Caspase enzymes promote cell death, but are also involved in sperm development in fruit flies. The discovery that, in sperm, caspase activation is restricted to the surface of organelles called mitochondria sheds light on this unusual role.
Protease enzymes called caspases are renowned killers, cleaving proteins to execute a program of apoptotic cell death. As such, the discovery1 in 2003 that caspase activation is required for sperm differentiation in the fruit fly Drosophila melanogaster came as a surprise. Writing in Developmental Cell, Aram et al.2 report that the restriction of caspase activity to the surfaces of organelles called mitochondria allows the enzymes to exert this alternative effect.
During the development of Drosophila sperm, precursors called spermatids that are linked to one another by cytoplasmic bridges mature simultaneously. Their nuclei elongate, their mitochondria fuse to form two large aggregates, new organelles form and membranes are added around each sperm cell3. At the end of this process, organelles and cytoplasmic materials that are not needed in the mature sperm are removed in vesicles, and the spermatids separate from each other in a process called individualization.
Individualization and disposal of cytoplasmic material both proceed from head to tail along spermatids, and caspases are activated in a head–tail gradient. The discovery that caspase inhibitors block individualization and cause male sterility provided the first evidence that caspases were involved in sperm development1.
In 2007, the biologist Eli Arama and colleagues screened sterile male flies for mutants that could not activate caspases during sperm individualization4. This revealed that a testis-specific enzyme called Cullin-3-based ubiquitin ligase (CRL3), which attaches ubiquitin molecules to proteins to modify the proteins' behaviour, is required for caspase activation. The group proposed that CRL3 ubiquitinates an apoptosis-inhibiting protein called Bruce, which is then degraded, enabling caspase activation.
Subsequently, Arama's laboratory identified a protein called Soti that functions as a CRL3 inhibitor5 by competing with CRL3 targets to bind to the enzyme. Soti is concentrated in the tail region of spermatids and its levels form a gradient opposite to that of activated caspases. Thus, it has been proposed that CRL3 determines the level of activated caspase and that Soti inhibits caspase activation5. This mechanism explains how a caspase gradient forms during individualization, but why activated caspase does not kill spermatids has remained a mystery.
In the group's latest study, Aram et al. found that a testis-specific version of the mitochondrial enzyme succinyl-CoA synthetase (SCS) mediates caspase activity in spermatids. SCS is a key enzyme in a process called the Krebs cycle, which generates energy in all aerobic organisms. The enzyme has α and β subunits, and the β subunit exists in two forms, which determine whether the enzyme synthesizes energy-carrying ATP (A-SCSβ) or GTP molecules (G-SCSβ) from the Krebs-cycle intermediate succinyl-CoA.
Aram and colleagues demonstrated that testis-specific A-SCSβ has a Krebs-cycle-independent role in caspase activation. They found that levels of A-SCSβ were elevated at the onset of spermatid individualization, and that the protein serves as an anchor to tie CRL3 to the surface of mitochondria (Fig. 1). Moreover, A-SCSβ prevented Soti from binding to CRL3, blocking its inhibitory activity. It seems, therefore, that A-SCSβ keeps caspase activity at low levels and limits it to a restricted zone in the central region of the cylinder-like spermatids. This prevents the proteins from spreading to the nucleus and plasma membrane, where many caspase targets are located.
This intriguing proposal is puzzling for several reasons. First, given that A-SCSβ is normally located within mitochondria, what mediates its release to the external surface? In mammalian apoptosis, proteins of the Bcl-2 family stimulate the release from mitochondria of another apoptosis-promoting protein, cytochrome c, which can then activate caspases. Although Bcl-2 family members do not seem to play a major part in fruit-fly apoptosis, some members are required for the death of around one-third of early-stage sperm precursors called spermatogonia as a normal part of sperm maturation6. Thus, Bcl-2 family proteins might also be involved in releasing A-SCSβ from spermatid mitochondria.
Another Krebs-cycle enzyme, fumarase, also has both mitochondrial and cytoplasmic functions7. Like A-SCSβ, fumarase is synthesized in the cytoplasm, and transported into mitochondria thanks to a mitochondrion-targeting sequence in its amino-terminal domain. However, some fumarase molecules are thought to return to the cytoplasm during the import process8. Similarly, some of the A-SCSβ that is synthesized at the onset of sperm individualization might remain at the mitochondrial surface.
Another question is how CRL3 activates caspases. As previously proposed4, CRL3-mediated ubiquitination of Bruce, followed by Bruce degradation, could be the mechanism for caspase activation. Another CRL3 target is cytochrome c (ref. 2). In mammals, cytochrome c — in complex with a scaffold protein — cleaves procaspase proteins into active caspases. In fruit flies, testis-specific cytochrome c (Cyt-c-d) activates caspases in spermatids1. Because ubiquitination can regulate protein–protein interactions9, perhaps ubiquitinated, but not de-ubiquitinated, Cyt-c-d can activate caspases.
During sperm individualization, cytoplasmic material is disposed of in vesicular waste bags. Cullin-mediated ubiquitination is involved in establishing the structure of an organelle called the Golgi, which packages cargo into vesicles and also has a role in sorting the proteins for vesicle packaging10. Because several proteins in the Golgi (and in the other organelles involved in vesicle transport, such as the endoplasmic reticulum and lysosomes) are targets of caspases, cytoplasmic protein disposal may be collaboratively controlled by CRL3-mediated ubiquitination and caspase-mediated protein cleavage.
Finally, the non-apoptotic activation of caspases has been observed in many other biological processes, in both vertebrates and invertebrates11. It will be fascinating to discover whether similar mechanisms or molecules are involved in these processes. Footnote 1
Arama, E., Agapite, J. & Steller, H. Dev. Cell 4, 687–697 (2003).
Aram, L. et al. Dev. Cell 37, 15–33 (2016).
Fabian, L. & Brill, J. A. Spermatogenesis 2, 197–212 (2012).
Arama, E., Bader, M., Rieckhof, G. E. & Steller, H. PLoS Biol. 5, e251 (2007).
Kaplan, Y., Gibbs-Bar, L., Kalifa, Y., Feinstein-Rotkopf, Y. & Arama, E. Dev. Cell 19, 160–173 (2010).
Yacobi-Sharon, K., Namdar, Y. & Arama, E. Dev. Cell 25, 29–42 (2013).
Monaghan, R. M. & Whitmarsh, A. J. Trends Biochem. Sci. 40, 728–735 (2015).
Yogev, O., Naamati, A. & Pines, O. FEBS J. 278, 4230–4242 (2011).
Mukhopadhyay, D. & Riezman, H. Science 315, 201–205 (2007).
Lu, A. & Pfeffer, S. R. Trends Cell Biol. 24, 389–399 (2014).
Kuranaga, E. & Miura, M. Trends Cell Biol. 17, 135–144 (2007).
About this article
Frontiers in Cell and Developmental Biology (2016)