Peperomia species revealed that several mutations have occurred in the intron since the horizontal transfer, but sequence alterations have not caused frameshifts or created stop codons in the intronic open reading frame. In addition, two coxI pseudogenes in Peperomia cubensis were discovered that lack a large region of coxI exon 2 and contain a truncated version of the group I intron that likely cannot be spliced out. Received: 29 May 1997 / Accepted: 1 November 1997     

Patterns of Base Composition Within the Genes of Drosophila melanogaster

Base composition is not uniform across the genome of Drosophila melanogaster. Earlier analyses have suggested that there is variation in composition in D. melanogaster on both a large scale and a much smaller, within-gene, scale. Here we present analyses on 117 genes which have reliable intron/exon boundaries and no known alternative splicing. We detect significant heterogeneity in G+C content among intron segments from the same gene, as well as a significant positive correlation between the intron and the third codon position G+C content within genes. Both of these observations appear to be due, in part, to an overall decline in intron and third codon position G+C content along Drosophila genes with introns. However, there is also evidence of an increase in third codon position G+C content at the start of genes; this is particularly evident in genes without introns. This is consistent with selection acting against preferred codons at the start of genes. Received: 24 February 1997 / Accepted: 10 November 1997     

Actin Phylogeny Identifies Mesostigma viride as a Flagellate Ancestor of the Land Plants

Green algae and land plants trace their evolutionary history to a unique common ancestor. This ``green lineage' is phylogenetically subdivided into two distinct assemblages, the Chlorophyta and the Streptophyta. The Chlorophyta includes the Chlorophyceae, Trebouxiophyceae, Ulvophyceae, and Prasinopohyceae, whereas the Streptophyta includes the Charophyceae plus the bryophytes, ferns, and all other multicellular land plants (Embryophyta). The Prasinophyceae is believed to contain the earliest divergences within the green lineage. Phylogenetic analyses using rDNA sequences identify the prasinophytes as a paraphyletic taxon that diverges at the base of the Chlorophyta. rDNA analyses, however, provide ambiguous results regarding the identity of the flagellate ancestor of the Streptophyta. We have sequenced the actin-encoding cDNAs from Scherffelia dubia (Prasinophyceae), Coleochaete scutata, Spirogyra sp. (Charophyceae), and the single-copy actin gene from Mesostigma viride (Prasinophyceae). Phylogenetic analyses show Mesostigma to be the earliest divergence within the Streptophyta and provide direct evidence for a scaly, biflagellate, unicellular ancestor for this lineage. This result is supported by the existence of two conserved actin-coding region introns (positions 20-3, 152-1), and one intron in the 5′-untranslated region of the actin gene shared by Mesostigma and the embryophytes. Received: 10 July 1997 / Accepted: 9 April 1998     

Evolution of Orthologous Intronless and Intron-Bearing Globin Genes in Two Insect Species

While globin genes ctt-2β and ctt-9.1 in Chironomus thummi thummi each have a single intron, all of the other insect globin genes reported so far are intronless. We analyzed four globin genes linked to the two intron-bearing genes in C. th. thummi. Three have a single intron at the same position as ctt-2β and ctt-9.1; the fourth is intronless and lies between intron bearing genes. Finally, in addition to its intron, one gene (ctt-13RT) was recently interrupted by retrotransposition. Phylogenetic analyses show that the six genes in C. th. thummi share common ancestry with five globin genes in the distantly related species C. tentans, and that a 5-gene ancestral cluster predates the divergence of the two species. One gene in the ancestral cluster gave rise to ctn-ORFB in C. tentans, and duplicated in C. th. thummi to create ctt-11 and ctt-12. From parsimonious calculations of evolutionary distances since speciation, ctt-11, ctt-12, and ctn-ORFB evolved rapidly, while ctn-ORFE in C. tentans evolved slowly compared to other globin genes in the clusters. While these four globins are under selective pressure, we suggest that most chironomid globin genes were not selected for their unique function. Instead, we propose that high gene copy number itself was selected because conditions favored organisms that could synthesize more hemoglobin. High gene copy number selection to produce more of a useful product may be the basis of forming multigene families, all of whose members initially accumulate neutral substitutions while retaining essential function. Maintenance of a large family of globin genes not only ensured high levels of hemoglobin production, but may have facilitated the extensive divergence of chironomids into as many as 5000 species. Received: 31 December 1996 / Accepted: 16 May 1997     

Preserved Close Linkage Between the Genes Encoding Troponin I and Troponin T, Reflecting an Evolution of Adapter Proteins Coupling the Ca2+ Signaling of Contractility

Ca²⁺-regulated motility is essential to numerous cellular functions, including muscle contraction. Systems with troponin C, myosin light chain, or calmodulin as the Ca²⁺ receptor have evolved in striated muscle and other types of cells to transduce the cytoplasm Ca²⁺ signals into allosteric conformational changes of contractile proteins. While these Ca²⁺ receptors are homologous proteins, their coupling to the responding elements is quite different in various cell types. The Ca²⁺ regulatory system in vertebrate striated muscle represents a highly specialized such signal transduction pathway consisting of the troponin complex and tropomyosin associated with the actin filament. To understand the molecular mechanism in the Ca²⁺ regulation of muscle contraction and cell motility, we have revealed a preserved ancestral close linkage between the genes encoding two of the troponin subunits, troponin I and troponin T, in the genome of mouse. The data suggest that the troponin I and troponin T genes may have originated from a single locus and evolved in parallel to encode a striated muscle-specific adapter to couple the Ca²⁺ receptor, troponin C, to the actin–myosin contractile machinery. This hypothesis views the three troponin subunits as two structure–function domains: the Ca²⁺ receptor and the signal transducing adapter. This model may help to further our understanding of the Ca²⁺ regulation of muscle contraction and the structure–function relationship of other potential adapter proteins which are converged to constitute the Ca²⁺ signal transduction pathways governing nonmuscle cell motility. Received: 15 April 1999 / Accepted: 15 July 1999     

Insect muscle actins differ distinctly from invertebrate and vertebrate cytoplasmic actins

Summary Invertebrate actins resemble vertebrate cytoplasmic actins, and the distinction between muscle and cytoplasmic actins in invertebrates is not well established as for vertebrate actins. However, Bombyx and Drosophila have actin genes specifically expressed in muscles. To investigate if the distinction between muscle and cytoplasmic actins evidenced by gene expression analysis is related to the sequence of corresponding genes, we compare the sequences of actin genes of these two insect species and of other Metazoa. We find that insect muscle actins form a family of related proteins characterized by about 10 muscle-specific amino acids. Insect muscle actins have clearly diverged from cytoplasmic actins and form a monophyletic group emerging from a cluster of closely related proteins including insect and vertebrate cytoplasmic actins and actins of mollusc, cestode, and nematode. We propose that muscle-specific actin genes have appeared independently at least twice during the evolution of animals: insect muscle actin genes have emerged from an ancestral cytoplasmic actin gene within the arthropod phylum, whereas vertebrate muscle actin genes evolved within the chordate lineage as previously described.Offprint requests to.: N. Mounier     

Distribution and evolution of introns in drosophila amylase genes

While the two amylase genes of Drosophila melanogaster are intronless, the three genes of D. pseudoobscura harbor a short intron. This raises the question of the common structure of the Amy gene in Drosophila species. We have investigated the presence or absence of an intron in the amylase genes of 150 species of Drosophilids. Using polymerase chain reaction (PCR), we have amplified a region that surrounds the intron site reported in D. pseudoobscura and a few other species. The results revealed that most species contain an intron, with a variable size ranging from 50 to 750 bp, although the very majoritary size was around 60–80 bp. Several species belonging to different lineages were found to lack an intron. This loss of intervening sequence was likely due to evolutionarily independent and rather frequent events. Some other species had both types of genes: In the obscura group, and to a lesser extent in the ananassae subgroup, intronless copies had much diverged from intron-containing genes. Base composition of short introns was found to be variable and correlated with that of the surrounding exons, whereas long introns were all A-T rich. We have extended our study to non-Drosophilid insects. In species from other orders of Holometaboles, Lepidoptera and Hymenoptera, an intron was found at an identical position in the Amy gene, suggesting that the intron was ancestral. Received: 23 October 1995 / Accepted: 5 March 1996     

Evolutionary History of the 11p15 Human Mucin Gene Family

The four human mucin genes MUC6, MUC2, MUC5AC, and MUC5B are located at chromosome 11p15.5. It has been demonstrated that the three mucins MUC2, MUC5AC, and MUC5B contain several Cys-subdomains of 108 amino acid residues. In contrast, little information is available concerning MUC6. These Cys-subdomains contain 10 cysteine residues that have a highly conserved position. We present here a coherent probable evolutionary history of this human gene family after comparison of the nucleotide sequences of these Cys-subdomains. The three MUC loci MUC2, MUC5AC, and MUC5B may have evolved from a common ancestral gene by two successive duplications. Moreover, we can postulate that MUC5AC and MUC5B have evolved in a concerted manner, while MUC2 has evolved separately. Received: 30 January 1997 / Accepted: 17 April 1997     

Genomic organization of the extracellular coding region of the human FGFR4 and FLT4 genes: Evolution of the genes encoding receptor tyrosine kinases with immunoglobulin-like domains

Receptor tyrosine kinases with five, seven, and three Ig-like domains in their extracellular region are grouped in subclasses IIIa, IIIb, and IIIc, respectively. Here, we describe the genomic organization of the extracellular coding region of the human FGFR4 (IIIc) and FLT4 (IIIb) genes and compare it to that of the human FGFR1(IIIc), KIT, and FMS (IIIa). The results show that while genes belonging to the same subclass have an identical exon/intron structure in their extracellular coding region—as they do in their intracellular coding region—genes of related subclasses only have a similar exon/intron structure. These results strongly support the hypothesis that the genes of the three subclasses evolved from a common ancestor by duplications involving entire genes, already in pieces. Hypotheses on the origin of introns and on the difference in the number of extracellular Ig-like domains in the three gene subclasses are discussed. Received: 19 August 1996 / Accepted: 2 January 1997     

Origin and Inheritance of Group I Introns in 26S rRNA Genes of Gaeumannomyces graminis

Studies of the distribution of the three group I introns (intron A, intron T, and intron AT) in the 26S rDNA of Gaeumannomyces graminis had suggested that they were transferred to a common ancestor of G. graminis var. avenae and var. tritici after it had branched off from var. graminis. Intron AT and intron A exhibited vertical inheritance and coevolved in concert with their hosts. Intron loss could occur after its acquisition. Loss of any one of the three introns could occur in var. tritici whereas only loss of intron T had been found in the majority of var. avenae isolates. The existence of isolates of var. tritici and var. avenae with three introns suggested that intron loss could be reversed by intron acquisition and that the whole process is a dynamic one. This process of intron acquisition and intron loss reached different equilibrium points for different varieties and subgroups, which explained the irregular distribution of these introns in G. graminis. Each of the three group I introns was more closely related to other intron sequences that share the same insertion point in the 26S rDNA than to each other. These introns in distantly related organisms appeared to have a common ancestry. This system had provided a good model for studies on both the lateral transfer and common ancestry of group I introns in the 26S rRNA genes. Received: 17 May 1996 / Accepted: 14 January 1997     

Ferritin gene organization: Differences between plants and animals suggest possible kingdom-specific selective constraints

Ferritin, a protein widespread in nature, concentrates iron ∼10¹¹–10¹²-fold above the solubility within a spherical shell of 24 subunits; it derives in plants and animals from a common ancestor (based on sequence) but displays a cytoplasmic location in animals compared to the plastid in contemporary plants. Ferritin gene regulation in plants and animals is altered by development, hormones, and excess iron; iron signals target DNA in plants but mRNA in animals. Evolution has thus conserved the two end points of ferritin gene expression, the physiological signals and the protein structure, while allowing some divergence of the genetic mechanisms. Comparison of ferritin gene organization in plants and animals, made possible by the cloning of a dicot (soybean) ferritin gene presented here and the recent cloning of two monocot (maize) ferritin genes, shows evolutionary divergence in ferritin gene organization between plants and animals but conservation among plants or among animals; divergence in the genetic mechanism for iron regulation is reflected by the absence in all three plant genes of the IRE, a highly conserved, noncoding sequence in vertebrate animal ferritin mRNA. In plant ferritin genes, the number of introns (n= 7) is higher than in animals (n= 3). Second, no intron positions are conserved when ferritin genes of plants and animals are compared, although all ferritin gene introns are in the coding region; within kingdoms, the intron positions in ferritin genes are conserved. Finally, secondary protein structure has no apparent relationship to intron/exon boundaries in plant ferritin genes, whereas in animal ferritin genes the correspondence is high. The structural differences in introns/exons among phylogenetically related ferritin coding sequences and the high conservation of the gene structure within plant or animal kingdoms suggest that kingdom-specific functional constraints may exist to maintain a particular intron/exon pattern within ferritin genes. In the case of plants, where ferritin gene intron placement is unrelated to triplet codons or protein structure, and where ferritin is targeted to the plastid, the selection pressure on gene organization may relate to RNA function and plastid/nuclear signaling. Received: 25 July 1995 / Accepted: 3 October 1995     

The Complete Amino Acid Sequence of Actins from Bovine Aorta, Bovine Heart, Bovine Fast Skeletal Muscle, and Rabbit, Slow Skeletal Muscle

Complete amino acid sequences for four mammalian muscle actins are reported: bovine skeletal muscle actin, bovine cardiac actin, the major component of bovine aorta actin, and rabbit slow skeletal muscle actin. The number of different actins in a higher mammal for which full amino acid sequences are now available is therefore increased from two to five. Screening of different smooth muscle tissues revealed in addition to the aorta type actin a second smooth muscle actin, which appears very similar if not identical to chicken gizzard actin. Since the sequence of chicken gizzard actin is known, six different actins are presently characterized in a higher mammal.
The two smooth muscle actins—bovine aorta actin and chicken gizzard actin—differ by only three amino acid substitutions, all located in the amino-terminal end. In the rest of their sequences both smooth muscle actins share the same four amino acid substitutions, which distinguish them from skeletal muscle actin. Cardiac muscle actin differs from skeletal muscle actin by only four amino acid exchanges. No amino acid substitutions were found when actins from rabbit fast and slow skeletal muscle were compared.
In addition we summarize the amino acid substitution patterns of the six different mammalian actins and discuss their tissue specificity. The results show a very close relationship between the four muscle actins in comparison to the nonmuscle actins. The amino substitution patterns indicate that skeletal muscle actin is the highest differentiated actin form, whereas smooth muscle actins show a noticeably closer relation to nonmuscle actins. By these criteria cardiac muscle actin lies between skeletal muscle actin and smooth muscle actins.     

Genomic organization and evolution of actin genes in the amphioxus Branchiostoma belcheri and Branchiostoma floridae

We previously described the cDNA cloning and expression patterns of actin genes from amphioxus Branchiostoma floridae (Kusakabe, R., Kusakabe, T., Satoh, N., Holland, N.D., Holland, L.Z., 1997. Differential gene expression and intracellular mRNA localization of amphioxus actin isoforms throughout development: implications for conserved mechanisms of chordate development. Dev. Genes Evol. 207, 203–215). In the present paper, we report the characterization of cDNA clones for actin genes from a closely related species, Branchiostoma belcheri, and the exon–intron organization of B. floridae actin genes. Each of these two amphioxus species has two types of actin genes, muscle and cytoplasmic. The coding and non-coding regions of each type are well-conserved between the two species. A comparison of nucleotide sequences of muscle actin genes between the two species suggests that a gene conversion may have occurred between two B. floridae muscle actin genes BfMA1 and BfMA2. From the conserved positions of introns between actin genes of amphioxus and those of other deuterostomes, the evolution of deuterostome actin genes can be inferred. Thus, the presence of an intron at codon 328/329 in vertebrate muscle and cytoplasmic actin genes but not in any known actin gene in other deuterostomes suggests that a gene conversion may have occurred between muscle and cytoplasmic actin genes during the early evolution of the vertebrates after separation from other deuterostomes. A Southern blot analysis of genomic DNA revealed that the amphioxus genome contains multiple muscle and cytoplasmic actin genes. Some of these actin genes seem to have arisen from recent duplication and gene conversion. Our findings suggest that the multiple genes encoding muscle and cytoplasmic actin isoforms arose independently in each of the three chordate lineages and that gene duplications and gene conversions established the extant actin multigene family during the evolution of chordates.