23

Features

February 2014 © Biochemical Society

Molecular Evolution

Abbreviations: CNE, cis-non-coding element; hpf, hours post-fertilization; SINE, short interspersed repetitive element; TE, transposable element

Key words: actinodin gene, cis-non-coding element, sarcopterygian, teleost, tetrapod, transposable element

Mariner4 who was the sole survivor of his ship’s crew, the living coelacanth is the sole survivor of an ancient lineage of fishes that had literally gone in absentia for 70 million years. We now know that there are two closely related species of coelacanths (Latimeria), both from the Indian Ocean: one from the south-east coast of Africa and the other from Indonesia, and that both are rare and endangered. As a child growing up in rural Hawaii, I read Old Four Legs5, Smith’s account of the living coelacanth’s discovery and his efforts to procure a second specimen; immediately I became smitten by this large prehistoric-looking fish with lobed-fins (Figure 1), an obsession that would continue throughout my life. We published our first coelacanth paper 20 years ago6, although I could never have fathomed at the time that its genome would eventually be sequenced and that we would have a chance to peer into its evolutionary history or to be able to do biology with this creature, albeit from the inside-out. A few of the more salient biological implications of the coelacanth genome are discussed below, particularly with reference to evolutionary transitions.

Biological implications of the coelacanth genome

An Ancient MarinerThe coelacanth is an iconic species that has captured the public’s imagination owing to its unusual appearance and its special history. The fish possesses paired fins resembling limb-like structures and other features that have placed it in an outgroup position to the terrestrial vertebrates. The recent sequencing of the coelacanth genome has enabled insights into the fin–limb transition and the origin of terrestriality. Here, I discuss some of the more salient findings and interpret them in an evolutionary context.

Charles Darwin was a brilliant naturalist who recognized seemingly disparate connections in the natural world; he saw evolutionary relationships among forms living and extinct, over wide geographical barriers, and he used powerful deductive reasoning and logic to establish the now obvious parallels of natural selection and the principles underlying domestication of plants and animals. Well before the publication of On the Origin of Species1, Darwin speculated on the mechanism by which evolutionary change occurred: “With belief of transmutation & geographical grouping we are led to endeavour to discover causes of change, the manner of adaptation... instinct & structure become full of speculation & line of observation”2. Darwin today would no doubt revel in our ability to draw biological inferences from genome sequences, especially for an organism such as the coelacanth, which holds a special place as a living ‘transitional’ form*.

The story of the living coelacanth is a remarkable one, filled with intrigue, disbelief, pathos and even geopolitical posturing3. Much like Coleridge’s Ancient

Chris T. Amemiya (Benaroya Research Institute and the University of Washington, USA)

*Darwin discusses the nagging problem of ‘Transitional Varieties’ in Chapter VI of On the Origin of Species2. Although the term is usually

referred to in a paleontological context, the living coelacanths and the six extant species of lungfish represent long evolutionary

lineages that are unique in their departure points in-between the ray-finned fishes (like zebrafish) and the tetrapods (like us).

24

Features

February 2014 © Biochemical Society

Molecular Evolution

Figure 1. African coelacanth photographs. (a) An African coelacanth and diver in a cavern off the coast of South Africa. Note the lobed fins and prominent scales. (b) A facial shot of Sydney, one of the members of the Sodwana Bay population of coelacanths. Sydney is a coelacanth named in honour of Dr Sydney Brenner, who is originally from South Africa and an early proponent of sequencing of the coelacanth genome. (c and d) Close-up views of Sydney’s pectoral and pelvic fins respectively. The paddle-like appearance of the fins was the impetus for Smith’s term, ‘Old Four Legs.’

Why sequence the coelacanth genome?

The overarching reason for sequencing the genome of the coelacanth is that it represents a key evolutionary departure point in-between the ray-finned fishes (such as carp and salmon) and the tetrapods (amphibians, reptiles, birds and mammals), a position occupied by only two extant lineages, one of which (lungfish) is particularly difficult to study due to the enormity of its genome. As a prelude to genome sequencing, our studies had shown, importantly, that selected coelacanth gene families were highly conserved in terms of their genomic organization and that their rates of molecular evolution were comparatively lower than the same genes in tetrapods7,8. These findings augured well for subsequent genome sequencing of the coelacanth, and supported the contention that it would be well suited to ‘informing’ aspects of tetrapod biology, namely the molecular changes that may have occurred during adoption of a terrestrial environment.

Evolutionary transitions: genomic losses and gains

Comparisons of genome sequences from different groups of species can provide candidate genes (or regulatory

elements) that have been lost or gained during evolutionary divergence in the respective lineages. Two different scenarios are given in Figure 2 for genomic components lost or gained among three different vertebrate lineages (teleost fishes, coelacanths and tetrapods). As depicted by the Venn diagrams, the patterns of losses and gains can implicate putative elements that may have been of importance in evolutionary processes.

In Figure 2(a), a scenario is given in which specific sequences have been lost in the lineage leading to tetrapods, but are retained among teleosts and coelacanths. Our analysis detected at least 50 genes that satisfied this criterion9, and included genes involved in development of the fins, eye, ear, kidney, brain and axial structures (trunk, somites and tail). How could loss of such genes facilitate or contribute to the evolution of terrestriality? One idea is that certain fish-specific structures are maladaptive in a terrestrial environment and that their losses led to evolution of structures better suited to life on land. Teleological arguments could be made to this end for several of the 50 genes lost among tetrapods; however, without empirical evidence, these remain ad hoc hypotheses. One particularly intriguing example for which there is some supporting evidence is provided by the actinodin genes.

The coelacanth has a single actinodin (And) gene, whereas teleost fishes have multiple copies due to ancestral gen(om)e duplications; there are no And genes in tetrapods. The And genes are involved in the formation of fibre-like collagenous structures in developing fins of fish and are necessary, ultimately, for the development of the soft bony fin rays, the lepidotrichia. Certainly, the fins of fish are extremely well-adapted to life in the aquatic environment and are used in many ways in addition to locomotion. A compelling study by Akimenko and colleagues showed that knockdown of two of the And genes in zebrafish results in structural disarray in the developing pectoral fin bud and subsequent misexpression of developmental genes long implicated in the fin–limb transition10. The overarching inference from these studies is that loss of actinotrichia and fin rays on the paired appendages during the evolution of tetrapods may have enabled the development of the endochondral skeletal structures that are necessary for terrestrial locomotion and land dwelling.

In contrast with the loss of genes during evolution, a scenario is given in Figure 2(b) in which those sequences that are newly arisen in the lineage that led to the lobe-finned vertebrates or sarcopterygians, are present in coelacanth and tetrapods, yet absent from teleosts. Although a definitive analysis has yet to be carried out for protein-coding genes, such an analysis has been carried out for the so-called cis-non-coding elements, or CNEs. These are stretches of DNA that

25

Features

February 2014 © Biochemical Society

Molecular Evolution

facilitators of developmental neuronal plasticity17–19.Well before the coelacanth genome project, Bejerano

et al.20 computationally analysed the small amount of Latimeria genomic data that had been deposited in GenBank®, and showed that the origin of a particular ‘ultraconserved’ element in the human genome could be traced to the lobe-finned fishes. This short interspersed repetitive element (SINE), importantly, was still conserved in the coelacanth and found in high copy number; whereas in humans and other tetrapods, this family had degenerated beyond recognition except in those instances where it had become ‘exapted’ to new functions in both coding and non-coding sequences. The authors concluded that the coelacanth, by preserving what in other species would be transient TE families, might serve as a ‘living molecular fossil’ and that the coelacanth genome could very well hold relicts of additional events that helped shape our own evolution. Indeed, additional studies showed similar patterns for two other coelacanth TEs: a SINE21,22 and a Harbinger-type transposon23. Thus it was with great anticipation when the entire Latimeria genome sequence became available, as it was assumed that it would provide a veritable treasure trove of evolutionary information on TEs.

The analysis of TEs in the coelacanth genome has since been extensively carried out and the initial prediction that the genome would consist of very high levels of TEs proved incorrect. It is now known that

are found by virtue of both their conserved sequence and proximity relative to genes in multiple sequence alignments. A substantial fraction of the vertebrate CNEs detected in this manner is enhancer elements that activate transcription11,12. Moreover, it is thought that cis-regulation by transcription factors has played an integral role in effecting evolutionary change13–15. Thousands of CNEs were detected that arose in the sarcopterygian lineage, i.e. after the divergence of coelacanths and tetrapods from the teleost fishes. Subjecting these datasets to gene ontology statistical analysis allows rough determination of the biological nature of the genes that these CNEs regulate. Among the highest scoring categories were those associated with embryonic morphogenesis, differentiation, regulation of transcription, sensory organ development, brain and endocrine system development, and limb development.

One of these newly arisen CNEs was found in a region upstream of the HOX-D homeobox gene cluster (Figure 3). This region contained six CNEs that were shown previously to consist of enhancer elements (blue ovals in Figure 3a). Previous analyses of these six enhancers using a mouse reporter transgene assay gave distinct patterns of expression in the developing forelimb (autopod) as shown in the illustration below the map. One of the enhancers (I1) was found in coelacanths and tetrapods, but not in teleost fish and thus represents a new sarcopterygian enhancer. Validation experiments using the coelacanth enhancer in the transgenic mouse reporter assay gave an expression pattern similar to that seen previously with the mouse version (Figure 3b). The evolutionary implication from this analysis is that, whereas the autopod is considered a tetrapod invention, an enhancer element that drives autopodial expression is already present in the genome of a lobe-finned fish. From these preliminary analyses, it is impossible to know to what extent such an element has played in the fin–limb transition as the observations are largely correlative. However, these kinds of experiments set the stage for further more hypothesis-driven research into the rudiments of tetrapod appendages.

Repository of transcriptionally active transposable elements

Transposable elements (TEs) are curious genetic sequences that have the ability to undergo mobilization within the genome, either by self-contained genetic machinery or via assistance from other intact TEs. It is estimated that the human genome comprises ~44% TEs, many of which have degenerated and are no longer mobile16. TEs are thought to influence myriad biological processes, including the regulation and diversification of the genome, and can act as mutagenic agents and

Figure 2. Inclusion of the coelacanth genome in comparative genome analysis can identify genes (and conserved non-coding sequences) that have been lost (red) or gained (green) in vertebrate evolution. (a) Loss of genes (and CNEs) from the tetrapod lineages, i.e. these sequences were present in teleost fishes and coelacanth, but absent (lost) in tetrapods. (b) Gain of genes (and CNEs) in the lobe-finned vertebrates (coelacanths plus tetrapods).

26

Features

February 2014 © Biochemical Society

Molecular Evolution

Figure 3. Identification of a coelacanth enhancer element that drives expression in a limb-specific manner. Long-range multiple alignments were computationally derived for several vertebrate genomes using the multi-Z tool and filtered extensively to identify both genes and CNEs, allowing delineation of lineage-restricted sequence patterns as per Figure 2(b). The region in (a) represents 1 Mb of the mouse HOX-D locus and surrounding area. Genes are labelled and depicted as grey boxes, and previously identified enhancers are shown as blue ovals. Enhancer-driven expression patterns of mouse transgenic reporter constructs are depicted in the illustrations below the map for the mouse limb bud (forelimb). One of these enhancers, I1, is notable in that it is shared by both tetrapods and coelacanths, but not by teleost fishes. (b) The coelacanth version of this enhancer element drives expression in the mouse limb bud in a similar manner to the mouse enhancer, and suggests that this is a sarcopterygian innovation. Modified from Figure 2 of the study in 2013 by Amemiya et al.9

TEs comprise approximately 25% of the coelacanth genome, lower than that for humans. However, more importantly, the number of different TE families in the coelacanth was shown to be markedly higher than in humans and that surprisingly, 23% of all coelacanth TEs were transcriptionally expressed24,25. Taken together, these observations suggest that the coelacanth genome is transpositionally dynamic, an inference that seems to be at odds with the slower relative rate of coding sequence evolution.

The underlying evolutionary biology of TEs is still largely a black box as many of the assumed biological effects of transposition have been difficult to assess directly. One feature of TEs is that they endow their neighbouring milieu with better chromatin accessibility. This would have effects on gene regulation of neighbouring genes and would certainly have evolutionary implications. One TE in the coelacanth, the LatiHarb1 transposon23, contains a transposase gene and a second gene, whose function is still unknown, but has been exapted into coding sequences of a few other vertebrate genes. This TE alone is found in >5% of the Latimeria genome and is highly conserved within the coelacanth genome. The use of a fluorescent transgene reporter system in zebrafish confirmed that indeed this transposon facilitates transcription in a very general sense and in many tissues (Figure 4), although its exaptation into the regulatory landscape of tetrapods has yet to be assessed. Two of the SINEs in the coelacanths have been shown to be exapted into transcriptional regulatory positions for neuronal genes in the mouse20,21. Clearly, we have just touched the surface insofar as our understanding of the evolutionary influence that TEs have had on tetrapod evolution.

Caveats and future prospects

As with all genome reports, there is the possibility of over-interpretation of the sequence data26. The genome of an extant species represents but one snapshot in time and this snapshot may not necessarily reveal anything tangible about morphological and physiological changes that have occurred over time. In addition, each species has also undergone its own evolution since divergence from a common ancestor, further complicating the analysis. As far as the lobe-finned fishes are concerned, there is a rich history of other fossilized forms (including ‘fishapods’) whose genomes, if they could be determined, would certainly be better suited than Latimeria for understanding the mechanisms involved in the evolutionary transition to a tetrapod bauplan and life history27–29. Genome reports are not meant to be a be-all and end-all evolutionary analysis, but rather a starting point for further biological inquiry and hypotheses.

Figure 4. Expression of a transposon-reporter construct in zebrafish. An intact LatiHarb1 transposon was placed upstream of a basal promoter and GFP reporter gene in the p339hsp70GFPrc vector32 and injected into single cell embryos of zebrafish. (a) Embryos at 9 hpf (hours post-fertilization) (late gastrula stage) showing early, although largely unspecific, expression of GFP. (b) Side view of a head of a 32 hpf embryo showing expression primarily in the integument (skin). (c) Side view of an embryo at 32 hpf showing GFP expression in the notochord.

27

Features

February 2014 © Biochemical Society

Molecular Evolution

Although we will never be able to replay evolution, creative experimentation can help us to unravel the rudiments of key steps that undoubtedly were necessary for evolutionary transitions10,30,31. This is our grand challenge in the field of evolutionary developmental biology: to incorporate and reconcile morphology and natural history (including the fossil record) with modern genetics and developmental biology. ■I thank Dr Jessica Alföldi, Dr Axel Meyer and Dr Kerstin Lindblah-Toh, and the other members of the Coelacanth Genome Sequencing Consortium for working together to sequence and analyse the coelacanth genome. All photographs of the African coelacanth were taken by Laurent Ballesta during mixed-gas deepwater dives in Jesser Canyon, Sodwana Bay, South Africa; these photographs are used with permission and are not to be duplicated and/or distributed without the express consent of Laurent Ballesta (www.coelacanthe-projet-gombessa.com). Dr Nicolas Rohner contributed the picture of the mouse forelimb expression with the coelacanth enhancer-reporter construct. Additional data discussed here were generated by Dr Igor Schneider, Dr Byrappa Venkatesh and Dr Jeramiah Smith.

1. Darwin, C. (1859) On the Origin of Species by Means of Natural Selection, Random House, New York

2. Darwin, C. (1838) Notebook B: Transmutation (1837–1838), Cambridge University Press, Cambridge. Also available at: http://darwin-online.org.uk

3. Weinberg, S. (2000) A Fish Caught In Time: The Search For The Coelacanth, HarperCollins, New York

4. Coleridge, S.T. (1883) The Rime of The Ancient Mariner, University Press, Cambridge

5. Smith, J.L.B. (1956) Old Fourlegs: The Story of the Coelacanth, Longman Green, London

6. Amemiya, C.T., Ohta, Y., Litman, R.T., Rast, J.P., Haire, R.N. and Litman, G.W. (1993) Proc. Natl. Acad. Sci. U.S.A. 90, 6661–6665

7. Amemiya, C.T., Powers, T.P., Prohaska, S.J. et al. (2010) Proc. Natl. Acad. Sci. U.S.A. 107, 3622–3627

References

Chris Amemiya is a Full Member at the Benaroya Research Institute and Professor of Biology at the University of Washington, Seattle, Washington USA. His lifelong fascination with natural history was fostered by being raised in Hawaii, a

natural laboratory for evolution, and by a father who was an avid fisherman. His laboratory studies developmental mechanisms and the evolution of vertebrates, with an eye for novel strategies that organisms have acquired to deal with their internal and external environments. email: camemiya@benaroyaresearch.org

8. Noonan, J.P., Grimwood J., Danke J. et al. (2004) Genome Res. 14, 2397–2405

9. Amemiya, C.T., Alfoldi J., Lee, A.P. et al. (2013) Nature 496, 311–316

10. Zhang, J., Wagh P., Guay D. et al. (2010) Nature 466, 234–237

11. Pennacchio, L.A. and Rubin, E.M. (2001) Nat. Rev. Genet. 2, 100–109

12. Visel, A., Bristow, J. and Pennacchio, L.A. (2007) Semin. Cell Dev. Biol. 18, 140–152

13. Carroll, S.B., Grenier, J.K. and Weatherbee, S.D. (2001) From DNA to Diversity: Molecular Genetics and The Evolution of Animal Design, Blackwell Science, Malden

14. Carroll, S.B. (2005) PLoS Biol. 3, e24515. Shapiro, M.D., Marks, M.E., Peichel, C.L. et al. (2004) Nature

428, 717–72316. Lander, E.S., Linton, L.M., Birren, B. et al. (2001) Nature 409,

860–92117. Thomas, C.A. and Muotri, A.R. (2012) Front. Biosci. (Elite

Ed.) 4, 1663–166818. Bundo, M., Toyoshima, M. and Okada, Y. (2014) Neuron,

doi:10.1016/j.neuron.2013.10.05319. Kazazian, Jr, H.H. (2004) Science 303, 1626–163220. Bejerano, G., Lowe, C.B., Ahituv, N. et al. (2006) Nature 441,

87–9021. Nishihara, H., Smit A.F. and Okada, N. (2006) Genome Res.

16, 864–87422. Xie, X., Kamal, M. and Lander, E.S. (2006) Proc. Natl. Acad.

Sci. U.S.A. 103, 11659–1166423. Smith, J.J., Sumiyama, K. and Amemiya, C.T. (2012) Mol.

Biol. Evol. 29, 985–99324. Forconi, M., Chalopin, D. and Barucca, M. et al. (2013) J.

Exp. Zool. B Mol. Dev. Evol., doi:10.1002/jez.b.2252725. Chalopin, D., Fan, S., Simakov, O., Meyer, A., Schartl, M. and

Volff, J.N. (2013) J. Exp. Zool. B Mol. Dev. Evol., doi:10.1002/jez.b.22521

26. Zimmer, C. (2010) Yet-another-genome-syndrome. Discover Magazine, The Loom Science Blog. Also available at: http://blogs.discovermagazine.com/loom/2010/04/02/yet-another-genome-syndrome/#.UtZelrQbGg8

27. Daeschler, E.B., Shubin, N.H. and Jenkins, Jr, F.A. (2006) Nature 440, 757–763

28. Shubin, N.H., Daeschler, E.B. and Jenkins, Jr, F.A. (2006) Nature 440 764–771

29. Clack, J.A. (2012) Gaining Ground: The Origin and Evolution of Tetrapods (2nd ed.), Indiana University Press, Bloomington

30. Abitua, P.B., Wagner, E., Navarrete, I.A. and Levine, M. (2012) Nature 492, 104–107

31. Carroll, S.B. (2005) Endless Forms Most Beautiful: The New Science of Evo Devo and the Making of the Animal Kingdom. W.W.Norton, New York

32. Mongin, E., Auer, T.O., Bourrat, F. et al. (2011) PLoS One 6, e19747