Integration Of Molecular And Morphological Data For Extant Clades

Knowledge of Mesozoic mammals has increased enormously since the publication of Lillegraven, Kielan-Jaworowska, and Clemens's book (1979), resulting in a wider range of taxonomic diversity, new records in previously blank geographic areas, and better documentation of anatomical characteristics. The morphological data for Mesozoic taxa are far more extensive than they were two decades ago, and recent studies have resulted in explicit, comprehensive hypotheses of relationships among most major groups of Mesozoic mammals. But new problems have arisen, especially in the area of molecular versus mor phological estimates for the timing and sequence of origination for extant mammalian clades having their roots in the Mesozoic.

Rapid recent advances in molecular evolutionary studies have made it possible to estimate the sequence and timing of the splits of major extant mammalian lineages, based on the assumption of a molecular clock. This has presented a welcome opportunity for molecular evolutionists to challenge the widely accepted outline for Meso-zoic mammal evolution as established by paleontologists and morphologists (e.g., Kumar and Hedges, 1998; Murphy, Eizirik, Johnson et al., 2001; Murphy, Eizirik, O'Brien, et al., 2001). Conversely, improvements in the fossil record have allowed paleontologists to test molecular hypotheses on early mammalian evolution by reference to the fossil record of the Mesozoic (see, e.g., Benton, 1999; Foote et al., 1999; Novacek, 1999; and compare with Kumar and Hedges, 1998; Esteal, 1999; Gee, 1999). Morphologists have countered that assumption of constancy of rate in molecular evolution, on which the earlier time estimates for divergence were established, are not necessarily reliable. It is also possible that calculated rates are biased, owing to calibration on the basis of spurious evidence from the fossil record (e.g., Benton, 1999; Foote et al., 1999).

There are great differences between the molecular trees of modern mammal orders and those based on morphological data. The latest molecular studies of placental mammals achieved a strong consensus on the phylogenies of superordinal clades of placentals. Of the four superordinal clades, Afrotheria (proboscideans, hyracoids, macroscelideans, and some others) represent the earliest-diverging clade, followed by Xenarthra, Laurasiatheria (artiodactyls, perissodactyls, carnivorans, lipotyphlans, and others), and Euarchonta (primates, chiropterans, dermopterans). The grouping of afrotheres and their placement as a basal clade among crown Placentalia by molecular evidence is contradicted by morphological phylogenies for these mammals (e.g., Novacek, 1992a; Fischer and Tassy, 1993; Prothero, 1993; Shoshani and McKenna, 1998; Allard et al. 1999; Asher, 1999). The case study of the ungulate-whale relationship represents another widely discussed example of incongruences in the morphological and molecular evidence. The paraphyly of artiodactyls suggested by molecular evidence has been criticized by morphologists (O'Leary and Geisler, 1999; Gatesy and O'Leary, 2001), although most recently, the finding from Pakistan in 2001 by two teams of paleontologists provided further evidence for the nesting of whales among artiodactyls (Gingerich et al., 2001; Thewissen et al., 2001), as suggested by molecular evolutionists.

Moreover, recent molecular studies have postulated a much earlier diversification of the superordinal placental clades than predicted from fossil records. The current molecular dating suggests that ordinal diversification of living placental groups ranged from 104 to 60 Ma (e.g., Waddell et al., 1999; Eizirik et al., 2001; Madsen et al., 2001; Murphy, Eizirik, Johnson, et al., 2001; Murphy, Eizirik, O'Brien, et al., 2001; but see Kumar and Hedges, 1998), whereas recognizable members of most living orders do not appear until the late Paleocene or early Eocene. A thorough assessment as to whether any of the earliest fossils of eutherians can be placed in one of the superordinal pla-cental groups on the basis of paleontological data is the key to resolving the conflict between molecules and fossils for the time estimate for the earliest placental diversification. There is no doubt that paleontologists will have to continue their difficult task of searching for ever earlier fossils of modern lineages. At the same time the molecular evolutionists might rethink the assumptions and their datasets for their time estimates, especially about the constancy and calibration of the molecular clock.

Relatively speaking, there is less discordance between molecular-based time estimates of marsupial diversification and the appearance of relevant clades of crown Marsupialia in the fossil record. Springer (1997) estimated a divergence of "Ameridelphia" from various groups of Australidelphia between 78 and 70 Ma ago and, within "Ameridelphia," a split of Paucituberculata from "Didelphi-morphia" between 65 and 61 Ma ago. The earliest known fossil taxa (Stagodontidae) that can be placed within the "Ameridelphia" crown group are about 80 Ma old, and the earliest putative member of Paucituberculata (Glasbius) is about 70 Ma old (chapter 12). According to Christian de Muizon, the earliest known South American didelphi-morphs are from the early Paleocene, about 63 Ma old (Marshall and Muizon, 1995; Muizon et al., 1997; Muizon, 1998; Muizon and Cifelli, 2001), although Rougier et al. (1998) placed these Late Cretaceous and Paleocene taxa outside the Marsupialia crown group.

There is better concordance between molecules and fossils in the placement of monotremes within mammals and in the timing of the divergence of monotremes from crown therians. Molecular phylogenies of the relationship of monotremes to other mammals differ depending on the types of molecules and genes examined. Although the debate is ongoing, the balance of evidence clearly favors the traditional hypothesis of a monophyletic therian group to the exclusion of monotremes.

Studies of DNA-DNA hybridization and mitochon-drial genomes suggest that monotremes may be the sister taxon to marsupials, to the exclusion of placental mammals—a relationship advocated by Janke, Kirsch, and their colleagues. This led some molecular evolutionists to argue for the resurrection of Gregory's (1947) classic "Marsupi-

onta"hypothesis (Janke et al., 1996,1997,2002; Penny and Hasegawa, 1997; Kirsch et al., 1997). From a morphological point of view, this hypothesis was too poorly supported to be taken seriously (see discussions by Zeller, 1999a; Szalay and Sargis, 2001; and Luo et al., 2002; among others).

All studies on nuclear genes and most studies on protein sequences have clearly rejected the "Marsupionta" hypothesis, instead uniting placentals and marsupials to the exclusion of Monotremata. The therian hypothesis is supported by an earlier study on protamine DNA, corroborated by studies of neurotrophin genes, ß-globin gene sequences, and the latest studies on imprint genes and retroposons (reviewed by Killian et al. 2000, 2001). A recent comparative study on the nuclear versus mitochondrial genomes by Springer et al. (2001) shows that the nuclear exon genes are more informative than the protein-coding mitochondrial genes for resolving deep mammal phylogeny, weakening the relative strength of the mito-chondrial evidence for recovering higher-level phylo-genetic history in mammals, at least for the case of deep phyletic splits of monotremes, marsupials, and placentals.

Although an earlier study on the combined protein sequences yielded an unresolved trichotomy for the three groups of living mammals (Czelusniak et al., 1990), the more recent evidence from the protamine protein sequence, the a-lactalbumin protein sequence, and the IgM proteins excluded monotremes from a monophyletic Theria. In summary, although some types of molecular data have provided limited support for the "Marsupionta" hypothesis, the prevailing evidence from nuclear genes, retroposons, and protein sequences favors the traditional hypothesis of a monophyletic crown Theria (Mammalia of Linnaeus).

The study of molecular evolutionary rates by a-lact-albumin protein and IgM protein sequences suggests that modern monotremes split from crown therians about 170 to 168 Ma ago (Messer et al., 1998; Belov et al., 2002). This molecular time estimate is in concordance with the latest morphological hypothesis incorporating relevant fossils. Recent analyses place the Middle Jurassic Ambondro and the Middle-Late Jurassic Asfaltomylos as stem taxa related to monotremes, with shuotheriids (also of Middle Jurassic age) as a sister group to australosphenidans (Luo, Kielan-Jaworowska, and Cifelli, 2001; Luo et al., 2002; Rauhut et al., 2002; Kielan-Jaworowska, Cifelli, and Luo, 2002). As such, the split sequence and time estimate by molecules are in good agreement with the morphological hypothesis based on fossils.

The mammalian evolutionary tree is fundamental to comparative mammalian genomics, as differences and similarities in the gene and protein sequences of diverse modern mammalian species are accumulated through their evolutionary history. The cross illumination of molecular evolutionary biology and mammalian paleontology can only enhance our understanding of the deep evolutionary history of mammals.

0 0

Post a comment