bacterial substrates until food becomes limited. This situation triggers a
differentiation program that involves aggregation of starving cells (mediated by cAMP
signaling and chemotaxis) to form a large aggregate that differentiates into a fruiting
body containing spore cells and a stalk. Cells forming the stalk are dead and exhibit
some features of apoptosis, such as nuclear condensation (Cornillon et al., 1994).
However, DNA is not degraded to nucleosome-sized fragments, and the dead stalk
cells are not phagocytized. Instead, they acquire a rigid cellulose cell wall reminiscent
of plant cell walls.
Caspase inhibitors have been shown to block certain stages of this differentiation
program, and active site labeling showed ‘caspase-like’ activity, a finding which,
however, could not be confirmed with the standard fluorogenic caspase substrates,
DEVD-AMV or YVAD-AMC (Cornillon et al., 1994). Moreover, cell death was not
inhibited by caspase inhibitors and therefore appeared to be caspase independent
(Olie et al., 1998).
Stalk-cell differentiation can be induced in vitro in suspensions of Dictyostelium
cells treated with DIF (differentiation-inducing factor). Such cells are more amenable
to experimental analysis, and Arnoult et al. (2001) have used them to demonstrate a
decrease in cell size, degradation of DNA to oligonucleosome-size fragments, and loss
of mitochondrial membrane potential in dying cells. They showed further that the
Dictyostelium homolog of AIF (apoptosis-inducing factor [Susin et al., 1999]) was
released from mitochondria during the differentiation process, and that it could
induce DNA fragmentation when added to nuclei in vitro. These experiments
strongly support the notion that a cell-death program accompanies stalk-cell
differentiation in Dictyostelium. This program has distinct characteristics that
distinguish it from caspase-dependent apoptosis in higher eukaryotes. It is caspase
independent and appears to be mediated by release of AIF from mitochondria.
AIF is a highly conserved mitochondrial protein with NADH oxidase activity
(Miramar et al., 2001). Despite its normal localization in mitochondria, AIF contains
a nuclear localization signal that appears to be responsible for its translocation to the
nucleus after release from mitochondria in apoptotic cells. AIF transport to the
nucleus leads directly to DNA degradation by a process different from that caused
by caspase-dependent apoptosis because it results in larger DNA fragments. AIF has
been shown to be essential for caspase-independent apoptosis in cleavage-stage mouse
embryos (Joza et al., 2001). Its presence in Dictyostelium cells suggests that it may
represent an ancient cell-death pathway that is evolutionarily conserved. In higher
animals, this pathway seems to coexist with the caspase-dependent apoptotic pathway.
In Dictyostelium, however, which apparently lacks a classical caspase gene, the AIF
pathway is used to organize the developmental cell-death program.
Caspase-mediated cell death and AIF-activated cell death have in common the fact
that they both rely on mitochondrial disintegration (see Chapter 7). Caspase
activation via the caspase 9/APAF1 pathway requires the release of cytochrome c from
mitochondria. AIF-activated cell death requires release of AIF from mitochondria.
In both cases, mitochondria are causally involved in apoptosis. This fact has led to
an interesting proposal by Ameisen (1996; 1998) for the evolution of cell death. He
EVOLUTION OF CELL DEATH 157