Tooth structure re-engineered

Mice deficient in the EDA protein lack normal tooth features. Restoring EDA in embryonic teeth at increasing doses has now been found to recover these dental features in a stepwise pattern that mimics evolution. See Article p.44

A fundamental connection between developmental changes and evolution has long been established1. This link is gaining renewed emphasis2 as molecular studies shed new light on evolution by revealing many genetic modifications that alter developmental processes, in turn changing an organism's shape and structure. On page 44 of this issue, Harjunmaa et al.3 report that, by simply tinkering with the genes and signalling pathways that control the shape of developing teeth, they have remade several different tooth structures in vitro. These structures draw a striking parallel with how teeth evolved from those of distant mammalian ancestors to the teeth of modern-day rodents.

Some lineages of therians (marsupial and placental mammals and their kin) that lived in the Mesozoic era, 252 million to 66 million years ago, had 'tribosphenic' molars4. The taller front end of the tribosphenic lower molar, called the trigonid, had three cusps — the raised points on the crown of the tooth — for shearing food. The lower back end, the talonid, had a basin-like surface for grinding food5,6. The trigonid and talonid of Mesozoic mammals are still recognizable in modern-day rodents, albeit in a highly modified form. As rodents arose from ancestral mammals and diversified into many lineages, cusps that were separate in the Palaeocene epoch 66 million to 56 million years ago7 became progressively connected by crests — which are more effective for chewing plant food — in a 'cusp-to-crest' dental evolution that occurred in many rodent groups8,9.

The ectodysplasin A (Eda) gene encodes a vertebrate signalling protein that is involved in the development of a wide array of structures, from hair to sweat glands10. In the embryonic tooth, the EDA protein is active in enamel knots, which are signalling centres and the precursors of adult tooth structures. EDA regulates the position and size of future tooth cusps and their connecting crests8. Mice that do not express Eda lack normal cusps and crests, and instead have only basic, generalized teeth10.

Harjunmaa and colleagues grew Eda-deficient teeth in vitro, and found that cusps and crests could be restored by adding EDA. They demonstrated, with the aid of computer models, that different doses of EDA alter tooth morphogenesis (the process by which structures are shaped as they develop), akin to the dental transformations that occurred as rodents evolved from their Mesozoic mammalian ancestors (Fig. 1). For example, the trigonid — the first part of the tribosphenic molar to have evolved — is regenerated with only a small dose of EDA. However, a higher dose of EDA is required to restore the talonid, which evolved more recently5.

Figure 1: Reconstructing tooth evolution in vitro.

a, As mammals evolved, their molars (one indicated in the jaw) became ever more complex, because extra tooth features evolved over time. Structures called trigonids (dark grey) evolved first, in early mammals such as the Mesozoic symmetrodonts, followed by talonids (light grey) in a group of Mesozoic therians called cladotherians. Hypoconulids (blue) evolved in Mesozoic therians, and anteroconids (yellow) in advanced rodents. The anteroconids are similar in structure to pseudo-talonids, which evolved separately (convergently) in pseudo-tribosphenic teeth in an early-divergent clade of mammals (the pseudo-talonid is also indicated in yellow). b, Deletion of the Eda gene in mouse embryos results in the loss of all of these tooth features. Harjunmaa et al.3 show in vitro that addition of the EDA protein to embryonic teeth from Eda-deficient mice in increasing doses can replay the steps of evolution. Furthermore, features that evolved longer ago respond in a less variable manner than features that evolved more recently.

The cusp-to-crest morphogenesis of mouse molars is controlled by a gene network that includes genes encoding the signalling proteins fibroblast growth factor 3 (Fgf3; ref. 9) and sonic hedgehog (Shh)11. An increase or decrease in Fgf3 causes over- or underdevelopment of tooth features, respectively9. Harjunmaa and co-workers found that reducing the concentration of SHH in Eda-deficient teeth regenerated the ancestral features of Palaeocene rodents, reversing the cusp-to-crest transformation of modern-day rodents. Thus, molecular manipulations that alter tooth morphogenesis in vitro can replay evolution, either forward, to mimic the fossil record, or in reverse.

Perhaps the most exciting insight from Harjunmaa and colleagues' work is that ancestral structures show a more uniform response to the addition or removal of EDA or SHH than those that evolved more recently or independently in different lineages (convergent evolution). For example, addition of a low dose of EDA reliably restored the trigonid, as expected of ancestral features, which are typically evolutionarily well conserved owing to their long history. By contrast, higher doses restored the talonid in many, but not all, teeth — development of this feature was more variable in response to EDA. This is consistent with the theory6 that the talonid basin evolved convergently in different mammal lineages, but has reduced in size in some carnivoran or insectivoran mammals.

The hypoconulid is a talonid cusp in some mammals, but is enlarged and forms a separate lobe in mice. The authors found that full development of this structure requires a higher dose of EDA than does the rest of the talonid, and shows even wider variation in its response to EDA. Finally, the anteroconid in mice — another feature that arose late in rodent evolution — requires the highest EDA dose to regenerate, and shows the broadest variation when regenerated. Its position on the tooth corresponds to the 'pseudo-talonid' that arose in some early-divergent mammals that died out before the end of the Mesozoic. Harjunmaa and co-workers' experiment therefore demonstrates that modern-day mice still have the developmental potential to replicate evolutionary events that occurred long ago, in the now-extinct mammals of the Mesozoic12.

The level of EDA required to give rise to individual molar characteristics therefore seems to provide information about how robust their development is. When studying how morphogenesis drives evolution13, it will be crucial to bear in mind that the sensitivity of a particular tooth feature to EDA activity may indicate the likelihood of an evolutionary transformation producing that feature. For example, as mentioned above, talonid-like features evolved twice — in the basal diversification of modern mammals and in early-divergent groups of the Mesozoic. Variable sensitivities to gene-expression dosage and signalling strength can serve as a measure of the evolutionary variability of each tooth feature, and may underpin the many convergences and reversals of tooth evolution observed in the mammalian fossil record.

Eda and Shh have varying effects on many vertebrate structures, so it can be hard to tease apart which evolutionary feature is controlled by which part of the gene network. Harjunmaa et al. have cleared this hurdle in a welcome development that brings us closer to being able to test how changes in morphogenesis affect the final shape of evolving teeth as seen in the fossil record. Genetic engineering of developmental processes in vitro is a fruitful way to decipher how the shape of organs or other biological structures is modified by evolution.


  1. 1

    Gould, S. J. Ontogeny and Phylogeny (Belknap, 1985).

    Google Scholar 

  2. 2

    Abzhanov, A. Trends Genet. 29, 712–722 (2013).

    CAS  Article  Google Scholar 

  3. 3

    Harjunmaa, E. et al. Nature 512, 44–48 (2014).

    ADS  CAS  Article  Google Scholar 

  4. 4

    Kielan-Jaworowska, Z., Cifelli, R. L. & Luo, Z.-X. Mammals from the Age of Dinosaurs: Origins, Evolution, and Structure (Columbia Univ. Press, 2004).

    Google Scholar 

  5. 5

    Crompton, A. W. Zool. J. Linn. Soc. 50 (Suppl. 1), 65–87 (1971).

    Google Scholar 

  6. 6

    Luo, Z.-X., Cifelli, R. L. & Kielan-Jaworowska, Z. Nature 409, 53–57 (2001).

    ADS  CAS  Article  Google Scholar 

  7. 7

    Meng, J. & Wyss, A. R. J. Mamm. Evol. 8, 1–71 (2007).

    Article  Google Scholar 

  8. 8

    Gomes-Rodrigues, H. et al. Nature Commun. 4, 2504 (2013).

    ADS  Article  Google Scholar 

  9. 9

    Charles, C. et al. Proc. Natl Acad. Sci. USA 106, 22364–22368 (2009).

    ADS  CAS  Article  Google Scholar 

  10. 10

    Mikkola, M. J. & Thesleff, I. Cytokine Growth Factor Rev. 14, 211–224 (2003).

    CAS  Article  Google Scholar 

  11. 11

    Cho, S.-W. et al. Development 138, 1807–1816 (2011).

    CAS  Article  Google Scholar 

  12. 12

    Luo, Z.-X., Ji, Q. & Yuan, C.-X. Nature 450, 93–97 (2007).

    ADS  CAS  Article  Google Scholar 

  13. 13

    Carroll, S. B., Grenier, J. K. & Weatherbee, S. D. From DNA to Diversity: Molecular Genetics and the Evolution of Animal Design 2nd edn (Blackwell, 2005).

    Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Zhe-Xi Luo.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Luo, Z. Tooth structure re-engineered. Nature 512, 36–37 (2014). https://doi.org/10.1038/nature13651

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing