Evolution of direct development in echinoderms

In chapter 14 of the Origin of Species, Darwin wondered about the whole process of metamorphosis. Some species undergo radical transformations from embryo to adult, passing through larval stages that are very different from the adult, while others proceed directly to the adult form. This process of metamorphosis is of great interest to both developmental and evolutionary biologists, because what we see are major transitions in form not over long periods of time, but within a single generation.

We are so much accustomed to see a difference in structure between
the embryo and the adult, that we are tempted to look at this
difference as in some necessary manner contingent on growth. But there
is no reason why, for instance, the wing of a bat, or the fin of a
porpoise, should not have been sketched out with all their parts in
proper proportion, as soon as any part became visible. In some whole
groups of animals and in certain members of other groups this is the
case, and the embryo does not at any period differ widely from the
adult: thus Owen has remarked in regard to cuttlefish, "There is no
metamorphosis; the cephalopodic character is manifested long before
the parts of the embryo are completed." Landshells and fresh-water
crustaceans are born having their proper forms, whilst the marine
members of the same two great classes pass through considerable and
often great changes during their development. Spiders, again, barely
undergo any metamorphosis. The larvae of most insects pass through a
worm-like stage, whether they are active and adapted to diversified
habits, or are inactive from being placed in the midst of proper
nutriment or from being fed by their parents; but in some few cases,
as in that of Aphis, if we look to the admirable drawings of the
development of this insect, by Professor Huxley, we see hardly any
trace of the vermiform stage.

Why do some lineages undergo amazing processes of morphological change over their life histories, while others quickly settle on a single form and stick with it through their entire life? In some cases, we can even find closely related species where one goes through metamorphosis, and another doesn't; this is clearly a relatively labile character in evolution. And one of the sharpest, clearest examples of this fascinating flexibility is found in the sea urchins.

Sea urchins have a characteristic pattern of development. They form a hollow ball of cells that subsequently forms a simple gut and mouth, develops an internal skeleton, and turns into a pluteus larva, which looks a little bit like a spiny space shuttle. The pluteus is planktonic, and uses cilia to sweep food particles into its mouth as it drifts, accumulating nutrients for the next stage of its growth, when it metamorphosis into a bottom-dwelling, radially symmetric creature that will spend most of its life scraping algae off of rocks, and occasionally spewing out quantities of gametes to begin the cycle anew. This is called indirect development, because the individuals have to pass through a very specific and very different larval stage before they can become adults.

This pattern is not universal, however. Some species of sea urchins, instead of requiring their progeny to glean a living from the sea before they can become adults, pack their eggs with enough nutrients that they can skip the free-floating larval stage, and go directly to the adult form. This is called direct development, plainly enough. It has the advantage that the progeny can bypass the long and risky larval stage, but the disadvantage that the parents have to invest much more in each egg (a direct developer's egg may have 100 times the volume of an indirect developer's) and consequently produce fewer eggs. It's a classic life-history trade-off: will you have many children to whom you give relatively little attention, hoping that one or two will get lucky, or do you produce fewer children, but put a lot of effort into each one so that most will succeed?

The sea urchins show different lineages following different strategies. In the diagram below, all have similar embryos, but some produce spiny larvae that feed (the indirect developers) and others produce fat barrel-shaped larvae that simply develop straight into the adult form.

Evolution of developmental mode in sea urchins. Adult and larval forms are mapped on a phylogeny containing the major living sea urchin orders. Adult morphology is conservative in most lineages, but sand dollars and heart urchins have greatly modified adults. The pluteus larva is maintained by most lineages, and is the primitive mode of development. Direct development has arisen independently in over 20 lineages of sea urchins.

Some of the best examples of strategic diversity are found in the genus Heliocidaris, illustrated below. Heliocidaris tuberculata is an indirect developer, with the standard pluteus larva. It's sibling species, Heliocidaris erythrogramma, is a direct developer: it's embryo forms a barrel-shaped blob that goes on to make an adult. The adults of these two species are essentially indistinguishable, but the route they take to that point is very, very different.

Evolutionary changes in development and evolutionary events in the actin gene family mapped on the phylogeny of some camarodont sea urchins. Cladistic events: (1) novel CyU gene, (2) duplication of CyII and CyIII genes, (3) reduced number of actin genes expressed embryonically, (4) loss of CyII embryonic expression, (5) loss of CyIII gene, (6) switch to direct development.

These radically different patterns of development are not particularly difficult to explain in evolution, and I'd always considered it a straightforward shift that could be explained by a passive loss of embryonic properties, as has been explained by Wray and is illustrated below. Variations that increase parental investment in progeny would be selected for if they helped speed up the rate of development, and led to more progeny reaching adulthood, so a lineage could gradually increase egg size, nutrient content, etc. The larval stage might still form, but it becomes increasingly superfluous—they don't need to eat, so mutations that knocked out feeding structures would not be deleterious, and would accumulate. The loss of those genes might also have an advantage in simplifying and further increasing the rate of development. This passive loss of unnecessary larval features would eventually lead to the blob-like schmoo larva.

Wray's theory of the transitional steps during the
evolution of direct-developing larvae. This model suggests that
developmental changes occur neutrally after feeding is lost and
that the reduction in time to metamorphosis is a relatively late
step in the process.

Now some work by Smith et al. suggests that the progression above may not be correct, at least for some species. The acquisition of direct development may be less a passive loss of unneeded larval features, and more a matter of selection for accelerated growth of adult features.

In the diagram above, there is a step where a "facultative feeding" larva is produced. This is an animal that has all the functional specializations of an indirect developing urchin, and can and will eat anything it can find, but if you starve it, it's fine—it will go on and develop into an adult on the energy Mom Urchin packed into the egg. One of the nice things about the biodiversity in urchins is that you can find species that exemplify the first step, the obligate feeding pluteus, and other species that represent the last, the obligate nonfeeding schmoo, and also intermediate species that produce the facultative feeding larva. One such species is Clypeaster rosaceus.

You need to know one more thing about urchins to interpret this next image. The larvae sets aside a piece of its internal structure as an adult rudiment that will eventually grow to form the bulk of the adult body. If you're familiar with Drosophila development, there's a similar phenomenon: the larva sets aside small chunks of tissue called imaginal discs that will eventually mature into the adult cuticle and other structures. In the echinoderms, the part that will make the adult is the left coelom, a hollow ball of cells tucked away inside.

In this series of photographs, the top row shows the adults — the middle column, Clypeaster rosaceus is most interesting as the species that will form the intermediate, facultatively feeding larva. The second row shows the larvae: on the left, the obligate feeder of the indirect developer, in the middle, the facultative feeder of C. roaceus, and on the right, the non-feeding schmoo larva. The bottom row is most interesting: it's a section sliced through each of the larvae so that you can see what the left coelom (circled) looks like.

Clypeaster rosaceus maintains the same elaborate larval outward morphology and ability to feed of the typical indirect developer
C. subdepressus (D,E) Surprisingly though C. rosaceus forms a large left coelom (H) before the end of gastrulation, which is remarkably
similar to the development of the direct developer H. erythrogramma (I). Regardless of the larval development, all three species form relatively
similar adults (A,B,C). First row: adults; second row: larvae; third row: larvae showing the left coelom. G is an image from the whole embryo,
and H and I are stained sections of the larvae. Plc, presumptive left coelom; lc, left coelom.

Notice that the two larvae on the left look most similar to one another on the outside, but it's the two on the right that most resemble each other on the inside. C. rosaceus, to all outward appearances, looks like a typical pluteus, but internally what it has done is committed to an early, substantial investment in the adult rudiment of the left coelom. This isn't a passive difference at all. C. rosaceus has acquired a mutation that shifts the timing of development, and accelerates the growth of the left coelom. This shift precedes any acquisition of the barrel-like morphology seen in the rightmost animal.

Now we have to rethink the earlier model. The transformation to a direct developer requires a couple of positive changes: greater maternal investment in the egg, and timing changes in the regulation of development of adult structures in the embryo and larva. There's almost certainly some positive feedback going on there, too—timing shifts would demand more available energy to build those structures, and greater maternal investment would permit the acceleration, which would in turn demand more energy, and the cycle would go on. In the diagram of Raff's model of the evolution of this transition, you can see that acceleration of the timing and the more rapid metamorphic transformation have been moved up to be among the earliest events.

Based on the observation that C. rosaceus, a
facultative feeder, forms a large left coelom relatively early in
development, this shows a new theory for the evolution of direct
development. We suggest that developmental changes happen
before the ability to feed is lost and these developmental
changes may be driven by selection for rapid metamorphosis.

Furthermore, if you look at the last stages of the change, it's no longer a passive loss of larval features: it's an active reallocation of cells in the embryo and larva to adult fates. There are obviously some losses (refer back to that Heliocidaris cladogram: we know that one of the later steps in H. erythrogramma was loss of an actin gene), but simple loss is insufficient to explain the dynamic redirection of cells and tissues to new roles.

I think this is also a reflection of a shift we're making in thinking about development. Often in the literature we see development taken for granted as a relatively passive process, where the intricate machinery that assembles the organism responds to accommodate functional changes; it's a black box that is assumed to react to input, rather than driving change. In the case of these sea urchins, what we're finding is that regulatory changes in early developmental mechanisms are preceding the overt morphological changes that first drew our attention to the particular case. We also saw a similar phenomenon in the blind cavefish where what was once thought to be a passive accident of loss of inessential genes is turning out to be a more complicated story of gene regulatory interactions — it's all about how genes talk to one another.

Raff RA (1996) The Shape of Life: Genes, Development, and the Evolution of Animal Form^{(amzn/b&n/abe/pwll)}. The University of Chicago Press, Chicago.

Smith MS, Zigler KS, Raff RA (2007) Evolution of direct-developing larvae: selection vs loss. BioEssays 29:566-571.