During vertebrate embryonic development, somites are formed from the unsegmented presomitic mesoderm (PSM) by a highly regulated process. They are formed in pairs in an anterior-to-posterior progression within the PSM, on either side of the notochord and are transient structures, which ultimately go on to form the axial skeleton, the dermis and the skeletal musculature. Using f-aequorin (a Ca²⁺-sensitive bioluminescent reporter; Shimomura et al. 1990) and an ultrasensitive photon-imaging microscope (Webb et al. 1997), we have begun to: (1) visualize and characterize the Ca²⁺ signals that occur in the trunk during the segmentation period; (2) to explore the developmental significance and function of the Ca²⁺ signals; and (3) to explore the source and release mechanism of the Ca²⁺ generating the signals.

Imaging data indicate that localized Ca²⁺ elevations are generated within the PSM before the somites are formed, within specific regions of the maturing somites after they have formed and within the notochord. Our analysis so far has concentrated on the Ca²⁺ transients generated by the maturing somites after they have formed. These transients were not seen in every embryo examined (n = 15) in a regular, repeating manner (i.e. on either side of the notochord within every pair of formed somites). However, although appearing to be chaotic, when they did occur, all the transients appeared at the same location with respect to somite morphology (i.e. at the medial and lateral extremities of the somites), and were of similar amplitude, duration, and involved approximately the same number of cells. In order to explore the significance and function of these seemingly chaotic signals, we loaded embryos with either photolabile caged Ca²⁺ or caged Ca²⁺ buffer, then illuminated specific regions of formed somites. The precocious elevation of Ca²⁺ in embryos loaded with caged Ca²⁺ (NP-EGTA) caused the formation of shorter somites along the medio-lateral axis, compared with untreated controls. Moreover, when regions of active buffer were generated by localized illumination of embryos loaded with the caged Ca²⁺ chelator, diazo-2, this resulted in the formation of elongated somites along the medio-lateral axis, compared with the untreated controls.

We suggest therefore, that these seemingly chaotic Ca²⁺ transients play a role in defining the medio-lateral boundaries of the maturing somites. As they do not occur in a regular, predictable manner, we propose that they are not the primary signal that defines these boundary limits, but are generated only when required. They act, therefore, to reinforce the primary medio-lateral boundary definition signal only in situations where the primary signal is ineffective.

It has been reported that Ca²⁺ transients observed during somitogenesis in Xenopus originate from both cADPR and IP₃ receptor-activated Ca²⁺ stores (Ferrari & Spitzer, 1999), rather than from Ca²⁺ entry via extracellular sources. When f-aequorin-loaded zebrafish embryos were incubated in Ca²⁺-free medium containing EGTA during the segmentation period, Ca²⁺ transients were still visualized in the trunk, and somitogenesis occurred normally for a while before the embryos dissociated. Thus these data suggest that zebrafish are similar to Xenopus where the source of Ca²⁺ during somitogenesis is intracellular rather than extracellular. However, while treating embryos with U73122 (a phospholipase C blocker) resulted in medio-laterally elongated somites, ryanodine (a cADPR receptor antagonist) had no effect on somite morphology. Thus these results suggest that unlike Xenopus, in zebrafish Ca²⁺ is released through IP₃ receptor-activated stores alone during early somite maturation. Furthermore, we have demonstrated the localization of IP₃ receptors within the maturing somites via immunohistochemistry and have shown that they are functional by visualizing intracellular Ca²⁺ release induced by localized photo-activation of caged IP₃.

This work was supported by Hong Kong RGC-CERG Grant HKUST 6106/01M awarded to A.L.M. We thank Dr O. Shimomura, Dr Y. Kishi, and Dr S. Inouye for supplying us with f-aequorin.