Got yolk?

While it has long been recognized that mammals are derived from egg-laying reptilian ancestors, the exact path that we took from laying eggs to nursing young still a lively area of active research. There must have been some genetic mechanism for yolk formation in our ancestors, but what happened to these genes during the shift to lactation? There is a very exciting paper out in PLoS Biology today that answers important questions about the genes involved in yolk production, helping us to better understand what has happened to them in the course of mammalian evolution.

First, a little background. Egg yolk is formed with the help of vitellogenin (VTG), a protein that is synthesized in the liver and acts as a delivery molecule that brings essential nutrients and minerals to developing egg cells. This protein is encoded by the VIT genes, and it appears that the number of copies of VIT genes is related to the speed and degree of yolk development in eggs. If animal X has more VIT genes than animal Y, it will be able to produce larger, more numerous eggs at a faster rate. Thus, measuring the presence and activity of VIT genes gives us a good indicator of the repro physiology of an animal and its degree of lecithotrophy.

One of the fascinating things about mammals is that there are living representatives of several progressive steps toward the evolution of full placental reproduction: as we all know, monotremes lay eggs, marsupials show primitive placentation and give birth to very precocial young, and eutherians have complex placentas that allow them to retain and nourish their embryos more extensively and to give birth to relatively developed offspring. I don't like the term "living fossils," because the living animals are not the same as the ones found in the fossil record and have been adapting to their environments just as long as we have. Nevertheless, monotremes and marsupials have conserved basal modes of reproduction, and this allows us to glimpse different stages of the evolution of placentation and lactation.

So, since we are fortunate enough to have this convenient continuum of reproductive strategies at our disposal, Brawand et al analyzed VIT regions from the genome sequences of a variety of mammals, with representatives from all three groups (monotremes, marsupials, eutherians). They isolated three known VIT genes (VIT1, VIT2, and VIT3) from the chicken genome, and searched for the same regions among the mammalian genomes. Although it was expected that mammals would have lost much or all of the function of the genes responsible for yolk production, the researchers could still search for pseudogenes, degraded versions of regions that have lost function over time.

So, that is what they did, but what did they find?

Monotremes: The platypus genome shows a loss of function of VIT1 (it has become a pseudogene), but does encode a second VIT gene that appears to encompass a region corresponding to both VIT2 and VIT3 in the chicken sequence. So, monotremes retain active VIT function, but not to the same degree as birds. This makes sense: although monotremes do lay eggs, they also have glands that produce milk, which offspring suck from hairs on the mother's belly.

Marsupials: All three VIT genes in the marsupial genomes showed frameshift mutations, which indicate loss of function. What is most interesting about the marsupial results is that it appears the inactivation of all three genes predates the divergence between Australian and American lineages, which occured roughly 70 million years ago. The three genes were not all inactivated at the same time, however. It appears that VIT3 was the first to go, around 170 mya, followed by VIT1 140 mya, and finally VIT3 around 60 mya, although this gene was lost earlier in the eutherian lineage, about 90-100 mya.

Eutherians: All of the VIT genes in humans were inactive and highly degraded, as expected for animals that dispensed with egg-laying many millions of years ago, but the VIT remnants were identifiable from the pseudogene remains of previously functional coding regions. The eutherians were represented by the armadillo, dog, and human. VIT1 was determined to have lost function before these lineages diverged, roughly, 90-100 mya, but inactivation times for the other two VIT genesweren't mentioned.

The dates are significant, because they mean that two VIT genes could have lost function before the split between meta- and eutherians. This is very likely for VIT3, but it is only a possibility for VIT1, the authors point out that the timing is such that it could have been rendered a pseudogene either in the common ancestor or independently in each lineage. VIT1 was lost very recently, 50 mya, in the platypus, meaning that it was independiently inactivated in monotreme and therian evolution. VIT2 was degraded much later, so it appears to have been independently inactivated in each lineage.

The researchers also searched for evidence of CSN genes. These encode for casein phosphoproteins, which are crucial components of mammalian milk and are good markers for the development of lactation. The results indicated that caseins were present in the common ancestor of all mammals, appearing sometime between 210-300 mya. This means that caseins were present by the time VIT genes were losing function, confirming that the development of lactation provided animals with alternative sources of nourishment, which is what made them less dependent on yolky eggs and thus allowed the VIT genes to degrade.

A few final notes to ponder: It seems intuitive that monotreme yolk genes show an intermediate state between ancestral dependence upon vitellogenesis and the derived lactation of modern mammals, but having definite sequences to confirm this expectation is important. I would be interested to see if the VIT genes in the echidna show the same patterns of inactivation as the platypus, maybe the echnida genome would give us a clearer picture of whether the mega-VIT they retain originates from VIT2 or VIT3, or if it is somehow a fusion of the two. VIT1 became a pseudogene around 50 million years ago, relatively recently. I am not sure of when the platypus/echidna split occurred within the monotreme lineage, if you happen to know please leave a comment and enlighten us! Echidnas usually lay only a single egg, while the platypus lays 1-3 at a time. It would be fascinating to see if echidnas show further VIT degradation, although I don't know how the total volume of egg laid scales to body size in each species.

One of the most fascinating things about the study is how recently some of the VIT genes were lost. The appearance of caseins predates the estimated eutherian inactivation of the last VIT gene by around 200 million years, meaning that mammals hung onto vitellogenin even after they had developed the ability to nourish their young with milk through lactation.

To sum it all up, Brawand et al have shown that VIT genes can be used as indicators for the switch from predominately yolk-based nourishment of offspring to the rise of lactation and placentation. Monotremes, which are the only mammals that still lay yolky eggs but also exhibit primitive lactation, have reduced one VIT region to a pseudogene but retain another functional region that allows VTG production for yolk development. This is yet another item to add to the list of intermediate features displayed by the monotremes. The study also shows that CSN genes were present in the common mammalian ancestor, but that VIT genes have been independently inactivated multiple times among the three major groups of modern mammals. I'd be interested to see results from multituberculates also...this is a branch of the mammalian family tree that has gone completely extinct, no living representatives of the clade are left. I wonder what our odds are of getting a Meniscoessus genome sequenced?

Brawand, D., Wahli, W., Kaessmann, H. (2008). Loss of Egg Yolk Genes in Mammals and the Origin of Lactation and Placentation. PLoS Biology, 6(3), e63. DOI: 10.1371/journal.pbio.0060063


(Image credits: Egg baby, DNA, Platypus, Wallaby, Baby and puppy)