COMMENTARY

microRNAs revive old views about jawless vertebrate divergence and evolution

Philippe Janvier1 Centre National de la Recherche Scientifique, Unité Mixte de Recherche 7207, Muséum National d’Histoire Naturelle, 75231 Paris, France

ost living vertebrates are hopeful third source of data [that is, mi- jawed vertebrates or gna- croRNAs (miRNAs), which are highly M thostomes, like sharks or conserved, noncoding genes that can be mammals. However, two treated in datasets as presence/absence, small fish groups, hagfishes (67 species) like most phenotypic characters, and ana- and lampreys (about 40 species), are de- lyzed with the same parsimony algorithms, void of jaws and are either ignored by with the additional advantage of being the public or regarded as pests. Both are rarely lost in evolution. Moreover, eel-shaped, cartilaginous, and scale-less, miRNAs are regulatory genes that are and their overall anatomy is roughly simi- strongly involved in cell differentiation lar to that of other fishes, with a brain, and thus play a key role in organogenesis spinal cord, and sensory capsule, although during development. Although the use of more simple than that of gnathostomes miRNAs for resolving deep divergences (Fig. 1). First regarded as intestinal worms between major animal groups is still in or degenerate jawed vertebrates (1) and its infancy, one can foresee that it will then gathered into a group, the cyclo- soon provide a powerful source of data for stomes (rounded mouth), on the basis of elucidating phylogenetic patterns and basic anatomical characteristics (horny throw light on the developmental and teeth, pouch-shaped gills, single nostril, metabolic processes involved during major and lack of paired fins) (2), these fishes evolutionary divergences and phenotypic became finally considered in evolutionary transformations. The authors (11) show times as a monophyletic offshoot of that the expression profiles of a number primitive vertebrates, sister group to jawed of miRNAs in various organs of the em- vertebrates (3), and this became the re- bryos of lampreys and a jawed vertebrates ceived view for most of the of the 20th (zebrafish) are basically similar, thereby century (4). supporting their homology (some of their In the late 1970s and with the rise of the homologs in hagfishes are present, but cladistic principles for assessing phyloge- their expression profiles are still un- netic relationships on the basis Fig. 1. Evidence for monophyly of the cyclostomes known). They also provide the most ex- (hagfishes and lampreys) based on microRNAs dis- of phenotypic characters, it was suggested fi tensive critical review of the phenotypic tribution and expression pro les among verte- (anatomical and physiological) characters that cyclostomes might be paraphyleytic; brates deprives phylogeneticists from the possibility that is, they are a basal grade in the ver- to explore the stepwise assembly of the vertebrate that were long used for supporting cyclo- tebrate tree, with lampreys being more body plan on the sole basis of phenotypic characters stome paraphyly and show that many of closely related to gnathostomes than to of these living jawless vertebrates. The pattern of them were spurious, wrongly coded, or hagfishes (5). This phylogenetic pattern the chordate tree now returns to what it was over inapplicable to the putative outgroups implied that cyclostomes could throw light a century ago, except for the relationships of extinct of the vertebrates. Their work notably on the early steps of the assembly of the groups, such as the jawless ostracoderms (e.g., het- shows that hagfishes and lampreys share vertebrate body plan and that hagfishes erostracans and osteostracans) and the jawed pla- four unique miRNA families, which is could document the most generalized coderms. The investigations by Heimberg et al. enough to provide the strongest possible (11) on microRNAs suggest that the cyclostomes, condition for a number of vertebrate particularly hagfishes, underwent a spectacular de- support to cyclostomes monophyly (Fig. characters (6). However, it soon raised generacy since their divergence from other verte- 1). The results of Heimberg et al. (11) heated debates, because an increasingly brates (that is, the loss of numerous phenotypic are certainly the most convincing contri- large number of molecular sequence data characters) and that the most recent common an- bution ever published in support of cyclo- provided increasingly strong support for cestor of the vertebrates was probably more com- stomes monophyly, and we can foresee the old theory of cyclostomes monophyly plex than the living cylostomes. This discovery may that they will soon be complemented by (7, 8); that is, hagfishes and lampreys again question the status of the ostracoderms as miRNA expression profiles from the † were actually sister groups that had di- jawless stem gnathostomes. , Palaeozoic groups. long elusive, but now available, hagfish verged in the early Paleozoic, up to 500 embryos (12). million years (Myr) ago. Morphologists of cladistics (9). Clearly, a third indepen- Although I was among the early sup- who defended cylostome paraphyly argued dent source of characters was needed porters of vertebrate paraphyly (6, 7), I am that molecular sequence-based trees were to resolve this conflict, and much hope impressed by the evidence provided by inconclusive because of the uncertainty was put in developmental data for de- Heimberg et al. (11) and prepared to as to the outgroups of the vertebrates (i.e., ciding whether hagfishes were actually their closest relatives and either tunicates primitive and had lost many characteristics of amphioxus) (Fig. 1) or because since their divergence from other verte- Author contributions: P.J. wrote the paper. of biases in the methods of molecular brates (10). The author declares no conflict of interest. tree reconstruction that were at odds with The report by Heimberg et al. (11) in See companion article on page 19379. the original standard parsimony methods PNAS provides the application of this 1E-mail: [email protected].

www.pnas.org/cgi/doi/10.1073/pnas.1014583107 PNAS | November 9, 2010 | vol. 107 | no. 45 | 19137–19138 Downloaded by guest on September 24, 2021 admit that cyclostomes are, in fact, erostracans or osteostracans (Fig. 1), but and dentine) are assumed to have arisen monophyletic. The consequence is that are now regarded as stem gnathostomes in vertebrates about 500–480 Myr ago they may tell us little, if anything, about (14, 15); that is, they are more closely re- among the jawless stem gnathostomes, the dawn of vertebrate evolution, except lated to jawed vertebrates that to cyclo- such as conodonts and early ostracoderms that the intuitions of 19th century zoolo- stomes. These authors even alluded to (15). The earliest presumed Paleozoic gists were correct in assuming that these the possibility that all jawless vertebrates, vertebrate fossils, such as the Cambrian odd vertebrates (notably, hagfishes) are fossil and recent, could have been derived, myllokunmingiids (Fig. 1) (16), and the strongly degenerate and have lost many through many character losses, from a characters over time (2, 4). This is an un- later presumed soft-bodied vertebrates, precedented case of character loss among fi such as euconodonts and euphaneropids living vertebrates. Unfortunately, fossil Hag shes and lampreys (13, 14), cannot be clearly proven to be cyclostomes are of no help in documenting share four unique stem cyclostomes. Nor is there any evi- this degeneracy. They are mere soft tissue dence left that cyclostomes are derived imprints dated to 360 Myr for lampreys miRNA families. from any ostracoderm group through and possibly, 300 Myr for hagfishes, and loss of characters. The contribution by they are tentatively interpreted as roughly Heimberg et al. (11) perhaps marks the similar to modern forms. common ancestor that was morphologi- end of the dream that, in common with the Heimberg et al. (11) conclude that the cally more similar to a jawed than a jawless phylogeny of other groups, living jawless latest common ancestor of all vertebrates vertebrate (13). vertebrates could document the stepwise may have been phenotypically more com- Current vertebrate phenotype-based rise of the vertebrate body plan. However, plex than living cyclostomes. Interestingly, phylogenies nevertheless assume that the this echoes the opinion expressed long lack of jaws and mineralized tissues is although it brings us back to very old ago by some paleontologists (13), who primitive for vertebrates (14, 15) and evolutionary views, it has the merit to show supported the theory that lampreys and that cyclostomes, be they monophyletic or that systematic theories are refutable, all hagfishes were derived from heavily ar- not, diverged in the Early Palaezoic from the more when character losses can be mored and ossified Paleozoic jawless fishes a naked and soft-bodied common verte- justified by biological processes and are, in referred to as ostracoderms, such as het- brate ancestor. Mineralized tissues (bone turn, experimentally refutable.

1. Linnaeus C (1758) The System of Nature Through the 7. Forey PL, Janvier P (1993) Agnathans and the origin of terrelationships of hagfish, lampreys, and gnathos- Three Kingdoms of Nature. Animal Kingdom (Laurentii jawed vertebrates. Nature 361:129–134. tomes and the nature of the ancestral vertebrate. Salvii, Stockholm). 8. Delarbre C, Gallut C, Barriel V, Janvier P, Gachelin G Proc Natl Acad Sci USA 107:19379–19383. 2. Duméril AMC (1812) Dissertation on the Cyclostome (2002) Complete mitochondrial DNA of the hagfish, 12. Ota KG, Kuraku S, Kuratani S (2007) Hagfish embryol- Fish Family Aimed at Demonstrating Their Relation- Eptatretus burgeri: The comparative analysis of mito- ogy with reference to the evolution of the neural crest. ships to Invertebrate Animals (Didot, Paris). chondrial DNA sequences strongly supports the cyclo- Nature 446:672–675. 3. Haeckel E (1866) General Morphology of Organisms stome monophyly. Mol Phylogenet Evol 22:184–192. 13. Jarvik E (1980) Basic Structure and Evolution of Verte- (Georg Reimer, Berlin). 9. Near TJ (2009) Conflict and resolution between phylog- brates (Academic, London), Vol 1. 4. Yalden D (1985) Feeding mechanisms as evidence for enies inferred from molecular and phenotypic data 14. Janvier P (2008) Early jawless vertebrates and cyclo- cyclostome monophyly. Zool J Linn Soc 84:291–300. sets for hagfish, lampreys, and gnathostomes. J Exp stome origins. Zoolog Sci 25:1045–1056. 5. Løvtrup S (1977) The Phylogeny of Vertebrata (Wiley, Zoolog B Mol Dev Evol 312:1–13. 15. Donoghue PCJ, Forey PL, Aldridge RJ (2000) Conodont New York). 10. Janvier P (2007) Evolutionary biology: Born-again hag- affinity and chordate phylogeny. Biol Rev Camb Philos 6. Janvier P (1981) The phylogeny of the Craniata, with fishes. Nature 446:622–623. Soc 75:191–251. particular reference to the significance of fossil ’agna- 11. Heimberg AM, Cowper-Sal·lari R, Sémon M, Donog- 16. Shu DG, et al. (2003) Head and backbone of the early thans’. J Vert Paleont 1:121–159. hue PCJ, Peterson KJ (2010) microRNAs reveal the in- Cambrian vertebrate Haikouichthys. Nature 421:526–529.

19138 | www.pnas.org/cgi/doi/10.1073/pnas.1014583107 Janvier Downloaded by guest on September 24, 2021