Snakes and ladders: the ups and downs of animal segmentation.

ABSTRACT

Segmentation, the division of the body into repetitive modular subunits or metameres, is ubiquitous throughout the animal kingdom. This morphological motif appeared several times in widely divergent phyla, some without common segmented ancestors (Bateson 1894; Willmer 1990). Segmentation presents challenges to standard evolutionary narratives (Minelli and Fusco 2004), in part because segments are discrete structures, like rungs in a ladder, that are added or subtracted in an all-or-none fashion, and also because large changes in segment number can occur in evolutionary lineages with little sign of intermediate forms. Two recent papers, one on snakes (Gomez et al. 2008) and one on centipedes (Vedel et al. 2008), shed some light on these important questions. In vertebrates, segmentation takes the form of somitogenesis, in which paired blocks of tissue known as somites bud off at regular time-intervals from the presomitic mesoderm (PSM) that fl anks the notochord and proceed to give rise to vertebrae, ribs, muscle and dorsal dermis (Dequéant and Pourquié 2008). The numbers of segments in mammals, birds and fi sh are not very different, all falling well under 100, within a factor of 2 of each other. Some other groups, such as snakes, however, stand out by possessing an enormous number of vertebrae (130-500, compared to 65 in mouse, 55 in chicken, 33 in human and 31 in zebrafi sh; Vonk and Richardson 2008; Marx and Rabb 1972). While there has been much speculation as to how this atypical (for vertebrates) segmental phenotype may have conferred adaptive advantages to snakes and their ancestors (Willmer 1990), it seems remarkable, particularly in the context of the incrementalist scenarios favoured by the standard selectionist framework, that generation of such an extreme morphology was even attainable. The developmental dynamics disclosed in the snake and centipede studies show vividly how evolution of form can take abrupt turns. First, Gomez et al. (2008) showed that their experimental animal, the corn snake, makes its somites in a fashion similar to that of fi sh, birds and mammals. As previously predicted by Cooke and Zeeman (1976) and later shown experimentally by Olivier Pourquié and his colleagues (reviewed in Dequéant and Pourquié 2008), the molecular–genetic mechanism that underlies this process consists of a biochemical oscillator (known as the segmentation clock) and a gradient, or wavefront. The clock is now known to comprise the periodic expression of Notch pathway signalling components and, depending on the vertebrate class, Wnt and fi broblast growth factor (FGF) pathway components as well (Dequéant and Pourquié 2008). The wavefront, with its source at the tailbud, consists, at a minimum, of FGF8 (Dequéant and Pourquié 2008). The FGF gradient serves as gate for the formation of the somites in the following fashion the PSM, which is locally synchronous with respect to the clock, reacts to attaining a specifi c clock-value (i.e. a critical concentration of one of the periodically changing components) by creating a fi ssure, but the tissue only does this when it is located at a point of the embryo’s axis where the FGF8 concentration is below a critical value. Because of the factor’s graded distribution, this position is substantially anterior to the tailbud. As the tailbud grows caudally, the shallow end of the gradient regresses in the same direction, progressively allowing new blocks of the PSM to bud off from the as-yet unsegmented region when the critical clockvalue next recurs in the newly disinhibited tissue. The snake embryo exhibited cyclic expression of Lunatic fringe (lfng), an enzyme of the Notch signalling pathway, as well as an FGF gradient (Gomez et al. 2008). The wavefront in snake embryos regressed caudally by one somite length every time a somite was formed, similar to what is observed in chicken, mouse and zebrafi sh models of somitogenesis (Dequéant and Pourquié 2008; Holley 2007). Thus.