Evolution and Morphogenesis

We shall never fully understand the process of evolution until we know how the environment affects the mechanisms that produce pattern and form in embryogenesis. Natural selection must act on the developmental programmes to effect change. We require, therefore, a morphological view of evolution, which goes beyond the traditional level of observation to a morphological explanation of the observed diversity. Later in this chapter we shall discuss some specific examples whereby morphogenesis has been experimentally influenced to produce early embryonic forms, early, that is, from an evolutionary point of view. This chapter has no mathematics per se and is more or less a stand-alone biological chapter. However, the concepts developed and their practical applications are firmly based on the models, and their analysis, presented and elaborated on in earlier chapters, particularly Chapters 2, 3,4 and 6.

Natural selection is the process of evolution in which there is preferential survival of those who are best adapted to the environment. There is enormous diversity and within species such diversity arises from random genetic mutations and recombination. We must therefore ask why there is not a continuous spectrum of forms, shapes and so on, even within a single species. The implication is that the development programmes must be sufficiently robust to withstand a reasonable amount of random input. From the extensive genetic research on the fruit fly Drosophila, for example, it seems that only a finite range of mutations is possible, relatively few in fact.

The general belief is that evolution never moves backward, although it might be difficult to provide a definition of what we mean by direction. If evolution takes place in which a vertebrate limb moves from being three-toed to four-toed, from a morphological view of evolution there is no reason, if conditions are appropriate, that there cannot be a transition 'back' from the four-toed to the three-toed variety. From our study of pattern formation mechanisms this simply means that the sequential bifurcation programme is different. In Section 7.4 we show an example where an experimentally induced change in the parameters of the mechanism of morphogenesis results in embryonic forms which, with the accepted direction of evolutionary change, means that evolution has moved backwards.

If we take the development of the vertebrate limb we saw in Section 6.6 in the last chapter that development was sequential in that the humerus preceded the formation of the radius and ulna and these preceded the formation of the subsequent cartilage patterns such as the phalanges. As a specific example, we argued that the formation of the humerus could cue the next bifurcation by influencing the geometry of the limb bud. We also saw how graft experiments could alter the pattern sequence and we showed how the result was a natural consequence when viewed from a mechanistic viewpoint.

So, intimately associated with the concept of bifurcation programmes, are discrete events whereby there is a discrete change from one pattern to another as some parameter passes through a bifurcation value. The possibility of discrete changes in a species as opposed to gradual changes is at the root of a current controversy in evolution, between what is called punctuated equilibrium and phyletic gradualism, which has raged for about the past few decades. (Neo-Darwinism is the term which has been used for punctuated equilibrium.) Put simply, punctuated equilibrium is the view that evolutionary change, or speciation and morphological diversification, takes place effectively instantaneously on geological time, whereas gradualism implies a more gradual evolution to a new species or a new morphology. The arguments for both come from the fossil records and different sets of data are used to justify each view—sometimes even the same set of data is used. Figure 7.1 schematically shows the two extremes.

From a strictly observational approach to the question we would require a much more extensive fossil record than currently exists, or is ever likely to be. From time to time newly discovered sites are described which provide fine-scaled palaeontologi-cal resolution of speciation events. For example, Williamson (1981a,b) describes one of these in northern Kenya for molluscs and uses it to argue for his view of evolution. Sheldon (1987) gathered fossil data, from sites in mid-Wales, on trilobites (crab-like marine creatures that vary in size from a few millimetres to tens of centimetres) and on the basis of his study argues for a gradualist approach. From an historical point of view, the notion of punctuated equilibrium was very clearly put by Darwin (1873) himself in the 6th and later editions of his book, On the Origin of Species, in which he said (see the summary at the end of Chapter XI, p. 139), 'although each species must have passed

Figure 7.1. Punctuated equilibrium implies that as we move through geological time changes in speciation occur very quickly (on geological time) as compared with stasis, the period between speciation events. Phyletic gradualism says that speciation and diversification are a gradual evolutionary process.

Figure 7.1. Punctuated equilibrium implies that as we move through geological time changes in speciation occur very quickly (on geological time) as compared with stasis, the period between speciation events. Phyletic gradualism says that speciation and diversification are a gradual evolutionary process.

through numerous transitional stages, it is probable that the periods, during which each underwent modification, though many and long as measured by years, have been short in comparison with the periods during which each remained in an unchanged condition.' (The corresponding passage in the first edition is in the summary of Chapter X, p. 139.)

From our study of pattern formation mechanisms in earlier chapters the controversy seems artificial. We have seen, particularly from Chapters 2 to 6, 4 and 5 that a slow variation in a parameter can affect the final pattern in a continuous and discrete way. For example, consider the mechanism for generating butterfly wing patterns in Section 3.3. A continuous variation in one of the parameters, when applied, say, to forming a wing eyespot, results in a continuous variation in the eyespot size. The expression (equation (3.24), for example) for the radius of the eyespot shows a continuous dependence on the parameters of the model mechanism. In the laboratory the varying parameter could be temperature, for example. Such a continuous variation falls clearly within the gradualist view of evolutionary change.

On the other hand suppose we consider Figures 2.14(b) and (c) which we reproduce here as Figures 7.2(a) and (b) for convenience. It encapsulates the correspondence be-

Figure 7.2. (a) Solution space for a reaction diffusion mechanism (system (2.8) in Chapter 2) with domain size, y , and morphogen diffusion coefficient ratio, d. (b) The spatial patterns in morphogen concentration with d, y parameter values in the regions indicated in (a). (From Arcuri and Murray 1986)

tween discrete patterns and two of the mechanism's dimensionless parameters. Another example is given by Figure 4.21 in Chapter 4 on the formation of teeth primordia where again the parameter space indicates abrupt changes in the pattern formed. Although Figures 7.2 and 4.21 are the bifurcation space for a specific reaction diffusion mechanism we obtain comparable bifurcation spaces for the mechanochemical models in the last chapter and other pattern formation mechanisms. In Section 6.6 in the last chapter we noted that the effect of a tissue graft on the cartilage patterns in the developing limb was to increase cell proliferation and hence the size of the actual limb bud. Let us, for illustrative purposes, focus on the development of the vertebrate limb. In Figure 7.2(a), if we associate cell number with domain size y we see that as y continuously increases for a fixed d, say, d = 100, we have bifurcation values in y when the pattern changes abruptly from one pattern in Figure 7.2(b) to another. So, a continuous variation in a parameter here effects discontinuous changes in the final spatial pattern. This pattern variation clearly falls within a punctuated equilibrium approach to evolution.

Thus, depending on the mechanism and the specific patterning feature we focus on, we can have a gradual or discontinuous change in form. So to reiterate our comment above it is clear that to understand how evolution takes place we must understand the morphogenetic processes involved.

Although the idea that morphogenesis is important in understanding species diversity goes back to the mid-19th century, it is only relatively recently that it has been raised again in a more systematic way by, for example, Alberch (1980) and Oster and Alberch (1982); we briefly describe some of their ideas below. Oster et al. (1988) presented a detailed study of vertebrate limb morphology, which is based on the notion of the morphological rules described in the last chapter. The latter paper presents experimental evidence to justify their morphogenetic view of evolutionary change; later in the chapter we describe their ideas and some of the supporting evidence.

Morphogenesis is a complex dynamic process in which development takes place in a sequential way with each step following, or bifurcating, from a previous one. Alberch (1980, 1982), and Oster and Alberch (1982) suggest that development can be viewed as involving only a small set of rules of cellular and mechanochemical interactions which, as we have seen from previous chapters, can generate complex morphologies. Irrespective of the actual mechanisms, they see developmental programmes as increasingly complex interactions between cell populations and their gene activity. Each level of the patterning process has its own dynamics (mechanism) and it in turn imposes certain constraints on what is possible. This is clear from our studies on pattern formation models wherein the parameters must lie in specific regions of parameter space to produce specific patterns; see, for example, Figure 7.2. Alberch (1982) and Oster and Alberch (1982) encapsulate their ideas of a developmental programme and developmental bifurcations in the diagram shown in Figure 7.3.

If the number or size of the mutations is sufficiently large, or sufficiently close to a bifurcation boundary, there can be a qualitative change in morphology. From our knowledge of pattern formation mechanisms, together with Figure 7.3, we can see how different stability domains correspond to different phenotypes and how certain genetic mutations can result in a major morphological change and others do not. Not only that, we can see how transitions between different morphologies are constrained by the topology of the parameter domains for a given morphology. For example, a transition be-

Figure 7.3. A schematic diagram showing how random genetic mutations can be filtered out to produce a stable phenotype. For example, here random genetic mutations affect the size of the various developmental parameters. With the parameters in a certain domain, 1 say, the mechanisms create the specific pattern 1 at the next level up; this is a possible phenotype. Depending on the size of the random mutations we can move from one parameter domain to another and end up with a different phenotype. There is thus a finite number of realisable forms. At the next stage selection takes place and the final result is a number (reduced) of realised phenotypes. (From Oster and Alberch 1982)

Figure 7.3. A schematic diagram showing how random genetic mutations can be filtered out to produce a stable phenotype. For example, here random genetic mutations affect the size of the various developmental parameters. With the parameters in a certain domain, 1 say, the mechanisms create the specific pattern 1 at the next level up; this is a possible phenotype. Depending on the size of the random mutations we can move from one parameter domain to another and end up with a different phenotype. There is thus a finite number of realisable forms. At the next stage selection takes place and the final result is a number (reduced) of realised phenotypes. (From Oster and Alberch 1982)

tween states 1 and 2 is more likely than between 1 and 5 and furthermore, to move from 1 to 5 intervening states have to be traversed. An important point to note is that existing morphological forms depend crucially on the history of their past forms. The conclusion therefore is that the appearance of novel phenotypic forms is not random, but can be discontinuous. As Alberch (1980) notes, 'We need to view the organism as an integrated whole, the product of a developmental program and constrained by developmental and functional interactions. In evolution, selection may decide the winner of a given game but development non-randomly defines the players.'

Developmental Constraints

In previous chapters we have shown that, for given morphogenetic mechanisms, geometry and scale impose certain developmental constraints. For example, in Chapter 3, Section 3.1 we noted that a spotted animal could have a striped tail but not the other way round. In the case of pattern formation associated with skin organ primordia, as discussed in the last chapter, we have mechanical examples which exhibit similar developmental constraints. Holder (1983) carried out an extensive observational study of 145 hands and feet of four classes of tetrapod vertebrates. He concluded that developmental constraints were important in the evolution of digit patterns.

Figure 7.4 (refer also to Chapter 4, Figure 4.10) shows some of the key mechanical steps in the early development of certain skin organs such as feathers, scales and teeth. In Section 6.6 we addressed the problem of generating cell condensation patterns which we associated with the papillae. In the model for epithelial sheets, discussed in Section 6.8, we saw how spatially heterogeneous patterns could be formed and even initiated by the

Figure 7.4. Key mechanical events in the dermis and epidermis in development of skin organ primordia. (After Oster and Alberch 1982)

dermal patterns. Odell et al. (1981) showed how buckling of sheets of discrete cells such as in Figure 7.4, could arise. From the sequential view of development we might ask whether it is possible to move onto a different developmental pathway by disrupting a mechanical event. There is experimental evidence that a transition can be effected from the scale pathway to the feather pathway, for example, Dhouailly et al. (1980), by treating the skin organ primordia with retinoic acid. In their experiments feathers were formed on chick foot scales.

Was this article helpful?

0 0

Post a comment