Major Divisions in Oscines Revealed by Insertions in the Nuclear Gene c-myc: A Novel Gene in Avian Phylogenetics

The order Passeriformes is a monophyletic group consisting of more than half of all living birds species (Raikow 1982). The major split of the passerines into the suboscines and oscines is well supported by morphological characters, although a few taxa (e.g. Acanthisittidae, New Zealand wrens) defy allocation to either suborder (see Sibley and Ahlquist 1990). Molecular analyses corroborate this dichotomy in passerines (Sibley and Ahlquist 1990, Edwards et al. 1991). Some oscine families are distinct, but convergent evolution apparently is common and has obscured phylogenetic relationships, making the subdivision of this group based on morphology difficult (Beecher 1953, Tordoff 1954, Ames 1971, Raikow 1978, Bledsoe 1988). In fact, even the delimitations of most families are uncertain, and only two families, the Alaudidae (larks) and the Hirundinidae (swallows and martins), are unambiguously defined (Mayr 1958). Consequently, oscine relationships at the family level and above are insufficiently known, and all taxonomic arrangements are controversial. Besides the larks and swallows, three main groups of oscines have been recognized based on morphology: (1) Old World insect-eaters and their relatives; (2) New World insect-eaters and finches; and (3) crows, birds-of-paradise, and associated families (Mayr and Greenway 1956, Voous 1985). Before the advent of quantitative biochemical methods, most systematists recognized these groups, and the major debate concerned how they were related (Voous 1985). Although all combinations of the three groups have been advocated at one time or another, a major issue is whether the crows and their allies constitute the sister group to all other oscines, or are nested within them. The fully developed double pneumatic fossae in the proximal end of the humerus present in many oscines, but not in crows and allies or in the suboscines (Bock 1962), suggests the existence of a clade including all oscines except crows and their allies. This dichotomy has been supported by DNA-DNA hybridization studies (Sibley and Ahlquist 1990, Harshman 1994, Sheldon and Gill 1996). In the classification of Sibley and Monroe (1990), the dichotomy is reflected by the division of the oscines into the parvorders Corvida and Passerida. The Passerida is further divided into the superfamilies Muscicapoidea, Sylvioidea, and Passeroidea. The DNA-DNA hybridization method as applied by Sibley and Ahlquist has been criticized on several grounds, and doubts concerning the validity of some of their results have been raised (Cracraft 1987, Houde 1987, Sarich et al. 1989, Sheldon and Bledsoe 1993). However, the currently favored method in molecular systematics, the comparison of nucleotide sequences, so far has generated few phylogenetic hypotheses at this high taxonomic level in oscines (but see Edwards et al. 1991, Groth 1998). Here, we present a hypothesis of phylogenetic relationships among oscines based on two previously undescribed insertions in exon 3 of c-myc. This hypothesis defines major groups of songbirds. C-myc is a nuclear proto-oncogene that encodes a protein transcription factor that plays a crucial role in the regulation of cell proliferation and apoptosis (Bouchard et al. 1998). The sequence of c-myc is highly conserved throughout the vertebrates, especially compared with the more rapidly evolving mitochondrial genes. Although no dates are known for splits between the evolutionary lineages studied herein, some of them might be very old, perhaps even of early Tertiary age (Feduccia 1995). Mutational saturation can reduce the resolving power of gene sequences and might be a problem when using mitochondrial genes to study ancient branching events in birds. In contrast, dissimilarities between c-myc sequences increase nearly linearly for evolutionary divergences well beyond 100 million years ago (Graybeal 1994). To investigate early avian divergences, we have sequenced about 500 base pairs of exon 3 of this gene for more than 150 species representing 65 nonpasserine and 36 passerine families. Our results confirm the slow rate of evolution of c-myc in birds. The maximum sequence divergence observed was about 11%, and only three indels occurred. Only one indel, an insertion of four amino acids relative to the published chicken sequence, has been observed outside the passerines. Representatives of 46 passerine families were selected for study (Table 1). Special emphasis was placed on sampling the superfamily Passeroidea sensu Sibley and Ahlquist (1990). If not stated otherwise, the usage of family and subfamily names follows Morony et al. (1975). However, at higher levels, i.e. superfamilies and parvorders, we use the terminology of Sibley and Ahlquist (1990) to facilitate comparisons between our results and their phylogenetic hypotheses. Distribution of taxa used in this study. Family names and lower taxonomic categories follow Morony et al. (1975), and higher categories follow Sibley and Monroe (1990). AM = Australian Museum, NRM = Swedish Museum of Natural History, ZMCU = Zoological Museum of the University of Copenhagen, and NCBI = National Center for Biotechnology Information (GenBank) Distribution of taxa used in this study. Family names and lower taxonomic categories follow Morony et al. (1975), and higher categories follow Sibley and Monroe (1990). AM = Australian Museum, NRM = Swedish Museum of Natural History, ZMCU = Zoological Museum of the University of Copenhagen, and NCBI = National Center for Biotechnology Information (GenBank) We extracted genomic DNA from tissue or blood using standard techniques of Proteinase K/SDS digestion followed by phenol chloroform extraction and ethanol precipitation, or by QIAamp DNA extraction kits following manufacturer's recommendations. Amplification was performed with primer pairs mycEX3A (CAAGAAGAAGATGAGGAAAT) and RmycEX3A (TTAGCTGCTCAAGTTTGTG), or mycEX3D (GAAGAAGAACAAGAAGAAGATG) and RmycEX3D (ACGAGAGTTCCTTAGCTGCT), developed by Thomas J. Parsons. Sequencing was performed with primers mycEX3A and RmycEX3A using Perkin Elmer Applied BioSystems 373 or 377 automated fluorescent sequencing instruments, and Perkin Elmer Applied BioSystems PRISM terminator cycle sequencing kits with AmpliTaq FS polymerase (either standard rhodamine and BigDye chemistries were employed). Sequence assembly was performed using the Perkin Elmer Applied BioSystems Sequence Navigator or the DNASTAR SeqMan II programs. Alignments of completed sequences were performed by eye. Indications of sequence positions throughout this report are relative to the numbering of the full-length protein-coding sequence of the chicken (Watson et al. 1983). Nucleotide sequences of exon 3 of c-myc have been studied in 80 species of suboscine and oscine passerines, representing 46 traditional families (Table 1). The sequences vary from 498 to 510 bases (corresponding to 166 to 170 amino acids) in length as a consequence of the presence or absence of two insertions consisting of one and three amino acids, respectively. These two insertions have not been observed among 65 nonpasseriform families, but they appear to exhibit consistent taxonomic distributions within the Passeriformes (with no reversals inferred on the portions of the tree where relationships are well established). Thus, they presumably represent unique and significant evolutionary events in passerine evolution. The ancestral state in passerines of no insertions was observed in all nonpasseriforms investigated and also was found in all suboscine and Corvida families (Table 2). All oscine families representing the parvorder Passerida that we examined possessed an insertion of a single amino acid at nucleotide position 793 relative to the chicken c-myc sequence (Watson et al. 1983). The occurrence of this insertion in all oscine passerines except the Corvida supports the hypothesis based on DNA-DNA hybridization of a sister-group relationship between the Corvida and all other oscines. In most families, this extra amino acid is a threonine. However, it is a proline in Hirundo and Sylvia and a serine in Certhia, Carduelis, and Icterus. Taxonomic distribution of the two insertions of amino acids in exon 3 of the nuclear c-myc gene in a survey of passerines. The insertions occur at positions 793 and 991 in the published c-myc sequence of the chicken (Watson et al. 1983). Genera, families, and subfamilies are based on the “traditional” classification of Morony et al. (1975), and superfamilies for oscines are from Sibley and Monroe (1990) based on DNA-DNA analysis Taxonomic distribution of the two insertions of amino acids in exon 3 of the nuclear c-myc gene in a survey of passerines. The insertions occur at positions 793 and 991 in the published c-myc sequence of the chicken (Watson et al. 1983). Genera, families, and subfamilies are based on the “traditional” classification of Morony et al. (1975), and superfamilies for oscines are from Sibley and Monroe (1990) based on DNA-DNA analysis At position 991, the Motacillidae, Fringillidae, Emberizidae, Parulidae, and Icteridae share an additional insertion of three amino acids relative to the chicken (Table 2). The first two of these are always a serine and a glycine. The third amino acid varies more among the families. Most taxa have a serine, but motacillids (Motacilla and Anthus) have threonine; Geothlypis, Parula, Carpodacus, and Icterus have leucine; Conirostrum has phenylalanine; and Carduelis has tryptophan. Some silent third-position variation in codon coding also occurs for this third inserted amino acid. We consider the passerine c-myc insertions described here to represent two unique evolutionary events, with no reversals evident in the taxa studied. This pattern is strongly suggested by the extreme rarity of indels in c-myc exon 3 throughout avian taxa. For example, among 102 nonpasserine species studied, representing 65 families, only one indel has been observed. This insertion of four amino acids relative to the chicken sequence occurs at position 796, i.e. at a different position than the passerine insertions reported here. The conservation in sequence length of c-myc may be due to the fact the myc protein has a helix-loop-helix structure that must form a heterdimeric complex with the regulatory Max protein. The central regulatory role of myc in cell division and development likely would tolerate little functional variation (Bouchard et al. 1998, Eilers 1999). Length changes may be rare owing to a requirement for radical compensatory changes in other genes, with reversals encountering an evolutionary hurdle of equivalent magnitude. Table 2 indicates that multiple amino-acid substitutions have occurred within the single amino-acid insertion, with possibly three substitutions of proline for threonine and two substitutions of serine for threonine. This further supports the low rate of indel mutations compared with the already slow rate of amino-acid sequence substitution. Likewise, the third amino acid of the three that are inserted displays substantial variation within related groups, whereas the length of insertion remains constant. The insertion involving a single amino acid observed in the c-myc sequence is a synapomorphy for all oscines that we studied, except species in the parvorder Corvida (Fig. 1). This observation supports the sister-group relationship of the corvids and their allies relative to other oscines, as suggested by DNA-DNA hybridization (Sibley and Ahlquist 1990, Harshman 1994, Sheldon and Gill 1996). Unfortunately, only one representative of the superfamily Menuroidea was available to us. Major divisions of passerines as indicated by insertions of amino acids in the nuclear gene c-myc. The first insertion is synapomorphic for the parvorder Passerida (sensu Sibley and Ahlquist 1990), whereas all representatives of the New World nine-primaried oscines, the primarily Old World finches, and the Motacillidae share a second insertion of amino acids Passerines typically have 10 primaries, which is generally agreed to be the ancestral condition. In several oscine families, the outermost primary is secondarily reduced or lost, and species in these groups are effectively nine-primaried. Which families are nine-primaried has been a matter of considerable confusion, however. Some families that are regarded as “nine-primaried” include species in which the tenth primary is in fact present, although vestigial. A long-recognized group of truly nine-primaried families is the so-called “New World nine-primaried oscines” that consist of the Parulidae, Emberizidae (Emberizinae, Thraupinae, Cardinalinae), and Icteridae (Raikow 1978, Feduccia 1996). Although not all of these families are confined to the New World, they are concentrated there. All representatives of the New World nine-primaried oscines that we analyzed (Parulidae, Emberizinae, Thraupinae, Cardinalinae, and Icteridae) possess the insertion of three amino acids at position 991 in the chicken sequence. This is a strong indication of the shared common ancestry of this group. Moreover, the Fringillidae and Motacillidae also share this insertion. The fringillids and motacillids are included in the Passeroidea by Sibley and Ahlquist (1990), along with the New World nine-primaried oscines. However, in other families in Passeroidea and studied herein (Alaudidae, Nectariniidae, Dicaeidae, Estrildidae, Passeridae, and Prunellidae), this insertion is absent. The c-myc data thus support a clade consisting of the New World nine-primaried oscines, the primarily Old World finches, and the wagtails and pipits. The Motacillidae have a vestigial tenth primary and traditionally have not been thought to be closely related to the New World nine-primaried oscines, although cytochrome-b sequence data suggest them to be closer to the Emberizidae than are the Fringillidae (Groth 1998). Cytochrome-b sequence data also suggest that the ten-primaried Passeridae are nested within this clade of emberizids, fringillids, and motacillids (Groth 1998). This arrangement is not supported by c-myc data, because the three species of Passeridae (=Ploceidae sensu Morony et al. 1975) we studied do not share the insertion of three amino acids with the rest of the group. It could be argued that the insertions reported herein, as single characters, should not be afforded more weight than other molecular characters. However, we believe that these insertions represent unique evolutionary events of unequivocal homology, with no reversal. As such, they present powerful evidence regarding relationships within passerines that have been difficult to resolve based on other potentially quite homoplastic characters. The greatly increased significance of unique molecular rearrangements has been recognized elsewhere (Batzer et al. 1996), and shared indels in protein-coding genes previously have been interpreted as strong markers for monophyly as long as the observations are based on wide taxonomic sampling (van Dijk et al. 1999). We studied sequences from more than 110 families of passerines and nonpasserines. The extreme low frequency of indels in c-myc, and the taxonomic distribution of insertions that we report, indicate that these should be considered highly significant characters for elucidating the evolution of passerines. Tissue and blood samples were kindly put at our disposal by the Zoological Museum, Copenhagen (Peter Arctander, Jon Fjeldså, and Jaime Garcia-Moreno); the Australian Museum, Sydney (Walter Boles); and the Department of Zoology, University of Gothenburg (Staffan Andersson). The other samples used belong to the Swedish Museum of Natural History. Most of these were collected during extensive fieldwork in Paraguay as part of collaboration with the Museo Nacional de Historia Natural del Paraguay, San Lorenzo. The Direccion de Parques Nacional y Vida Silvestre issued the necessary collecting and exports permits in Paraguay. The Magnus Bergvalls Stiftelse, Olle och Signhild Engkvists Stiftelser, and the Swedish Natural Science Research Council (grant B-AA/BU 01913-304) funded this project. Michael J. Braun and the Smithsonian Laboratory of Molecular Systematics provided initial material support to TJP and strongly encouraged investigation of the phylogenetic potential of the c-myc gene for resolving avian relationships. Gullevi Bergkvist and Mari Källersjö provided advice and assistance in the laboratory. The paper has benefitted significantly from the comments of Scott V. Edwards, Frederick H. Sheldon, and an anonymous reviewer.