S13.1: Molecular evolution of the mtDNA control-region in galliform birds

The mitochondrial genome (mtDNA) of vertebrates is a small, circular molecule of about 16000 - 21000 nucleotides (nt), with conserved structural organisation but with a rate of sequence evolution that is, on average, faster than nuclear genes (Brown et al. 1986). MtDNA contains the genes for two ribosomal RNAs (rRNA), 22 transfer RNAs (tRNA), 13 enzymes and the control-region (CR), which regulates the replication of the H-strand and transcription of all mtDNA genes (Clayton 1992).

The structural organisation of mammalian and avian mtDNA is similar, except for two important differences: (1) avian mtDNA apparently lacks the non-coding hairpin sequence in the WANCY tRNA cluster, which has been identified as the promoter of L-strand replication; and (2) the 5’ region flanking the CR has a novel gene order in birds, involving the cytochrome b, NADH-dehydrogenase 6 and three tRNA genes (Desjardins & Morais 1990; Quinn & Wilson 1993). In consequence, the avian CR is flanked by tRNA-Glu and tRNA-Phe genes, and not by tRNA-Pro and tRNA-Phe as is usual for vertebrates.

The CR is considered the most variable part of mtDNA; however, nucleotide substitutions and insertion/deletion of short nucleotide strips or longer repeats are not distributed randomly, but are concentrated at particular hypervariable sites and domains (Randi & Lucchini 1998). The avian CR can be divided into three domains, corresponding to the variable peripheral domains I and III, and to the central conserved domain II of vertebrates (Baker & Marshall 1997; Randi & Lucchini 1998). Domain I, adjacent to the tRNA-Glu gene, contains conserved sequences with functional roles in controlling the termination of H-strand DNA replication (TAS), and can include a variable number of tandemly repeated sequences (VNTR). Domain II is very conserved, with blocks of conserved sequences that are present in most vertebrates studied so far. Although the functional roles of these conserved sequence blocks are still unknown, their evolutionary stability suggests that they are constrained by strong selective pressures. Domain III contains the the light and heavy strand promoters of transcription (LSP and HSP), the origin of H-strand (O_H) replication, and other conserved sequence blocks (CSB1, CSB2 and CSB3) that regulate the replication of mtDNA (Clayton 1992; Baker & Marshall 1997; Randi & Lucchini 1998).

Interspersed among the conserved blocks are hypervariable sequences and nucleotide sites that contribute most of the intraspecific diversity and interspecific divergence (Baker & Marshall 1997; Randi & Lucchini 1998). Sequence variability is generated by point mutations and strand slippage during DNA replication. Strand slippage might produce length differences, making alignments in certain parts of the CRs problematic and increasing homoplasy in phylogenetic inference (Randi & Lucchini 1998).

Although the hypervariable parts of CR have been successfully used for high resolution genetic studies at the population level (see the review by Baker & Marshall 1997; Lucchini & Randi 1998), the CR apparently evolves at a remarkably slow rate, on average, in birds. Its average substitution rate is comparable to the rate of cytochrome b and other mtDNA protein-coding genes (Randi & Lucchini 1998; Zink and Blackwell 1998). Notwithstanding the existence of regions of problematic alignment and hypervariability, previously published case-studies suggested that avian CR sequences can express significant phylogenetic signal, comparable to that of mtDNA protein-coding genes (Kimball et al. 1997; Kidd & Friesen 1998; Randi & Lucchini 1998; Zink & Blackwell 1998). Therefore, the avian CR can be reliably used for phylogenetic inference. In a previously published study (Randi & Lucchini 1998), we used complete CR sequences to reconstruct phylogenetic relationships among Alectoris partridges that were entirely congruent with a cytochrome b phylogeny (Randi 1996). These results have stimulated us to develop a more extensive project to study the patterns and rates of CR molecular evolution in the Galliformes, and to explore CR’s phylogenetic content in a range of about 60 million years (MYR) of avian evolutionary divergence.

Total DNA was extracted from feather roots and tissue samples, stored in 95% ethanol, using guanidinium thiocyanate and diatomaceous silica particles (a procedure modified after Gerloff et al. 1995). We collected DNA samples of 89 species of Galliformes representing 45 genera of Phasianidae (13 Perdicini and 22 Phasianini, including Meleagris), Numida (Numididae), 4 genera of Odontophoridae, and two Megapodiidae. We follow Sibley & Ahlquist’s (1990) taxonomy of Galliformes. To control for authenticity of mtDNA sequences and exclude possible nuclear-copy DNA contaminations (Kidd & Friesen 1998), we have obtained purified mtDNA preparations through alkaline lysis of mitochondrial pellets (Palva & Palva 1985) in a selected subset of samples. In all cases, CR sequences obtained from total DNA and from purified mtDNA were identical, suggesting mitochondrial authenticity and not nuclear-copy contaminations.

The entire mtDNA CR was amplified using the Polymerase Chain Reaction (PCR) following the protocol described by Randi & Lucchini (1998). Purified PCR products were sequenced by double-stranded DNA cycle-sequencing with ABI Prism Dye Terminator chemicals in an ABI 373 automated DNA sequencer.

The CR sequences were first aligned automatically using CLUSTAL X (Thompson et al. 1997). Alignments were then adjusted manually in regions of hypervariability and length heterogeneity within domains I and III, assuming that: (1) transitional (Ti) substitutions are more common than transversions (Tv); and (2) short insertion/deletions (indels) are more common that Tv. The computer program LOOPDLOOP (Gilbert 1992) was used to draw potential secondary structures in single-stranded regions of domain I. Nucleotide composition and genetic distances were computed using MEGA 1.01 (Kumar et al. 1993), with pairwise exclusion of gaps. Phylogenetic reconstructions were obtained by: (1) neighbor-joining (NJ; Saitou & Nei 1987) clustering of genetic distance matrices, using MEGA; (2) maximum-parsimony procedures as implemented in PAUP 3.1.1 (MP; Swofford 1993), with unordered character states, 10 replicates of random-addition of terminal sequences and branch-and-bound option; and (3) maximum-likelihood procedures as implemented in PUZZLE 4.0 (ML; Strimmer and von Haeseler, 1996), with different models of nucleotide substitution. Robustness of phylogenies was determined by 500 bootstrap replicates (Felsenstein 1985) using the appropriate options in MEGA.

The CR of Galliformes is 1157 nt long, on average, ranging from 1118 nt in Alectura lathami (Megapodiidae) to 1181 in Scleroptila levaillantii (Perdicini). Short indels were introduced to optimise the alignments in the two peripheral domains, and particularly in the 3’ terminal part of CR, upstream to the LSP/HSP sequence block. In this region relatively long strips of poly-T and poly-A nucleotides might promote extensive strand slippage at DNA replication, with consequent insertion/deletion of one or few nt. Automatic alignments resulted in exceptionally high numbers of Tv (A/T substitutions), which can be minimized by ad hoc insertion of gaps. Of course, there are no objective criteria for inserting gaps, and therefore, this terminal part of the alignment was not used for phylogenetic inference. Nucleotide substitutions were also extensive in the hypervariable part of domain I, making this region unsuitable for phylogenetic inference. Except these two relatively short hypervariable regions, the alignment of the CRs was straightforward, and their length was remarkably conserved among galliforms. Only the four species of Gallus showed a VNTR sequence, about 60 nt long, in domain I, as previously described by Fumihito et al. (1994).

Domain I of the CR of galliforms can be divided into a conserved part A and a hypervariable part B, followed by a conserved central domain II and a variable domain III (Randi & Lucchini 1998). In part A of domain I all sequenced galliforms showed a very conserved ‘goose hairpin’ sequence (Quinn & Wilson 1993; Randi & Lucchini 1998) that is followed by two rather conserved blocks with motifs similar to the mammalian termination associated sequences (TAS and ETAS; Sbisà et al. 1997). These blocks include TCCC, GYRCAT and CSB-like motifs, which should have functional roles in regulating the replication of mtDNA H-strands (Sbisà et al. 1997; Randi & Lucchini 1998). Single-stranded domain I can form potentially stable secondary hairpin structures. Although details of these structures can be different in the different species, the ‘goose hairpin’ and GYRCAT sequences are always involved in main loops and stem hairpins. Secondary structure is supposed to have a functional role in controlling mtDNA replication (Baker & Marshall 1997). Part B of domain I contains the ‘pseudo-copy’ and the ‘original’ copy of a VNTR sequence that is tandemly duplicated only in the four species of Gallus (Fumihito et al. 1994; Randi & Lucchini 1998).

Central conserved domain II was highly conserved among the galliforms and showed sequence similarity to the conserved F, E, D, and C boxes of vertebrates. However, these boxes do not include the most conserved parts of domain II of galliforms. There are other very conserved blocks at the 3’ end of domain II, were there is a sequence block 35 nt long which is completely conserved among galliforms. This block is just four nt downstream of the putative O_H, and could have functional, albeit still undescribed, importance in controlling the start of H-strand replication.

Peripheral domain III showed sequences with moderate similarity to O_H, CSB1, CSB2/3 and LSP/HSP. These extensive comparative sequence analyses suggest that regulatory functions in CR domain III might not be associated with strongly conserved sequence motifs. The terminal part of domain III is rich in variable numbers of poly-T and poly-A strings, which produce interspecific length variation and can be aligned only by ad hoc insertions of gaps to minimize an otherwise exceptionally high number of T/A transversions.

The aligned CRs of galliforms (about 860 nt, excluding three hypervariable blocks at nt 1 - 15, 190 -350, and 1030 to the end of the alignment) express significant phylogenetic signal, and NJ, MP and ML trees show rather concordant topologies which are characterised by:

(1) A basal position of the Megapodiidae (in trees rooted using the CR of Struthio, or rooted at the midpoint of the longest internal branch. Unfortunately, we have not gathered CR sequences of the Cracidae. CR sequences support the megapodes as the sister lineage of all the other galliforms, in accordance with all current taxonomic arrangements based on morphologic and genetic data (Sibley & Ahlquist 1990).

(2) The New World Quail, Odontophoridae, constitute the next strongly supported (100% bootstrap values) basal group, and are the sister lineage of all the other galliforms. The phylogenetic position of the Odontophoridae as reconstructed by CR sequences is congruent with most of the available genetic data, including DNA-DNA hybridization distances (Sibley & Ahlquist 1990).

(3) The next distinct branch leads to Numida meleagris. This basal position relative to the remaining galliforms is supported by bootstrap values of 67%. Once again, the CR gives support to current morphological and molecular taxonomy of Numididae as a distinct and basal family of the Galliformes (Sibley & Ahlquist 1990).

(4) Although CR sequences define (by significant bootstrap values) the monophyly of four families of galliforms (Megapodiidae, Odontophoridae, Numididae and Phasianidae), they are not able to resolve basal relationships within the Phasianidae. In particular, CR sequences can not sort the Phasianini and Perdicini into monophyletic clades. Therefore, the basal relationships among the Phasianinae are represented by an unresolved polytomy including six groups plus seven individual lineages.

Phylogenetic analyses of CR sequences support the following clades in the Phasianidae:

(1) The Turkey, Meleagris gallopavo, representatives of three genera of Perdicini (Xenoperdix, Arborophila and Perdix), and representatives of three genera of Phasianini (Pucrasia, Lophophorus, Ithaginis) constitute distinct lineages that are basal to almost all the other Phasianidae, and not obviously linked to any other group. Their positions are not resolved in terms of bootstrap support. These seven species are distantly related to all the other Phasianidae and can not be significantly linked to any other clade using CR sequence information.

(2) The grouse (family Tetraonidae or subfamily Tetraonini, according to different taxonomic arrangements) constitute a strongly supported monophyletic assemblage (> 90% bootstrap support), with Bonasa as sister lineage to all the other genera.

(3) Tragopan and the Peacock-pheasants, Polyplectron, are sister lineages, joined in a clade supported by low bootstrap values (56%).

(4) Gallus and the Quail-francolins (Scleroptila, Peliperdix, Francolinus; according to Bloomer & Crowe 1998) join with Bambusicola thoracica in a strongly supported clade (98% bootstrap support).

(5) A clade of Phasianini (75% bootstrap support) includes genera Catreus, Syrmaticus, Phasianus, Crysolophus, Lophura and Crossoptilon.

(6) The peacocks (Pavo), Afropavo, Argusianus and Rheinardia constitute a strongly supported clade (91% bootstrap).

(7) The Partridge-francolins (genus Pternistis; Bloomer & Crowe 1998) and Perdicula asiatica are sister lineages (96% bootstrap value) and join Coturnix, Margaroperdix, Tetraogallus and the monophyletic Alectoris in a strongly supported clade (84% bootstrap).

Phylogenetic analyses of CR sequences do not reconstruct any monophyletic assemblage of taxa corresponding to the traditional Perdicini. Some Perdicini are distantly related to all the other taxa, and are basal in the phylogenetic tree of Galliformes. All the other Perdicini join in two monophyletic clades, one of which includes Gallus. Phylogenetic relationships among these two clades are not resolved by CR sequences. Assuming that Perdicini exists as a monophyletic assemblage, we have analysed their relationships through a CR alignment which excluded all the Phasianini. Phylogenetic trees showed that: (1) Xenoperdix, Arborophila, Perdix, Coturnix and Margaroperdix are in basal positions; (2) the Quail-francolins split into two clades: Scleroptila and Francolinus plus Peliperdix, which join Bambusicola and Gallus in a 100% supported clade; (3) the Partridge-francolins are monophyletic and linked to Tetraogallus and Alectoris. CR relationships within monophyletic Alectoris, Partridge-francolins, Quail-francolins and Gallus are concordant with relationships described in previously published genetic studies (Randi 1996; Randi & Lucchini 1998; Bloomer & Crowe 1998; Fumihito et al. 1994).

Phylogenetic analyses of CR sequences also do not reconstruct any monophyletic assemblage of taxa corresponding to the traditional Phasianini. Assuming the Phasianini exists as a monophyletic assemblage, we have analysed their relationships through a CR alignment excluding all the Perdicini. Phylogenetic trees showed that: (1) the grouse are monophyletic and Bonasa is their basal lineage; (2) Pavo and Afropavo are closely related, and Argusianus and Rheinardia are their sister lineages; (3) Tragopan might be related to Pavo and Polyplectron; (4) Lophophorus, Ithaginis, and Pucrasia are basal to the other pheasants.

The mtDNA CR of Galliformes evolves as a patchwork of hypervariable and strongly conserved sites. Comparative analyses of variability across the CR allow mapping some conserved sequences of putative functional importance in domains I and III. These conserved motifs are intermixed with highly variable short regions in domain I and in the terminal part of domain III. Hypervariable sites might quickly accumulate multiple substitutions. Other hypervariable regions, particularly at domain III, might evolve by strand slippage during DNA replication. Hypervariability and slippage makes it difficult to obtain objective alignments of these regions of the CR, and might generate homoplasy in phylogenetic reconstructions. These hypervariable regions (corresponding to about 300 nt over an average length of 1157 nt of the CR of galliforms) can not be used in phylogenetic analyses. All the other parts of the CR evolve slowly, and in particular the central domain II is very conserved among the galliforms. Overall, excluding the hypervariable sites, the CR evolves at remarkably slow rate, comparable to rates in cytochrome b, and can be reliably used for phylogenetic inference.

CR sequences reconstructed phylogenetic relationships among the four galliform families that are concordant with other genetic data, in particular DNA-DNA hybridisation, and with currently accepted taxonomic arrangements (Sibley & Ahlquist 1990): (1) the Megapodiidae are the sister lineage of all the other galliforms; and (2) the Odontophoridae and Numididae are the sister lineages of all the other Phasianidae (including Meleagris). On the other hand, basal phylogenetic relationships among the Phasianidae are not resolved by CR sequences. At the moment, it is not clear if this polytomy is the consequence of the low resolving power of CR sequences or is the consequence of a sudden and almost contemporaneous evolutionary radiation of the Phasianidae.

This basal polytomy makes it impossible to resolve the traditionally recognised subfamilies Perdicini and Phasianini into two well-supported monophyletic clades. In particular, the strong association among Gallus, Bambusicola and the Quail-francolins contrasts with hypothesized monophyletic Perdicini and Phasianini. The resolution of basal relationships among Phasianini and a deeper understanding of the puzzling relationships among Gallus, Bambusicola and the Quail-francolins require additional genetic data.

Phylogenetic resolution of the CR is concordant with conclusions from other genes or methods. In particular, CR sequences strongly support: (1) the polyphyly of traditional francolins, as suggested by other molecular and morphological character sets (Bloomer & Crowe 1998); (2) relationships within Alectoris, as suggested by allozymes and mtDNA cytochrome b sequences (Randi 1996); (3) a close relationship of Afropavo with Pavo as suggested by cytochrome b and partial CR sequences (Kimball et al. 1997); (4) the relationships within the New World Quails (Callipepla) as suggested by cytochrome b sequences (Zink and Blackwell 1998).

The results presented in this paper are part of an ongoing project to study phylogenetic relationships and the population and conservation genetics of Galliformes at the Italian Institute of Wildlife Biology, Ozzano dell’Emilia (BO), Italy, in collaboration with the World Pheasant Association (WPA), T. Crowe and coll. (Percy FitzPatrick Institute, University of Cape Town), and Rebecca Kimball and coll. (Department of Biology, University of New Mexico, Albuquerque), who are deeply acknowledged for sharing tissue and DNA samples. We thank the many members of WPA, in particular P. Garson, A. Hennache. G. Robbins, R. Sumner and H. Hassink, for collaborating with sample collection. This work was supported by the Italian Institute of Wildlife Biology.

Baker, A.J. & Marshall, H.D. 1997. Mitochondrial control region sequences as tools for understanding evolution. In: Mindell, D. P. (ed) Avian molecular evolution and systematics. San Diego; Academic Press: 51-82.

Bloomer, P. & Crowe, T.M. 1998. Francolin phylogenetics: Molecular, morphobehavioral, and combined evidence. Molecular Phylogenetics and Evolution 9: 236-254.

Brown, G.G., Gadaleta, G., Pepe, G., Saccone, C. & Sbisà, E. 1986. Structural conservation and variation in the D-loop containing region of vertebrate mitochondrial DNA. Journal of Molecular Biology 192: 503-511.

Clayton, D.A. 1992. Transcription and replication of animal mitochondrial DNAs. In Wolstenholme, D.R. and Jeon, K.W. (eds) Mitochondrial Genomes. San Diego; Academic Press: 217-323.

Desjardins, P. & Morais, R. 1990. Sequence and gene organisation of the chicken mitochondrial genome. A novel gene order in higher vertebrates. Journal of Molecular Biology 212: 599-634.

Felsenstein, J. 1985. Confidence limits on phylogenies: An approach using the bootstrap. Evolution 39: 83-791.

Fumihito, A., Miyake, T., Sumi, S-I., Takada, M., Ohono, S. & Kondo, N. 1994. One subspecies of the red jungle fowl (Gallus gallus gallus) suffices as the matriarchic ancestor of all domestic breeds. Proceedings of the National Academy of Science, USA 91: 12505-12509.

Fumihito, A., Miyake, T., Takada, M., Ohono, S. & Kondo, N. 1995. The genetic link between the Chinese bamboo partridge (Bambusicola thoracica) and the chicken and junglefowls of the genus Gallus. Proceedings of the National Academy of Science, USA 92: 11053-11056.

Gerloff, U., Schlotterer, C., Rassmann, K., Rambold, I., Hohmann, G., Frutth, B. & Tautz, D. 1995. Amplification of hypervariable simple sequence repeats (microsatellites) from excremental DNA of wild living Bonobos (Pan paniscus). Molecular Ecology 4: 515-518.

Gilbert, D.G. 1992. LOOPDLOOP, a Macintosh program for visualizing RNA secondary structure. Ftp.bio.indiana.edu.

Kidd, M. G. & Friesen ,V.L. 1998. Sequence variation in the guillemot (Alcidae: Cepphus) mitochondrial control region and its nuclear homolog. Molecular Biology and Evolution 15: 61-70.

Kimball, R.T., Braun, E. L. & Ligon, D.J. 1997. Resolution of the phylogenetic position of the Congo peafowl, Afropavo congensis: A biogeographic and evolutionary enigma. Proceedings of the Royal Society, London B 264: 1517-1523.

Kumar, S., Tamura, K. & Nei, M. 1993. MEGA: Molecular Evolutionary Genetic Analysis, Version 1.01. Pennsylvania State University.

Lucchini, V., & Randi, E. 1988. Mitochondrial DNA sequence variation and phylogeographic structure of rock partridge (Alectoris graeca) population. Heredity (in press.)

Palva, T.K. & Palva, E.T. 1985. Rapid isolation of animal mitochondrial DNA by alkaline extraction. FEBS Letters 192: 267-270.

Quinn, T.W. & Wilson, A.C. 1993. Sequence evolution in and around the mitochondrial control region in birds. Journal of Molecular Evolution 37: 417-425.

Randi, E. 1996. A mitochondrial cytochrome b phylogeny of the Alectoris partridges. Molecular Phylogenetics and Evolution 6: 214-227.

Randi, E. & Lucchini, V. 1998. Organisation and evolution of the mitochondrial DNA control-region in the avian genus Alectoris. Journal of Molecular Evolution (in press.)

Saitou, N. & Nei, M. 1987. The neighbor-joining method: A new method for reconstructing phylogenetic trees. Molecular Biology and Evolution 4: 406-425.

Sbisà, E., Tanzariello, F., Reyes, A., Pesole, G. & Saccone, C. 1997. Mammalian mitochondrial D-loop region structural analysis: Identification of new conserved sequences and their functional and evolutionary implications. Gene 205: 125-140.

Sibley, C.G. & Ahlquist, J.E. 1990. Phylogeny and classification of birds. A study in molecular evolution. New Haven: Yale University Press: 976 pp.

Strimmer, K. & von Haeseler, A. 1996. Quartet puzzling: A quartet maximum likelihood method for reconstructing tree topologies. Molecular Biology and Evolution 13: 964-969.

Swofford, D.L. 1993. PAUP: Phylogenetic Analysis Using Parsimony, Version 3.1. Champaign, Illinois: Natural Histhory Survey.

Thompson, J.D., Gibson T.J., Plewniack, F., Jeanmougin, F., & Higgins, D.G. 1997. The CLUSTAL X windows interface: Flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acid Research (in press.)

Zink, R.M. & Blackwell, R.C. 1998. Molecular systematics and biogeography of aridland gnatcatchers (genus Polioptila) and evidence supporting species status of the California gnatcatcher (Polioptila californica). Molecular Phylogenetics and Evolution 9: 26-32.