The rod-shaped notochord is one of the first prominent structures to form in the early embryo, and its evolutionary development was a key advance that allowed early chordates to grow larger and swim in the primordial oceans. In modern vertebrates, the notochord consists of two cell layers: an inner core of cells that each contain a large, fluid-filled vacuole and a surrounding epithelial layer that secretes an extracellular matrix sheath (1). The vacuolated cells persist in adults as part of the intervertebral discs that connect adjacent vertebrae, yet the origin and function of the notochord vacuoles remains unknown. Ellis et al. now show that the vacuoles are lysosome-related organelles that control both embryonic elongation and spine morphogenesis (2).

Signals from the notochord help guide embryonic patterning, but the structure also plays a mechanical role in development, acting as a skeleton before the embryo forms bones. One study using dissected notochords has suggested that osmotic swelling of the notochord vacuoles could provide a mechanical force to lengthen and straighten embryos along their anterior-posterior axis (3). This idea intrigued Michel Bagnat, from Duke University in Durham, NC, who studies how fluid secretion contributes to tissue morphogenesis in zebrafish. “We’re interested in fluid pressure as a developmental force, and [notochord vacuoles] are huge fluid-filled structures,” Bagnat says.

“The vacuoles… provide a biophysical framework for putting together the spine.”

Bagnat and his colleagues Kathryn Ellis and Jennifer Bagwell first examined the formation of notochord vacuoles in zebrafish embryos (2). In theory, these large membrane-bound compartments, which occupy most of the inner notochord cells’ volume, could be formed by either endocytic membrane transport or by biosynthetic vesicle traffic through the Golgi apparatus. Ellis et al. found that blocking transport to or from the Golgi disrupted vacuole biogenesis, whereas inhibiting endocytosis had no effect on vacuole formation or maintenance.

Mutations in several genes involved in sorting and transport from the Golgi to late endosomes and lysosomes—including components of the H+-ATPase that acidifies intracellular compartments—also inhibited notochord vacuole formation. But to determine the vacuole’s identity, Ellis et al. needed to find a marker for the organelle. “We saw that some lysosomal markers, like LAMP1, localized to the vacuole,” Bagnat recalls. “But other lysosomal markers weren’t found there.”

This suggested that notochord vacuoles might be lysosome-related organelles (LROs), a class of intracellular compartments that includes cell-specific organelles like melanosomes. Indeed, Ellis et al. found that Rab32a, a GTPase required for melanosome biogenesis (4), localized to notochord vacuoles and that a dominant-negative version of Rab32a inhibited vacuole formation. “If you put all the elements together, it meets all the criteria for an LRO,” Bagnat says. “So then, having defined what the compartment is, we could start targeting it.”

To investigate the vacuole’s function, Ellis et al. disrupted its function in zebrafish embryos by expressing dominant-negative Rab32a or by hyperactivating the Notch signaling pathway to inhibit differentiation of the inner, vacuolated notochord cells. As predicted (3), embryos lacking notochord vacuoles were shorter than controls. “But the embryos weren’t curved, so vacuoles aren’t important for embryo straightening,” Bagnat explains.

The real surprise, however, came when Ellis and Bagwell let the embryos develop further and form vertebrae. Fish lacking notochord vacuoles developed kinked spines, possibly because their notochords are unable to resist the force generated by osteoblasts as they invade the notochord to initiate bone formation. “So, in early development, the vacuoles help elongate the embryonic axis,” Bagnat says, “and later, they provide a biophysical framework for putting together the spine.”

“So now we’re using live imaging to understand spine formation and how defects in the vacuoles result in kinks,” Bagnat continues. His team also wants to understand vacuole biogenesis by identifying the ion channels that regulate osmotic swelling and by determining how membrane trafficking is controlled so that inner notochord cells form one large vacuole instead of multiple small ones.

1.
Yamamoto
M.
et al
.
2010
.
Development.
137
:
2527
2537
.
2.
Ellis
K.
et al
.
2013
.
J. Cell Biol.
.
3.
Adams
D.S.
et al
.
1990
.
Development.
110
:
115
130
.
4.
Wasmeier
C.
et al
.
2006
.
J. Cell Biol.
175
:
271
281
.

Author notes

Text by Ben Short