Deuterostome Origins and the Dorsal-Ventral Inversion Hypothesis

Lecture by Chris Cameron

BIOL 606 Session, University of Alberta, January 27, 1999.

Rapporteur: Tricia Abe

Chris Cameron began the lecture by posing the question, Where along the Deuterostome clade did dorsal-ventral inversion happen? The history of this investigation dates back to 1830 in France, when St. Hilaire proposed that the body plan of arthropods and vertebrates represent a common plan. Rathke and Kolliker (1842), discovered through embryogenesis that the ventral side of arthropods corresponds to the dorsal side of vertebrates. Leydid and Semper (1842) and Dohrn (1895) compared certain features of annelid organization to those of vertebrates. In 1926 Naef proposed the annelid hypothesis, which hypothesized that vertebrates pass through an annelid-like stage during development. Further, he conjectured that dorsal-ventral inversion first occurred in a hemichordate worm.

More recently, a homologous relationship between vertebrates and invertebrates is supported by a comparison of amino acid sequences of Xenopus and Drosophila. Arendt and Nabler-Jung found molecular similarities between these two groups (e.g. hydrophobic segments, cysteine-rich repeats) and suggested that these genes came from a common ancestor. Furthermore, homologous genes found in Drosophila (sog) and Xenopus (chordin) are similarly expressed in the ventral and dorsal surfaces, respectively. When Drosophila sog is ventrally injected into Xenopus gastrulae, it causes dorsalization of the ventral part of the gastrula, thus demonstrating molecular conservation of dorsal-ventral patterning mechanisms between arthropods and chordates. This supports St. Hilaire's idea that the dorsal side of vertebrates is equivalent to an inverted ventral side of arthropods.

In order to determine where the dorsal-ventral inversion happened along the deuterostomes, Cameron presented a survey of extant chordates. Dorsal and ventral surfaces were identified as those that evolved in relation to an organism's orientation with respect to a surface or to gravity. The echinoderms demonstrate no dorsal-ventral preference axis. Since body orientation is extremely plastic, it is unlikely that a dorsal-ventral inversion occurred in this clade. The hemichordates (Class Entrepneusta) have a stomochord in the anterior part of the body, and they also have a dorsal nerve cord. One might argue that these features are homologous to the vertebrate notochord and nerve cord, respectively. However, there is also a circumpharyngeal nerve cord extending along the ventral side of the animal. Further, the endostyle is located dorsally, while in chordates it is ventral. Thus the hemichordates are also uninformative with respect to elucidating the origins of dorsal-ventral inversion. The urochordates and cephalochordates are also shown to be highly uniformative for detecting a preference for a dorsal-ventral axis. In these clades, the animals are extremely small and swim erratically without a fixed dorsal-ventral orientation.

In light of weak evidence for dorsal-ventral axis preference in protovertebrates, it seems likely that the functional inversion of the dorsal-ventral axis probably occurred in the vertebrate clade. Although a labile dorsal axis is seen in other clades (e.g. ascidian tadpoles, cephalochordates), the dorsal-ventral axis appears to be constrained by the large body sizes found in vertebrates.

References: Chris Cameron provided this list of Dorso-ventral inversion references.


Discussant: Grant McIntyre

During the break, Keith Jackson told Cameron that in his labs he teaches students that the stomochord in hemichordates is NOT homologous to the vertebrate notochord. Cameron noted some of their similarities and differences. However, it appears to be a difficult task to determine homology of the stomochord and notochord based on embryology and morphology. Mark Wilson responded to this by saying that there is weak evidence for a homologous relationship between stomochords and notochords. Cameron added that a hemichordate is not even segmented, but only has paired gill slits going along the body.

Curt Strobeck made the point that he failed to see a switch from dorsal and ventral body surfaces in the tree shown during the talk. It appeared that in the Drosophila lineage, dorsal-ventral positioning evolved in one direction, and in the vertebrate lineage, it evolved in another direction. However, there was weak evidence for the proposed reversal of this character between lineages. Cameron replied that there were many other possible clades that he did not show, and it is often difficult to determine axis preference within other clades. The difficulty of circular reasoning when characterizing the dorsal/ventral surfaces was raised.

The point was made that in the hypothetical ancestral animal (an amphistome), dorsal/ventral surfaces were defined in relation to the gut and gastrulation process, so other animals (without clearly defined preferences for a dorso-ventral axis apparent to us) can have a defined dorso/ventral axis. Grant McIntyre asked what trait is more useful for determining dorsality; the an-veg axis or the lip of the blastopore? Rich Palmer wondered if the dorsal-ventral interior axis must be coupled to an external dorsal-ventral axis. McIntyre noted the potential for experiments addressing this question.

Discussion surrounding the definition of dorsal and ventral surfaces followed. Strobeck made the point that it is difficult to argue about a switch in the dorsal-ventral axis if we cannot define what constitutes dorsal and ventral surfaces. Although the orientation of the mouth has been used to define ventral surfaces, this is not a good character for definition. Sean Graham noted that the focal paper used the neural side as a defining character. An argument for defining the dorsal-ventral axis by mapping these traits onto cladograms was made.

The question of who the sister group of deuterostomes is, if not the arthropods, was asked. Bruce Heming said that comparative work with genes suggests the metazoans, although it is not clear how these genes evolve. Comparative analysis is needed. Strobeck asked how many genes the authors had mapped. Cameron was not sure. Although the common ancestor may not have had a nerve cord, perhaps it had the gene to set up the functional potential.

Palmer put up a recently published phylogenetic tree based on total evidence. The arthropods and annelids were shown to be surprisingly far apart, causing a hush to fall over an incredulous audience. Once the disquietude had settled, discussion of how to determine homology in morphological and molecular data followed.