Original cartoon by Alex Martin
Source: Original cartoon by Alex Martin

In any placental mammal a finely tuned immune system protects its body against invasion. Detection of foreign proteins triggers rejection. Yet half the genes of a fetus in the mother’s womb stem from the father, producing many distinctive proteins. Why does the mother’s body tolerate this alien presence? The fetus somehow circumvents her immunological defences, avoiding rejection. Peter Medawar, the father of immunology, first noted this immunological paradox in 1953. He drew a direct parallel between a fetus and a transplanted “allograft”  —  a tissue or entire organ  —  from another individual, stimulating a huge body of research over the past six decades.

But similarity between a fetus and a transplant goes only so far. Both mother and fetus have a vested interest in pregnancy, so it entails more than acceptance or rejection. Instead, a compromise is reached between nurturing the fetus and avoiding runaway invasion. The mother-fetus relationship involves close cooperation, with a unique interaction in the placenta between fetal cells and maternal white blood cells. But, despite many advances, reviews by Ashley Moffett and Charlie Loke in 2004 and 2006 revealed that the immunological paradox still awaits proper resolution.

Evolution of the placenta

Adapted from an illustration in Martin (1990)
Diagram of fetal membranes in a placental mammal:  Abbreviations: C = chorion; V = yolk sac; Al = allantois; Am = amnion; E = embryo.
Source: Adapted from an illustration in Martin (1990)

As their name indicates, all placental mammals have a well-developed placenta serving as the primary interface between mother and fetus. Of four fetal membranes serving different functions, the outermost  —  the chorion  —  always encloses the entire system. As the external barrier in contact with the womb it necessarily engages in any resistance to the mother’s immunological defences. But the placenta shows intriguing variation. On one hand, there is a wide spectrum of different types; on the other, each major group (order) of mammals is generally characterized by just one type, indicating early commitment to ancestral conditions. Recognition of three basic placenta types by Otto Grosser over a century ago has proven its worth: non-invasive, moderately invasive or highly invasive. In contrast to the invasive types, no breakdown of the womb’s inner lining occurs in a non-invasive placenta. In a highly invasive placenta maternal blood directly contacts the chorion. Taking examples among placental mammal orders, the placenta is non-invasive in both even-toed (artiodactyl) and odd-toed (perissodactyl) hoofed mammals, moderately invasive in carnivores and elephants, and highly invasive in rodents and rabbits. Variation within an order usually involves moderately or highly invasive types. But primates are a striking exception: the non-invasive placenta of lemurs and lorises is utterly unlike the highly invasive placenta of tarsiers and higher primates.

Adapted from an illustration in Martin (1990)
Diagram of 3 basic types of placenta, according to degree of invasion of the womb. Fetal tissues (pale pink) and maternal tissues (dark pink) are always separated by the chorion. Uterine glands (blue), producing “uterine milk”, are best developed in species with a non-invasive placenta and provide an alternative source of fetal nutrients.
Source: Adapted from an illustration in Martin (1990)

Many attempts have been made to trace the evolution of the placenta. For several decades, a dominant notion was that a placenta’s efficiency increases as it becomes more invasive, because reducing barriers between maternal and fetal blood favours exchange. A non-invasive placenta is accordingly seen as least efficient and most primitive. However, I have long argued that this interpretation is misguided. To cite just one problem: Dolphins  —  now known to be close relatives of hippopotamuses, nested among artiodactyls  —  have a purportedly “inefficient” non-invasive placenta and yet show rapid fetal growth, including development of a particularly large brain. As broad-based DNA trees for placental mammals became increasingly available after 2001, a revolutionary new consensus rapidly emerged. Four independent studies (including my own) all concluded that a non-invasive ancestral placenta was highly unlikely because far more changes would be required in subsequent evolution. A moderately invasive condition in ancestral placental mammals requires the least evolutionary change. I concluded that an explanation for the evolution of alternative placenta types must lie in trade-offs between womb invasion and overcoming immunological defences.

Bugs in the genome

It was originally thought that DNA in a cell nucleus consisted of long sequences of genes, each coding for a particular protein. Surprisingly, it gradually emerged that mammal DNA predominantly consists of non-coding sequences called “junk DNA” because most of them have no known function. In the human genome, for instance, only 1% of the DNA sequences code for about 25,000 genes, while another 7% may be associated with gene function in some way. Of the remaining 92%, “jumping genes” (mobile elements) make up almost half of the human genome, and about a sixth of these are derived from mostly inactive retroviruses. An invading retrovirus inserts DNA into the host’s genome and can initially be very dangerous. But the host species gradually takes control and over time the inserted sequences typically degenerate and undergo extensive rearrangement. Only the most recent retroviruses are intact and active, a well-known example being the human immunodeficiency virus (HIV) responsible for AIDS.

National Institute of Allergy and Infectious Diseases (NIAID), through National Institutes of Health (NIH) [Public domain, via Wikimedia Commons]
Structure of HIV, a typical retrovirus. The gag gene codes for a precursor of the shell, the pol gene yields enzymes that convert viral RNA into DNA for insertion into the host genome (reverse transcriptase and integrase), and the env gene codes for surface protein in the virus’s envelope.
Source: National Institute of Allergy and Infectious Diseases (NIAID), through National Institutes of Health (NIH) [Public domain, via Wikimedia Commons]

Jumping genes that enter the germ line are transmitted from one generation to the next. Although they are commonly dismissed as mere “genetic parasites”, accumulating evidence indicates that some (notably retroviruses) have been repeatedly recruited for beneficial functions. A typical retrovirus genome includes only 3 genes: a gag gene coding for a precursor of viral shell constituents, a pol gene coordinating production of components needed to convert viral RNA into DNA for insertion into the host’s genome, and an env gene coding for protein molecules embedded in the virus’s outer envelope. Over evolutionary time, successive retroviral amplifications generate families of repeated sequences. In certain rare cases, individual retroviral genes have been conserved for millions of years while the remaining sequences have degenerated. Retention of a single functional gene of retroviral origin in a cluster of related species indicates a selective benefit to the hosts.

Viral genes in the placenta

In a major breakthrough it was discovered that env genes from retroviruses have been repeatedly “captured” to serve key functions in the mammal placenta. Placenta-specific genes coding for envelope proteins, in each case derived from members of different retrovirus families, have been identified in the genomes of mammals belonging to several different groups. Envelope proteins of retroviruses play an essential rôle during infection through fusion with the host cell membrane. Moreover, experiments have shown that those proteins also suppress the host’s immune response. Convergent evolution has occurred in several mammal groups to “domesticate” env genes of retroviruses (renamed as syncytins) and exploit their properties of fusion and immunosuppression in the placenta. Among higher primates, one syncytin gene occurs only in Old World monkeys, apes and humans, while another occurs in New World monkeys as well. So the latter was presumably already present in the initial common ancestor of all higher primates, while the former emerged later only in the Old World branch. Evidence for “purifying selection” acting on both genes points to an important function. In separate developments, captures of syncytin genes coding for retroviral envelope proteins in the placenta have occurred twice in the mouse group of rodents, once in rabbits and hares, once in carnivores and once in tenrecs. In addition, various ruminants ranging from cows to giraffes  —  but not other artiodactyls  —  have a placenta-specific syncytin gene. Although artiodactyls generally have a non-invasive placenta, in ruminants the syncytin gene is involved in a very limited process of cell fusion.

Modified version of a tree in Cornelis et al. (2014).

Simplified evolutionary tree of the placental mammals, showing main placenta types (blue = non-invasive; orange = moderately invasive; red = highly invasive). Inverted mauve triangles indicate 8 independent insertions of syncytin genes in at least 6 different lineages.

Source: Modified version of a tree in Cornelis et al. (2014).

All syncytin genes identified to date were captured long after the origin of placental mammals, so they tell us nothing about the initial ancestral condition. Aiming to fill this gap, Lavialle and colleagues proposed that emergence of placental mammals was accompanied by capture of an original retroviral env gene that was subsequently replaced in various lineages through capture of different env genes following successive independent  infections by new retroviruses. A logical implication is that evidence for “lost syncytins” should be present in the genomes of placental mammals. Preliminary evidence has, indeed, been found in another retroviral envelope protein gene in the human genome; but additional confirmation is needed to complete the story.

There is clearly much more to be discovered. But what we already know provides a neat illustration of the way evolution works. Through “tinkering”, existing material (in this case viral envelope genes) can be modified for new purposes. Moreover, if a valuable new function results  —  as in expression of “captured” viral genes in the placenta for cell fusion and immunosuppression  —  that evolutionary modification may take place independently in several different lineages. And it all goes to show that fact is truly stranger than science fiction.


Carter,A. (2013) An ancient endogenous virus. Blog post in The Evolving Placenta: http://www.placentalevolution.blogspot.ch/2013/08/an-ancient-endogenous-...

Cornelis, G., Vernochet, C., Carradec, Q., Souquere, S., Mulot, B., Catzeflis, F., Nilsson, M.A., Menzies, B.R., Renfree, M.B., Pierron, G., Zeller, U., Heidmann, O. & Heidmann, T. (2015) Retroviral envelope gene captures and syncytin exaptation for placentation in marsupials. Proceedings of the National Academy of Sciences U.S.A. 112: E487-E496.

Cornelis, G., Vernochet, C., Malicorne, S., Souquere, S., Tzika, A.C., Goodman, S.M., Catzeflis, F., Robinson, T.J., Milinkovitch, M.C., Pierron, G., Heidmann, O., Dupressoir, A. &  Heidmann, T. (2014) Retroviral envelope syncytin capture in an ancestrally diverged mammalian clade for placentation in the primitive Afrotherian tenrecs. Proceedings of the National Academy of Sciences U.S.A. 111:E4332-E4341.

Lavialle, C., Cornelis, G., Dupressoir, A., Esnault, C., Heidmann, O., Vernochet, C. & Heidmann, T. (2013) Paleovirology of 'syncytins', retroviral env genes exapted for a role in placentation. Philosophical Transactions of the Royal Society of London B 368, Art. No. 20120507:1-10.

Martin, R.D. (1990). Primate Origins and Evolution: A Phylogenetic Reconstruction. New Jersey: Princeton University Press.

Martin, R.D. (2008) Evolution of placentation in primates: Implications of mammalian phylogeny. Evolutionary Biology 35:125-145.

Medawar, P.B. (1953) Some immunological and endocrinological problems raised by the evolution of viviparity in vertebrates. Symposia of the Society for Experimental Biology 7:320-338.

Moffett, A. & Loke, Y.W. (2004) The immunological paradox of pregnancy: a reappraisal. Placenta 25:1-8.

Moffett, A. & Loke, Y.W. (2006) Immunology of placentation in eutherian mammals. Nature Reviews Immunology 6:584-594.

Zimmer, C. (2012) Mammals made by viruses. Blog post in The Loom: http://phenomena.nationalgeographic.com/2012/02/14/mammals-made-by-viruses/

You are reading

How We Do It

Masturbation: Self-Abuse or Biological Necessity?

A dispassionate view of the natural history of self-gratification.

Bigfoot Parents Have Small Brains

The cautionary tale of Australia’s peculiar incubator birds

A Sperm’s Obstacle Course to the Egg

Men and women have quite different tactics for sperm numbers