Carl Zimmer wrote on evolution in jellyfish, with the fascinating conclusion that they bear greater molecular complexity than was previously thought. He cited a recent challenging review by Seipel and Schmid that discusses the evolution of triploblasty in the metazoa—it made me rethink some of my assumptions about germ layer phylogeny, anyway, so I thought I’d try to summarize it here. The story is clear, but I realized as I started to put it together that jeez, but we developmental biologists use a lot of jargon. If this is going to make any sense to anyone else, I’m going to have to step way back and explain a collection of concepts that we’ve been using since Lankester in the 19th century.

First, triploblasty. This term refers to the embryonic origins of the tissues of your body. Early in development, the vertebrate embryo organizes itself into a three-layered (hence “triplo”) structure, each layer of which is the progenitor or source (hence “blast”, which is Greek for bud, or germ) of a defined set of adult tissues. These layers are also called the germ layers, with “germ” in the sense of a source or starting point, as in the word “germination”. Your original three germ layers consisted of ectoderm, which developed into your epidermis and central nervous system; endoderm, which forms the lining of your gut and associated organs; and mesoderm, which contributed to all the stuff in between your gut tube and your skin, muscle, bone, connective tissue, blood, etc.

That all sounds neat and tidy, but this is biology, so of course it also gets fuzzy around the edges, and nothing is quite as discrete as three, simple, sharp germ layers. You will also sometimes hear the word “mesendoderm”; this refers to the tissues that have not yet separated into mesoderm and endoderm, and most accurately, we should say that most vertebrate muscle is mostly derived from mesendoderm, for instance. Another important and intermediate tissue is mesectoderm, cells derived from ectoderm that wander off into deeper layers and adopt diverse fates. And of course, all three germ layers interact extensively throughout development—it’s mesoderm communicating to ectoderm that persuades part of it to form a brain instead of skin, for instance.

What about diploblasty? That should be easy to figure out at this point: two germ layers. The diploblasts are animals, such as jellyfish and hydra, that are supposedly derived from just two germ layers, and in the adult, consist of an outer layer of cells (a skin), an inner layer (a gut), surrounding a gelatinous and sometimes acellular layer called the mesoglea. Don’t be confused by the fact that adult jellyfish have more than two layers, anymore than you should be baffled by the fact that adult humans have more than three layers of cells—the terms are all about how many layers were present in the early embryo, and that’s one of the things that this paper by Seipel and Schmid is discussing, whether the embryonic diploblasts might have more than two germ layers (and to cut the story short, their answer is going to be a strong “yes!”).

This has evolutionary implications, so let’s also introduce a little general taxonomy, and get familiar with the players. Here are the major animal groups we’re concerned with in this paper.

This organization suggests some evolutionary patterns. Look at just the superphylum column, for instance. The Parazoa have no differentiated tissues, so you might think of them as derived from a single germ. The Radiata are defined as having two germ layers. And there are the Bilateria, with three germ layers. One, two, three…to our pattern-forming brains, that suggests an evolutionary sequence of stepwise addition of new germ layers. Or, alternatively, that our pattern-forming brains have imposed a simple sequential pattern that isn’t really there. The comment in the table for the Ctenophora should make it clear that there are some…complications here, and the Seipel and Schmid work suggests a different pattern: one, three, then some go to two.

Whatever way it goes, though, the Radiata are a particularly interesting group to study. They branched off the lineage leading to us at about the time the fundamental features of the bilaterian body plan were evolving. Those features we share in common with jellyfish are likely to have been present in the Urbilaterian, the organism ancestral to all of the bilaterian phyla, and figuring out what those features are will tell us something about the early stages in our lineage’s evolution. In the figure below, the various features that led to our pattern are indicated by the series of black dots labeled “Basic Bauplan”, and you can see the Cnidaria branching off in the middle of them. What do we have that they don’t? What do we share?

Overview of the evolutionary relations in the metazoan kingdom. The evolutionary position of the Basic Bauplan is indicated by black dots. The adult stages of the four cnidarian classes are indicated by schematic drawings. While the Cubozoa always have polyp and medusa stages, either one can be reduced in the Hydrozoa and Scyphozoa. The Anthozoa have only the polyp stage.

The life cycles of the hydrozoans P. carnea (B) and Hydractinia echinata (C) and a representative anthozoan (D).

The main stages of
medusa development in P. carnea, with medusa bud stages 2, 3-4, 5-6, and 8-9. Abbreviations: e, entocodon (brown); ec, entocodon cavity; ex, exumbrella; go, gonads; m, mesoglea (ECM); ma, manubrium (feeding organ); maa, manubrium anlage; o, migrating oocytes; p, plate endoderm; rc, radial canal; sm, smooth muscle of the manubrium (dark green); sr, smooth muscle over the radial canal (light green); st, striated muscle (red); t, tentacles; v, velum.

Most cnidarians have a complex life cycle. Typically, the eggs develop into a two-layered planula larva, which then develops into a sessile polyp stage (a form familiar to many a high school biology class as the freshwater hydra), and finally a stage that forms by budding from the polyp, the medusa—the typical bell-shaped organism we think of when we hear the word “jellyfish”.

The interesting stuff is going on in the development of the medusa. The medusa has greater complexity, new cell types, and also seems to have a distinct third germ layer—calling it a diploblast seems to be a misnomer.

The diagram to the right is a series of sections at different developmental stages of a maturing medusa bud. The structure you see is a bud ballooning out from the body wall of a polyp, with etoderm on the outside (the stippled area) and endoderm inside (the columns of cells). At about stage 2, some cells separate from the ectoderm to form another layer, called the ectocodon, colored brown in the first two diagrams. The ectocodon is a distinct and separate cell layer, separated from the other two layers by extracellular matrix, so at this stage the medusa is inescapably a three-layered organism.

What’s going on in stage 3-4 is remarkably reminiscent of what occurs in us chordates. In chordates, the embryonic mesoderm in our bodies cavitates, or opens up in the middle, separating into two sets of mesodermal derivatives. The part adjacent to the gut is going to form the muscle and connective tissue of the gut, while the part next to the ectoderm will form body wall muscle.

In the medusa, the ectocodon also cavitates, and the part near the feeding organ (manubrium, ma in the stage 5-6 diagram) will form a layer of smooth muscle (green), while the part towards the outside will form the smooth and striated muscle lining the umbrella-shaped bell of the jellyfish. It’s a pattern that is remarkably similar to what we see in canonical triploblasts.

The similarities run deeper than a superficial morphological correspondence. Cnidarians regulate their muscle development with the same genes we do: Twist and Snail, Brachyury and various homeodomain genes. Their muscle structural genes, myosin and tropomyosin, resemble the same proteins found in bilaterians. They even use similar genes in neurogenesis, such as the Achaete/Scute complex. While much simpler in organization, cnidarians are setting up their muscles and nerves using a toolbox much like ours.

Now, it is possible that cnidarians evolved mesodermal derivatives like muscle completely independently, in a case of convergent evolution that recruited similar regulatory genes. That isn’t what Seipel and Schmid argue, though: instead, they postulate a last common ancestor that was relatively sophisticated.

Both the jellyfish and bilaterian striated muscles are derived from mesoderm-like primordia in a common ancestor established before the Zootype with clustered Hox genes evolved. In this case, the ancestor was not a diploblast planula type organism, but an organism with advanced anatomy including striated muscle.

Here is their diagram of the course of evolution in the metazoa.

Hypothetical model of the early metazoan evolution. Schematic drawings of sagittal sections with upward orientation of the oral opening illustrate the discussed hypothetical anatomies of metazoan ancestors. (A) Ancestral metazoan with flagella (thin black lines), adhesive structures (thick black spikes), digestive area (blue), gametogonia (orange), and primordial myocytes (light green and light red). (B and C) Intermediate stages of the basic metazoan Bauplan. (B) Intermediate stage formed from panel A. The digestive area has formed a pocket and primordial myocytes are found between the digestive and the flagellated epithelia. (C) Intermediate stage formed from panel B (or from panel A). It has a through gut and anterior – posterior polarity, primordial myocytes start aligning along the digestive tube. (D) Ctenophore ancestor, derived from B, a massive ECM has formed, myocytes mostly differentiate into the smooth muscle type (green) with a subepidermal location across the ECM. Locomotion is generated by specialized flagella, the comb plates. (E) Anthozoan polyp ancestor might have formed directly from panel B (or from panel F). (F) Cnidarian jellyfish-like ancestor, derived from B, or by closing the through gut from C (dashed arrow). Premordial myocytes differentiate into radial smooth muscle (green) and striated muscle (red) with a circumferential orientation. The digestive region is centralized in the radial animal. The inner layer is formed by a flagellated and smooth muscled epithelium. (G) Zootype ancestor, derived from panel C. Myocytes differentiate to smooth (green) and striated (red) muscle aligned in parallel around the digestive tube, at the ends sphincter muscles (red) control
the aperture of the digestive tube. This stage develops into the bilaterian line.

The ancestral metazoan (A) would have had fairly few specializations, but it would have had cells set aside for digestion (blue), others that formed an outer protective layer (white), and an inner mass containing gametes (orange) and muscle (pink and light green). They speculate that one function of the primitive muscle would have been to generate contractions to expel gametes. Invagination of the digestive area would have created a pouch-like gut (B), or if the invagination punches all the way through, a complete gut (C). From that point, the generic contractile tissue differentiated into smooth and striated muscle, and there were major reorganizations of morphology. In the bilateria, of course, an important innovation was the establishment of clustered Hox genes that specified positional information along the longitudinal axis.

The fascinating thing about this diagram is all the arrows going every which way, illustrating the possible paths evolution took. The solid arrows represent the authors’ best interpretation of the likely paths, but the dashed arrows are also reasonable possibilities. Whee! So much left to learn!

The authors do conclude that diploblasty is a derived state.

In summary it appears that Cnidaria derive from a motile
pre-zootype metazoan featuring mesodermate and possibly
bilaterian elements of anatomy. In this scenario the
evolution of the basic Bauplan did not include a diploblast
stage. The evolution of anatomical elements able to generate
rapid locomotion required the simultaneous emergence of a
digestive support system and a concurrent connection of
musculature and nervous system. Additionally there were
placement constraints for the basic anatomical elements
within the body and with respect to each other. All the basic
anatomical elements probably co-evolved as integrated
functional units in the basic Bauplan. This scenario reflects
a simplified evolutionary process leading to the major
animal phyla.

At the beginning, I mentioned the old appealing idea of a simple sequence of germ layers in evolution, from one to two to three. This model suggests something different, that we went from one straight to three; is that a bothersome leap? I don’t think so. Development is not simply about stacking tissues together like legos, but rather, about setting up and executing interactions. The processes that evolved to allow for specialized tissues of any number beyond one would have created the potential for combinatorial relationships to generate many cell identities—evolution proceeds not by serial addition of discrete elements, but by coevolution of multiple entities in parallel.

Seipel K, Schmid V (2005) Evolution of striated muscle: Jellyfish and the origin of triploblasty. Developmental Biology 282:14-26.