Theoretical Criteria For Evolvability

The recent synthesis of evolutionary and developmental biology, or EvoDevo (Hall, 1998), has yielded important new insights into the molecular- and cellular-level properties that facilitate morphological change in evolving lineages. Here the focus is on four of these basic properties, reviewed by Conrad (1990), Gerhart and Kirschner (1997), and Kirschner and Gerhart (1998), that increase the potential for evolutionary change in the limb, particularly within the autopodial region. The forelimb autopod includes the carpus, metacarpus, and fingers, whereas the hindlimb autopod includes the tarsus, metatarsus, and toes. The first property of the autopod, which increases its potential for evolutionary change, is modularity or compartmentalization. Modules are defined in different ways by different researchers (Winther,

2001). Modularity can be identified from embryological studies in which a module is an isolatable, transplantable, and well-characterized landmark on the embryo (Gilbert et al., 1996). Such modules are discrete units of development requiring specific selector gene expression within a bounded spatial domain (Gilbert et al., 1996; see also Carroll et al., 2001, for a discussion of selector genes). A module can also be recognized in a group of related organisms as a subset of body plan elements that exhibits adaptive variation more or less autonomously (Von Dassow and Munro, 1999). The most important feature of modularity is autonomy in which one aspect of form may "explore" new structural variants without affecting another (Gerhart and Kirschner, 1997).

The second important prerequisite for evolvability is redundancy. In the case of a module, genetic redundancy protects old functions while at the same time allowing for the acquisition of new ones. In the case of regulatory genes, redundancy and duplication allow different «i-regulatory (e.g., enhancer) elements to be acquired over time leading to slight differences in spatiotemporal gene expression (Figure 1; Carroll et al., 2001; Chiu and Hamrick,

2002). Genetic redundancy is, therefore, a means of maintaining a "pool" of evolutionary novelty at the biochemical level (Wagner, 1996). As discussed in a later section, redundancy is best recognized from gene knockout experiments. One example comes from the Msx genes, a group of homeobox genes that are often expressed in overlapping domains during the formation of ectodermal organs such as teeth and hair (Noveen et al., 1995). They are functionally redundant at early stages of organ development and knockouts of either Msxl or Msx2 still form hair follicles. The expression of these genes differs at later stages of hair morphogenesis, when Msxl expression is down-regulated and Msx2 expression shifts from the germinal matrix to the root sheath (Satokata et al., 2000). Mice homozygous for the disrupted Msx2 sequence are viable and mice heterozygous for the disrupted Msx2 sequence show no phenotypic differences compared to normal mice. Thus, redundancy appears to reduce the lethality of mutation and may also facilitate divergence of gene function in the form of change in spatiotemporal expression.

The third criterion for evolvability is weak linkage. This refers to the dependence of one stage in a metabolic or transduction pathway on another. The genes of eukaryotes, particularly of metazoans, have large and complex «¿•-regulatory regions that function to activate (e.g., enhancers) or suppress (e.g., repressors) transcription of the gene of interest (Figure 1). Many enhancer proteins are known to bind with relatively low specificity to enhancer sequences (Kirschner and Gerhart, 1998). Moreover, there is evidence to suggest that «¿-regulatory regions, such as enhancer sequences, can evolve quite rapidly through processes, such as de novo evolution, from previously nonfunctional DNA sequences, duplication, and then divergence from existing regulatory sequences or modification of existing regulatory sequences (Carroll et al., 2001; Chiu and Hamrick, 2002). As Carroll et al. (2001) have noted, the probability of mutational change in enhancer sequences is relatively high; so high, in fact, that the probability of de novo

Coding sequence


-1 Coding sequence I---1 Coding sequence


Coding sequence

Figure 1. Evolutionary change in the expression of regulatory gene function via gene duplication followed by enhancer sequence (E1, E2, and E3) divergence. Modified from Carroll et al. (2001).

Coding seqqence enhancer evolution in Drosophila is approximately once per gene. Thus, genes expressed during the patterning of morphogenetic fields may acquire new spatiotemporal patterns of expression via changes in their cis-regulatory regions. This flexibility of transcriptional regulation facilitates evolutionary change in morphogenetic pathways via mutation in cis-regulatory sequences. These cis-regulatory mutations usually have relatively subtle phenotypic effects (Stern, 2000), such as altering the number of bristles occurring on the legs of fruit flies (Stern, 1998a).

The final requirement for evolvability is robusticity, which refers to the ability of a module to withstand mutational changes in cell number, cell arrangement, etc., during development so that these mutations result in nonlethal phenotypes outside the structural norm (Conrad, 1990). The complex cellular arrangements that characterize morphogenetic fields in vertebrates are precisely sculpted by the processes of cell adhesion, cell proliferation, and programmed cell death (apoptosis). Yet, relatively subtle variations in cellular patterning during morphogenesis can yield potentially significant (adaptively) changes in morphology. In the case of tooth morphology, Jernvall's (2000) patterning cascade model of tooth development predicts that minor variations in the diffusion of molecular signals from the primary enamel knot produce slight variations in the locations of secondary enamel knots that later form smaller cusps. The effect is cumulative, where the last cusps to form tend to be the smallest and most variable in terms of size and position. These data explain the high degree of intraspecific variability observed in tooth cusp formation among seals, and may also explain the repeated convergent evolution of small cusps such as the hypocone in early mammalian evolution (Hunter and Jernvall, 1995; Jernvall et al., 1996). Mammalian molar tooth cusp topography is, therefore, relatively robust so that differences in cusp size and number can be variable even within a single seal species (Phoca hispida) yielding relatively subtle, nonlethal variation upon which selection may act (Jernvall et al., 2001).

0 0

Post a comment