Embryology Biology 441 Spring 2007 Albert Harris


"Evo-Devo" = Evolutionary Developmental Biology

This will be the very last lecture in the course

This is becoming one of the most active and popular sub-fields of embryology.

For example, there is a weekly "Evo-Devo" seminar at Duke, which anyone is welcome to attend. Many books and scientific papers have been written on aspects of development related to evolution, and aspects of evolution related to how embryos develop. Many national and international scientific meetings focus on this general area. It doesn't have a precise definition, however - lots of different kinds of research are considered to be evo-devo. In recent years, this area has increased greatly in popularity, and some of the best scientists are focusing on it.

Consider the following simple facts:

  • Evolution occurs by random mutations in genes.
  • But the effects of these mutations on the organism's phenotype (body structure, etc.) is what determines whether organisms with the mutant gene will survive, or not.
  • Therefore, we should expect that evolutionary changes in anatomy will be limited by peculiarities of the embryological mechanisms by which genes cause anatomy.
  • It may be possible to deduce facts about embryological mechanisms based on studies of genes, and vice versa.

This subject contains many speculations; but these would be very important if true.

There was a much earlier period in the history of embryology in which evolutionary implications dominated research. This began soon after 1860 and continued into the early 1900s.
This earlier dominance had the slogan "Ontogeny recapitulates phylogeny" - in other words: The sequence of anatomical changes during embryological development should be interpreted as (Somehow! For some unstated reason!) repeating the evolutionary sequence of anatomical shapes of any given species' ancestors.
Ontogeny means embryonic development.
Phylogeny means evolutionary origin (for example, in a "phylogenetic tree".)
Recapitulates means repeats.

The inventor of this "Biogenetic Law" was a German professor named Ernst Haeckel, who invented lots of other fancy words, including "Ecology". (To get this more in historical perspective, the words "genetics" and "gene" weren't invented until 45 years later). He first became famous for his drawings of animals discovered during the cruise of an oceanographic ship, and some of these drawings are frequently reproduced, mostly of species that people now suspect he may have invented.

Among specific examples was the hypothesis that the reason mammal and bird embryos pass through stage which have gill slits is that they (& we) evolved from fish. Scientists still accept this particular belief, about gill slits. And we believe we are descended from sea squirts, because their larvae have notochords. Furthermore, we probably evolved from neotenic sea squirts, that lost their metamorphosis into the adult form. Notice the analogy to neotenic salamanders, like axolotls, that no longer undergo metamorphosis. Also supporting this implausible phylogeny is the fact that actual neotenic species of sea squirts do exist.
There was also a (now almost forgotten) theory that humans are neotenic apes.

On the other hand, many comparable hypotheses have been disproved, or otherwise abandoned.
A good example was the theory that bone replaces cartilage during our embryonic development because this is a recapitulation of an evolutionary sequence in which our more distant ancestors' skeletons were made out of cartilage (as in adult salamanders and sharks, etc.) and then bone supposedly evolved later. It is now believed that bone actually evolved before cartilage, and that cartilage evolved later, as a means of allowing the skeleton to grow by internal swelling instead of being restricted to surface deposition (which is the only way bone can enlarge, because of its rigidity).

In the late 1800s, most research on embryos was motivated by hopes of discovering more about the evolutionary origins (phylogeny) of different kinds of animals. Barnacles must be crustaceans because their larvae swim and are almost indistinguishable from adults of some swimming crustacean. UNC's justly famous H.V. Wilson studied teleost fish embryology from this point of view (but was rather sarcastic about it!).

From about 1890 on, dominance shifted toward mechanistic research and concepts about embryology. Replacing the idea that embryonic development is some kind of accelerated repeat of all the stages of a given animal's evolution, scientists began to look for causal mechanisms by which anatomy is formed, and then to consider connections between genes and embryos.

For example, Thomas Hunt Morgan was primarily an embryologist, before and after his research using fruit flies to discover that genes are located along chromosomes, etc. His thesis research was embryological, and he wrote an embryology textbook, but his Nobel Prize was for his discoveries about genes. Many of the earliest geneticists began as embryologists. Morgan explained his Nobel Prize by saying that embryology was just a more difficult (and more interesting) subject, as compared with genetics.

Scientists then began to wonder what genetic reasons might explain recapitulation. One idea is that evolution tends to add major changes onto the end, as it were, of an animal's sequence of developmental changes. The embryos of related species therefore resemble each other more than the adult stages look alike. Likewise, if we believe that the great majority of random mutations cause harmful changes, then the more different effects a gene has, the less likely it is that mutating that gene can escape causing harmful changes in one or the other of the gene's multiple effects. Thus, genes with multiple effects can be expected to change much more conservatively (if at all), because not just 9 out of 10, but 99 out of a hundred, or 99,999 out of a hundred thousand mutations in such a gene will have some lethal effects.

Genes that control earlier and/or more fundamental embryological stages should therefore evolve the most slowly, which would tend to result in actual recapitulation. That was the explanation I learned in high school and college.

In the last ~30 years, scientists invented methods for direct study of the genes that control the earliest stages of development. This has led to a new explosion of scientific interest in connections between embryology and evolution (no longer called "recapitulationism" but now called by the somewhat trashy name Evo-Devo). If only some of the evodevoists knew a little Greek or Latin, they could have invented a classier name.

Key methods include * genetic screens of genes in which mutation causes relatively early death of embryos (primarily in flies), * "in situ" hybridization of complementary nucleic acids, so that you can selectively stain just those parts of an embryo where mRNA of a certain gene is localized, and also * monoclonal antibodies, * DNA base sequencing methods, not to mention methods for targeted inactivation of genes and for transfer of certain DNA sequences from one organism to another.

The most exciting discovery has been that many of the transcription factor genes have almost the same base sequences in vertebrates as they do in insects, and likewise in nearly all multicellular animals. You remember that transcription factors are the proteins whose function is to bind selectively to promoter regions of DNA, for the purpose of controlling which genes will be transcribed, and which won't ("turning genes on and off"), at different stages and in different parts of the body.

For example, although the structure of vertebrate eyes could hardly be more fundamentally different than the structure of insect eyes, very similar genes (similar enough to be given the same name, and to substitute for each other!) have been discovered in both vertebrates and in flies. This gene is called Pax 6.

Flies have a Pax 6 gene; Mutating this gene causes a failure of eye development; Over-activating this gene causes flies to develop extra eyes in various wrong places. We and other vertebrates have a gene with nearly the same base sequence, which is therefore called by this same name: Pax-6. Mutating a vertebrate's Pax-6 gene also causes abnormal development of eyes. Furthermore, and even more impressive, Pax-6 genes from one kind of animal can substitute functionally for missing or mutated Pax-6 genes in animals belonging to very different phylogenetic groups (e.g. humans and insects).

There are many such examples of genes discovered in flies, for which very similar cousins later got discovered in mammals, etc. (Wingless genes in humans! Not to mention Sonic Hedgehog, and all that).

The degree of this genetic similarity surprised everyone, but remember it is at the level of the genes whose proteins control the expression of other genes. These latter genes are the ones that actually build the eyes and the wings, and other organs - and these other genes are much less similar in insects as compared with vertebrates. The genes that control organ locations are the ones that have changed so little in evolution.

It had long been assumed that eyes and legs and wings evolved separately in the two main phylogenetic branches of the animal kingdom, the Arthropods as compared with the Vertebrates. Some scientists now doubt that; Others think that separate evolutionary origin of organs is not inconsistent with this discovery that the control genes, at the deepest levels of control, are almost the same.

In my own early education, the wildest phylogenetic speculation that we were taught about was that (maybe, maybe...) vertebrate bodies are in some sense upside-down relative to arthropod bodies, and vice versa. Arthropod hearts are along their backs; their neural trunks are along their ventral sides. In vertebrates, the heart is ventral and the neural tube dorsal. In these ways and a few others, we a somewhat like upside down versions of each other. One difference is that the Arthropod brain is circular, and surrounds the anterior end of the digestive tract, whereas no vertebrate has anything like that.

Molecular evidence has been discovered supporting even this crazy old theory, however. Among the transcription factors that control which sides of embryos will become dorsal, and which will become ventral, there are some close similarities between vertebrate and arthropod transcription factors, and their genes. Except these similarities of the transcription factors are partly upside-down, in the sense that some of the "back-causing" genes of insects more closely resemble the "ventral-causing" genes of vertebrates, and vice versa. Apparently none of these 120-year-old theories is too wild for molecular evidence to be found in their support!

The most tantalizing of all questions in this field is currently that of the spatial patterns of expression (= location of the transcription factor proteins) of the family of hox genes.

Hox genes were discovered in flies, but genes with very similar sequences were soon found in vertebrates, and almost all other multicellular animals. The fly genome has a family of 8 hox genes, divided between two locations on their chromosomes. These were discovered by the fact that mutations that caused homeotic changes in imaginal disk development always mapped to one or other of these 8 genes. Then all 8 genes (and many other genes that also code for transcription factors) were found to have almost the same DNA base sequences in the part of the gene that codes for the part of the protein that actually binds to the DNA of other genes' promoter regions. The name "homeobox" was invented to refer to these DNA sequences, but the term "Hox gene" has been reserved as a name for that original 8, and for their closest equivalents in vertebrates and other animals.

Mammals have four clusters of hox genes, on different chromosomes. Each cluster had a maximum of 13 members, all linked very closely so each other that they form a row on the chromosome. These clusters are called A, B, C and D, and the members of the clusters are named A1, A2, A3 etc. For each of these 4 families of linked genes, the number ones have more similar base sequences to the other number ones, etc. for each of the 13 numbers. The exceptions to this rule fit a pattern which can be expressed by saying that number such-and-such is missing from family A, etc. The fly hox genes have names rather than numbers, but each of their base sequences matches particular vertebrate hox genes, and these similarities are to genes in the same relative sequential locations on the chromosomes. For example, A1 is more similar to the fly hox gene at one end of the linkage group, and A13 is more similar to the fly hox gene at the other end of the other fly linkage group.

Furthermore, and most tantalizing of all, the spatial locations of these genes on their chromosomes (almost!) always matches the sequence of locations in the embryo where those genes are (mostly) expressed. For example, using in situ hybridization to find the location of mRNAs of the A1 hox gene, these RNAs were found to have their highest concentration at a certain place near the head. The A2 hox gene's mRNAs are more concentrated in parts of the embryo a few somites posterior to where the A1 gene's mRNAs were most concentrated. Actually, there are complications, such as that the most anterior region of expression of particular hox genes is often different in the neural tube as contrasted with the adjacent mesoderm, and also that there a sharp anterior borders of each of these regions of gene expression (= location of the mRNAs and location of the transcription factor proteins), but then the concentration of these molecules grades off gradually toward the rear of the animal.

Nobody really knows what mechanistic sense it makes that the relative locations of hox genes on their chromosomes should be in the same order as the locations where these genes are transcribed in the developing embryo. In the earliest days of genetics, many people expected that maybe all the genes used in the arms might map together on one chromosome, and the genes used in making the liver might map together somewhere else. But patterns like that were not found, in general. The hox genes are sort of an exception, in which the locations on the chromosome match the relative locations where a set of genes is active in the embryo.

Experiments have confirmed that damaging one hox gene of the series will result in abnormalities in the part of the body where that gene is most transcribed. That was more or less how homeotic genes were discovered in flies. Nobody has produced a mutant chicken that has legs where it wings should be, or anything as dramatically similar to homeotic abnormalities in flies. [I sometimes fantasize that flies might produce a horror movie titled "The Geneticist"!]

The closest analogies in vertebrates are that some hox gene mutations produce abnormal development of gill slits and others cause certain vertebral bones to have shapes that are supposed to occur in vertebrae of other parts of the body. Notice that both are cases of segmental errors, tissues developing shapes that are abnormal for them, but normal for cells in a nearby segment of the body.

Something very, very important and fundamental is waiting here to be discovered. The biggest breakthroughs are yet to be made. It is not yet understood why having genes next to each other on a chromosome is a good way to cause the proteins coded for by those genes to be synthesized in a matching geometric pattern in the developing embryo. Some theories I have heard include the idea of an enhancer region some distance away from one or the other end of the row of hox genes, with some kind of competition between the promoter regions of the hox genes. Another general idea (stated with complete confidence) was that the chromosomes gradually become de-inhibited from one end to the other... Or something!

No one knows how to find out. No one can think of experiments to distinguish between alternative theories about how all this might work.

In addition, no one knows why having different (but similar) transcription factors is a good way to control the anatomical shaping of parts of the body, from anterior to posterior.

And, lastly two facts (? I think they are facts. Scientists I respect have told me.) that seem not to fit anybody's theories are that these hox transcription factor proteins are much too large to diffuse from one cell to another (so they can't be forming multicellular diffusion gradients); and also, these anterior boundaries of hox gene activity move gradually (relative to cell movements, I mean) in the sense that different cells become the anterior boundaries.

Maybe all these problems will answer themselves once people have run a few thousand more in situ hybridizations, or RNAi experiments. Maybe no creative, inventive thinking will be needed. Almost everyone seems to be hoping so.



back to syllabus

back to index page