Segmentation of the Trunk
Development of segmental units can be seen as the combination of three processes: elongation of the anteroposterior (AP) axis, segregation of cells into segmental units, and patterning of each segmental unit (Balavoine 2015). In most species in which segmentation has been studied, axial elongation of the body is achieved by cell proliferation at the posterior end of the body and is followed by the formation of boundaries that separate and specify fields of cells to form consecutive metamers. As detailed in other chapters, these metamers may arise sequentially (as is the case with vertebrates and short germ insects) or simultaneously (as in long germ insects) along the AP axis, and may correspond to segmental, para-segmental, or double- segmental primordia. The actual patterns of cell division leading to the formation of morphological pattern elements within the units are often variable (the variable and irregularly shaped clones of cells within Drosophila compartments, for example), and cell clones arising within one metamer are restricted from crossing boundaries separating them from adjacent metamers. This is analogous to creating a repeating pattern in a row of bushes by trimming them so that their branches do not intermingle, without regard for the branching patterns of the individual bushes. We refer to these processes as boundary-driven segmentation mechanisms.
Segmentation by boundary-driven mechanisms is so widespread among the commonly studied animal models that one may be forgiven in assuming that this is the only way to generate metamers and segments. However, most organisms used to study segmentation are either arthropods or vertebrates. In contrast, clitellate annelids are known to generate segments by another mechanism, in w'hich the components of each segment arise as the clonal descendants of individual cells. We call this mechanism lineage-driven segmentation. Returning to the analogy of the row of bushes, if the bushes are carefully pruned to achieve the same branching pattern for each bush, then the row of bushes will present a repeating pattern even if the branches of adjacent bushes intermingle. In groups showing lineage-driven segmentation, axial elongation and segment specification are achieved by the same process: cell division.
Among annelids, lineage-driven segmentation has been unambiguously described for clitellate annelids, particularly in leeches (see Chapter 7). In what appears to be the ancestral mode of development for this group, large yolky embryos exhibit a modification of the D quadrant cleavage pattern so that the equivalents of micromeres 2d and 4d are large cells (designated as DNOPQ and DM, respectively). Their further divisions yield a fixed set of ten large stem cells known as teloblasts, amenable to microinjection of cell lineage markers. This approach has been used to show that the teloblast progeny also undergoes stereotyped patterns of cell division to generate segments in a lineage-dependent manner, but there is still considerable debate about whether the remaining annelid groups also show a similar mechanism of segment formation. At the heart of the issue is the question of whether segment formation in polychaete annelids segments is also lineage-driven, or boundary-driven instead. Answering this question has proven difficult because most polychaete annelids have smaller embryos than clitellates and develop indirectly to produce an intermediate larval form. Depending on the group, larval stages have from zero to several segments, which can be more or less specialized (Anderson 1973; Rouse 2006). Having a mobile, active larval stage greatly complicates tracing cell fates. Furthermore, in many groups, activation of post-larval segment development requires larvae to reach metamorphosis, a process that is often difficult to induce under artificial culture conditions. For these reasons, embryonic development of marine annelids is much better studied than development at later life stages, including post-metamorphic development of the trunk’s segmental units. Despite these challenges, the last few decades have seen the emergence of two major non-clitellate annelid models: the errant nereid Platynereis dumenlii and the sedentary capitellid Capitella teleta (Zantke et al. 2014; Seaver 2016). Application of modern cell-tracing strategies are starting to clear the picture of how segmental units develop in marine annelids, although studies have given somewhat conflicting evidence.
Embryonic and larval development studies from both live specimens and histological sections were made for representatives of several families of marine annelids during the late 1800s and first half of the 1900s (Anderson 1973). Most of these studies show that the ectodermal component of trunk segments forms from proliferation of a ring of ectoteloblasts derived from the second quartet blastomere 2d, while the mesodermal component of trunk segments arises from ventrolateral bands that result from teloblastic proliferation of the paired M mesoteloblasts, descended from bilateral equal division of the fourth quartet blastomere 4d (Figures 4.2 and 4.3). However, a lack of cell-tracing tools at post-embryonic stages made it impossible to test if segment formation in these groups is lineage-driven (as seen in clitellate embryos) or else boundary-driven. While early presence of ecto- and mesoteloblasts similar to those of clitellates was evident in embryos, and developmental series showed likely clonal segment formation (Anderson 1973), most post-metamorphic annelids show no large, teloblast-like cells at their posterior growth zone (Seaver, Thamm, and Hill 2005). This conundrum, however, began to be solved with the application of a novel approach combining endogenous cell cycle reporters, in vivo time-lapse fluorescence imaging, and in silico cell tracing (Ozpolat et al. 2017). This work showed that in the nereid Platynereis dumerilii, the M cells daughter of the 4d cell initially divide as teloblasts to generate primordial germ cells and four larval segments, but after their eighth division, they divide more or less equally to form a ring of cells located immediately in front of the pygidium and internal to the ectot- eloblast rings—becoming the mesodermal stem cells of the posterior growth zone. In other words, ectodermal and mesodermal stem cells derived from 2d and 4d, respectively, initially use lineage-driven segmentation and then transition to what currently is best described as boundary-driven segmentation (see Chapter 10).
The number of segments formed during embryonic development can range from none (in indirect-developing species with larvae without segments, e.g., Oweniidae, Chaetopteridae), a few, often specialized larval segments (e.g., Serpulidae, Nereididae, Eunicidae, Tomopteridae), to several (e.g., Capitellidae, Spionidae, Pectinariidae, non-leech clitellates), or even the complete adult complement (e.g., leeches). Below we describe in varying degree of detail what is known about segment formation in representatives of four families that cover a broad phylogenetic spectrum: (1) Oweniafusiformis, a member of the early branching family Oweniidae (Figure 4.4A-B); (2) Platynereis dumerilii, a member of the family Nereididae of errant polychaetes (Figure 4.4E-F); (3) Capitella teleta, a member of the family Capitellidae of sedentary polychaetes (Figure 4.4G-H); and (4) Tubifex tubifex, a member of the family Naididae of clitellates (Figure 4.4I-J). Segment formation in glossiphoniid leeches (Helobdella spp.) is treated separately in Chapter 7.