Armadillos, anteaters, and sloths (Xenarthra) Frédéric Delsuca,b,* and Emmanuel J. P. Douzerya,b the three xenarthran lineages: some grouped armadil- aUniversité Montpellier II, Place Eugène Bataillon, 34095 Montpellier los and sloths together (5, 6), whereas others suggested Cedex 05, France; bCNRS, Institut des Sciences de l'Evolution (UMR that anteaters are the closest relatives of sloths (7, 8). 5554), CC064-Place Eugène Bataillon, 34095 Montpellier Cedex 05, Early molecular studies based on evolutionary compari- France sons of protein sequences of alpha crystallin-A (9) and *To whom correspondence should be addressed ([email protected]) immunological distances derived from serum albumins (10) did not contribute much to this debate. 7 e most recent classiA cation (4) nevertheless groups anteaters Abstract (Vermilingua) and sloths (Folivora) into a clade called Pilosa, referring to their coat which constitutes a derived Armadillos, anteaters, and sloths (31 sp.) are grouped into character in Xenarthra since it is interpreted as a reversal fi ve families within the mammalian Order Xenarthra. The to the ancestral condition of mammals. xenarthran timetree shows that armadillos diverged from Xenarthran molecular systematics has been conducted anteaters + sloths ~70 million years ago (Ma), which was using a variety of molecular markers and has resulted in a followed by the divergence of anteaters and sloths ~60 well-resolved interfamilial phylogeny (Fig. 2). 7 e earliest Ma. Molecular dating also reveals that the pygmy anteater study analyzed a combination of the mitochondrial 12S diverged from other anteaters ~40 Ma, and that the two liv- and 16S ribosomal RNA (rRNA) genes and the nuclear ing genera of sloths diverged ~20 Ma. Molecular dates sug- exon 28 of the von Willebrand Factor gene (VWF) for eight gest that the paleoenvironmental changes that occurred of the 13 living xenarthran genera (11). Subsequently, during the Cenozoic (66–0 Ma) in South America have taxon sampling was increased to 12 genera in ana- i n fl uenced the evolutionary history of Xenarthra. lyses of three protein-coding nuclear genes, adding the complete ά-2B adrenergic receptor gene (ADRA2B) and Despite their heterogeneous morphologies, armadil- exon 11 of the breast cancer susceptibility gene (BRCA1) los (Cingulata), anteaters (Vermilingua), and sloths to the exon 28 VWF data (12). 7 is nuclear data set (Folivora) form a monophyletic group of curious placen- tal mammals, the Order Xenarthra. Indeed, all living and extinct members share the presence of supplementary intervertebral articulations termed “xenarthrales” in the posterior dorsal vertebrae (1). Extant xenarthrans are currently represented by 31 species of South American origin (2). 7 is reduced taxonomic diversity is the result of the last mass extinction event just 10,000 years ago that wiped out most of the product of an evolutionary radi- ation triggered by the isolation of South America during most of the Cenozoic era, 65–3 Ma (3). Five families are generally recognized: Dasypodidae (armadillos; Fig. 1), Cyclopedidae and Myrmecophagidae (anteaters), and Bradypodidae and Megalonychidae (sloths) (4). Here, we review the phylogenetic relationships and divergence times among the A ve families of extant xenarthrans. Few studies have been dedicated to reconstructing the phylogenetic relationships of xenarthrans. 7 e A rst attempts based on morphological and anatomical char- Fig. 1 A Southern Three-banded Armadillo (Tolypeutes matacus), acters have conP icted regarding the interrelationships of Family Dasypodidae, from Argentina. Credit: F. Delsuc. F. Delsuc and E. J. P. Douzery. Armadillos, anteaters, and sloths (Xenarthra). Pp. 475–478 in e Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009). HHedges.indbedges.indb 447575 11/28/2009/28/2009 11:29:55:29:55 PPMM 476 THE TIMETREE OF LIFE Bradypodidae 4 Megalonychidae 2 Folivora Cyclopedidae Pilosa 3 1 Myrmecophagidae Dasypodidae Vermilingua Cingulata K Paleogene Neogene MZ CENOZOIC 50 25 0 Million years ago Fig. 2 A timetree of armadillos, anteaters, and sloths (Xenarthra). Divergence times are shown in Table 1. Abbreviations: MZ (Mesozoic) and K (Cretaceous). was later enlarged by the addition of two mitochondrial by fossils, they have been classiA ed into distinct families genes (12S rRNA and NADH dehydrogenase 1 [ND1]) (Bradypodidae and Megalonychidae, respectively) (19). (13). Finally, retroposed elements and their P anking 7 is taxonomic distinction is supported by immuno- regions have been studied for representatives of all 13 logical data revealing considerable evolutionary dis- extant genera (14). tance between their albumins (10). Ancient DNA studies 7 e monophyly of Xenarthra is morphologically well based on mitochondrial 12S and 16S rRNA fragments supported by characters generally thought to reP ect adap- from fossil ground sloths also provide some support tations toward fossoriality and myrmecophagy (1). 7 e for the diphyly hypothesis (20, 21). 7 ese studies sug- common ancestry of the three xenarthran lineages was gest that modern three-toed sloths (Bradypodidae) are evidenced in early molecular studies (9, 10). Subsequent closely related to the Shasta Ground Sloth Nothrotheriops sequence-based phylogenetic studies focused on xenar- shastensis (Megatheriidae), whereas two-toed sloths thrans have also found strong statistical support for their (Megalonychidae) appear closer to the Giant Ground monophyly (11, 12), with the notable occurrence of a rare Sloth Mylodon darwinii (Mylodontidae). diagnostic three amino acid deletion in their α-crystallin Within anteaters (Vermilingua), the Pygmy Ant- A protein (15). eater (Cyclopes didactylus) is considered morphologic- 7 e four molecular studies sampling all anteater and ally divergent from the closely related Giant Anteater sloth genera provided unambiguous support for Pilosa (Myrmecophaga tridactyla) and tamanduas (genus Tam- (11–14), conA rming the results of large-scale analyses andua) (7, 22, 23). Based on this morphological diver- including fewer taxa (16–18). 7 ese results contradicted gence, C. didactylus is classiA ed in its own distinct family studies of the ear region (5) and cephalic arterial patterns Cyclopedidae (4). All molecular results conA rm this (6) that supported a basal position of anteaters within arrangement by supporting a close relationship between Xenarthra. 7 e extreme specialization of the skull Myrmecophaga and Tamandua (Myrmecophagidae) toward myrmecophagy in anteaters might have con- (11–14). founded these early morphological studies. Subsequent 7 e newly established phylogenetic framework has cladistic studies of morphological and anatomical fea- been used to derive a molecular timescale for xenar- tures (7) including the structure of the ear region (8) thran evolutionary history. 7 e earliest molecular dating favored Pilosa and provided shared-derived characters study used a maximum likelihood local molecular clock such as the interruption of the zygomatic arch and the approach with three calibration points to calculate diver- intra-pelvic location of the testes. gence dates among eight xenarthran species from amino 7 e two living genera of three-toed (Genus Bradypus) acid sequences of the VWF exon 28 (11). A Bayesian and two-toed (Genus Choloepus) sloths (Folivora) are relaxed molecular clock method approach (24) was sub- strictly arboreal and virtually unknown in the fossil sequently used to analyze a large combination of both record (3). On the basis of numerous morphological dif- mitochondrial and nuclear genes but including only one ferences and a presumably diphyletic origin suggested representative per xenarthran lineage (25). Finally, the HHedges.indbedges.indb 447676 11/28/2009/28/2009 11:29:58:29:58 PPMM Eukaryota; Metazoa; Vertebrata; Mammalia; Xenarthra 477 Table 1. Divergence times (Ma) and their confi dence/credibility intervals (CI) among armadillos, anteaters, and sloths (Xenarthra). Timetree Estimates Node Time Ref. (11) Ref. (25)(a) Ref. (25)(b) Ref. (26) Time CI Time CI Time CI Time CI 170.566.083–4969.978–6281.496–6864.775–55 2 60.0 56.7 71–42 61.4 70–52 66.8 83–52 55.2 65–46 339.739.349–29––––40.049–32 4 19.6 18.3 23–13 – – – – 20.8 28–15 Note: Node times in the timetree represent the mean of time estimates from different studies. Divergence time from ref. (11) is the mean estimates from three calibration points. Times estimates from ref. (25) are based on the analysis of (a) nuclear and (b) mitochondrial genes. The 95% Bayesian credibility intervals are reported for two studies (25, 26). same dating method has also been applied to a combined Eocene of Patagonia and Antarctica (~40 Ma) but they data set of three nuclear genes (ADRA2B, BRCA1, and cannot be precisely assigned to any of the recognized VWF) encompassing 12 of the 13 living genera (26). 7 e lineages (29). 7 e deep molecular estimate conA rms the age of the xenarthran last common ancestor, correspond- considerable divergence between the two modern sloth ing to the separation between Cingulata and Pilosa, was genera and supports their taxonomic distinction at the estimated at 70.5 Ma (Table 1). 7 is estimate, preceding family level. the Cretaceous–Paleogene boundary (Fig. 2), is slightly Finally, the molecular estimates obtained for xenar- more recent than previous ones suggesting a date ~80 thran divergence dates suggested a potential role played Ma (10, 20). It is also more compatible with the A rst by paleoenvironmental changes