The “Leech Perspective” on the Counting Problem in Segmentation
Comparisons of indeterminate and determinate growth in clitellate annelids highlight two related and frustratingly unanswered questions of general interest in segmentation. First, how do embryos that do exhibit determinate growth count out the appropriate number of segments for their species. A related second question, that might be expected to shed light on the first, is how precisely they count out their segments. Many oligochaetes make well over 100 segments, and there appear to be species-specific differences in the number of adult segments. The exact number of segments varies among individuals, however, in a roughly normal distribution around some value. For example, newly hatched Lumbricus terrestris were reported as having 130-160 segments, with 150 as the most common value (Evans 1946), equivalent to a natural variation of around 10% from the mean value. In leeches, as described earlier, the number of segments is much smaller (32) and is also more tightly controlled. We are aware of no naturally occurring variants of this value, whereas a 10% variation would be expected to produce animals ranging from 29 to 35 segments or so.
In the 1970s, when annelids and arthropods, as segmented protostomes, were considered as likely sister taxa, hypotheses were based in part on efforts to homologize the segmentation process in Helobdella with what was known about Drosophila. Thus, one model to explain the precise control of segment number in leeches was that the teloblast nuclei might undergo five rounds of syncytial nuclear proliferation (in M and О/P teloblasts, or six in N and Q teloblasts) after which the established set of 32 or 64 segmental founder cells would be counted out by successive cellu- larization events. This idea was ruled out when the (then) new technique of staining nuclear DNA with fluorescent Hoechst dyes revealed that the teloblasts are mono- nucleate and produce blast cells by standard mitoses (Zackson 1982).
Another significant outcome from this work was the discovery that the teloblasts in each lineage undergo additional mitoses after their full complement of segmental founder cells has been produced—moreover, the number of these so-called supernumerary blast cells varies from teloblast to teloblast within and among embryos
(Zackson 1982; Desjeux and Price 1999; Yoo et al. unpublished). Thus, while the total number of segments is tightly controlled, the numbers of blast cells produced is not. One interpretation of these observations is that variability in the number of stem cell divisions is an ancestral feature in clitellate annelids, and that the leech clade has superimposed a distinct secondary patterning process that gives precise control over the number of segments produced. The molecular bases of both of these inferred processes remain to be determined.
It should also be noted that the specific values cited for segment numbers in leech taxa, e.g., 15 for branchiobdellids, 29 for acanthobdellids, and 32 for true leeches, are based on necessarily subjective interpretations of morphological analyses (Purschke et al. 1993). Such analyses are particularly problematic for the head region; in leeches, for example, the number of segments is also reported as being 33 or 34 based on different interpretations of head morphology (Verdonschot 2015; Sawyer 1986). The value of 32 segments used in this chapter is based on the count of serially homologous neural metamers (4 in the head, 21 in the midbody, and 7 in the tail) (Muller et al. 1981). This value was reinforced by results from cell lineage tracing that showed that only 32 serially homologous segmental primordia arise from the teloblasts of the PGZ and that other head structures have a distinct embryonic origin.