Physical constraints

There are only about three dozen animal phyla, constituting the major body plans of the animal kingdom. One can easily imagine other types of body plans and animals that do not exist. (Science fiction writers do it all the time.) Why aren't there more major body types among the animals? To answer this, we have to consider the constraints that development imposes on evolution. There are three major classes of constraints on morphogenetic evolution.

First, there are physical constraints on the construction of the organism. The laws of diffusion, hydraulics, and physical support allow only certain mechanisms of development to occur. One cannot have a vertebrate on wheeled appendages (of the sort that Dorothy saw in Oz) because blood cannot circulate to a rotating organ; this entire possibility of evolution has been closed off. Similarly, structural parameters and fluid dynamics forbid the existence of 5-foot-tall mosquitoes.

The elasticity and tensile strengths of tissues is also a physical constraint. The six cell behaviors used in morphogenisis (cell division, growth, shape change, migration, death, and matrix secretion) are each limited by physical parameters, and thereby provide limits on what structures animals can form. Interactions between different sets of tissues involves coordinating the behaviors of cell sheets, rods, and tubes in a limited number of ways (Larsen 1992).

Morphogenetic constraints

There are also constraints involving morphogenetic construction rules (Oster et al. 1988). Bateson (1894) and Alberch (1989) noted that when organisms depart from their normal development, they do so in only a limited number of ways. Some of the best examples of these types of constraints come from the analysis of limb formation in vertebrates. Holder (1983) pointed out that although there have been many modifications of the vertebrate limb over 300 million years, some modifications (such as a middle digit shorter than its surrounding digits) are not found. Moreover, analyses of natural populations suggest that there is a relatively small number of ways in which limb changes can occur (Wake and Larson 1987). If a longer limb is favorable in a given environment, the humerus may become elongated, but one never sees two smaller humeri joined together in tandem, although one could imagine the selective advantages that such an arrangement might have. This observation indicates a construction scheme that has certain rules.

The rules governing the architecture of the limb may be the rules of the reaction-diffusion model (outlined in Chapter 1; Newman and Frisch 1979). Oster and colleagues (1988) found that the reaction-diffusion model can explain the known morphologies of the limb and can explain why other morphologies are forbidden. The reaction-diffusion equations predict the observed succession of bones from stylopod (humerus/femur) to zeugopod (ulna-radius/tibia-fibula) to autopod (hand/foot). If limb morphology is indeed determined by the reaction-diffusion mechanism, then spatial features that cannot be generated by reaction-diffusion kinetics will not occur.

Evidence for this mathematical model comes from experimental manipulations, comparative anatomy and cell biology. When an axolotl limb bud is treated with the anti-mitotic drug colchicine, the dimensions of the limb are reduced. In these experimental limbs, there is not only a reduction in the number of digits, but a loss of certain digits in a certain order, as predicted by the mathematical model and from the “forbidden” morphologies. Moreover, these losses of specific digits produce limbs very similar to those of certain salamanders whose limbs develop from particularly small limb buds (Figure 22.26; Alberch and Gale 1983, 1985). The self-organization of chondrocytes into nodules can be modelled by the Turing equations, and TGF-β2 appears to have the properties of the activator molecule postulated by this hypothesis (Miura and Shiota 2000a,Miura and shiota 2000b). Thus, the use of reaction-diffusion mechanisms to construct limbs may constrain the possibilities that can be generated during development, because only certain types of limbs are possible under these rules.

Figure 22.26

Relationship between cell number and number of digits in salamanders. (A) The hindlimb of an axolotl (Ambystoma mexicanum) with its five symmetrical digits. (B, C) Digits on the axolotl hindlimb after the hindlimb bud was incubated in colchicine to reduce (more...)

Phyletic constraints

Phyletic constraints constitute the third set of constraints on the evolution of new types of structures (Gould and Lewontin 1979). These are historical restrictions based on the genetics of an organism's development. For instance, once a structure comes to be generated by inductive interactions, it is difficult to start over again. The notochord, for example, which is still functional in adult protochordates such as amphioxus (Berrill 1987), is considered vestigial in adult vertebrates. Yet it is transiently necessary in vertebrate embryos, where it specifies the neural tube. Similarly, Waddington (1938) noted that although the pronephric kidney of the chick embryo is considered vestigial (since it has no ability to concentrate urine), it is the source of the ureteric bud that induces the formation of a functional kidney during chick development (see Chapter 14).

Until recently, it was thought that the earliest stages of development would be the hardest to change, because altering them would either destroy the embryo or generate a radically new phenotype. But recent work (and the reappraisal of older work: Raff et al. 1991) has shown that alterations can be made to early cleavage without upsetting the final form. Evolutionary modifications of cytoplasmic determinants in mollusc embryos can give rise to new types of larvae that still metamorphose into molluscs, and changes in sea urchin cytoplasmic determinants can generate sea urchins that develop without larvae but still become sea urchins. In fact, while all the vertebrates arrive at a particular stage of development called the pharyngula, they do so by very different means (see Figure 1.5). Birds, reptiles, and fishes arrive there after meroblastic cleavages of different sorts; amphibians get to the pharyngula stage by way of radial holoblastic cleavage; and mammals reach the same stage after constructing a blastocyst, chorion, and amnion. The earliest stages of development, then, appear to be extremely plastic. Similarly, the later stages are very different, as the different phenotypes of mice, sunfish, snakes, and newts amply demonstrate. There is something in the middle of development, however, that appears to be invariant.

Raff (1994) argues that the formation of new body plans (Baupläne) is inhibited by the need for global sequences of induction during the neurula stage (Figure 22.27). Before that stage, there are few inductive events. After that stage, there are a great many inductive events, but almost all of them occur within discrete modules. During early organogenesis, however, there are several inductive events occurring simultaneously that are global in nature. At this stage, the modules overlap and interact with one another. In vertebrates, to use von Baer's example, the earliest stages of development involve specifying axes and undergoing gastrulation. Induction has not yet happened on a large scale. Moreover, as Raff and colleagues have shown (Henry et al. 1989), there is a great deal of regulative ability at these stages, so small changes in morphogen distributions or the position of cleavage planes can be accommodated. After the major body plan is fixed, inductions occur all over the body, but are compartmentalized into discrete organ-forming systems. The lens induces the cornea, but if it fails to do so, only the eye is affected. Similarly, there are inductions in the skin that form feathers, scales, or fur. If they do not occur, the skin or patch of skin may lack these structures, but the rest of the body is unchanged. But during early organogenesis, the interactions are more global (Slack 1983). Failure to have the heart in a certain place can affect the induction of eyes (see Figure 6.4). Failure to induce the mesoderm in a certain region leads to malformations of the kidneys, limbs, and tail. It is this stage that constrains evolution and that typifies the vertebrate phylum. Thus, once a vertebrate, it is difficult to evolve into anything else.

Figure 22.27

Mechanism for the bottleneck at the pharyngula stage of vertebrate development. (A) In the cleaving embryo, global interactions exist, but there are very few of them (mainly to specify the axes of the organism). (B) At the neurula to pharyngula stages, (more...)

Box

Canalization and the Release of Developmental Constraints.

Footnotes

†

Leibniz, probably the philosopher who most influenced Darwin, noted that existence must be limited not only to the possible but to the compossible. That is, whereas numerous things can come into existence, only those that are mutually compatible will actually exist (see Lovejoy 1964). So although many developmental changes are possible, only those that can integrate into the rest of the organism (or which can cause a compensatory change in the rest of the organism) will be seen.