of the various genes from this screen along a signaling pathway is still lacking, it is
tempting to speculate that DAP-3, which seems to be localized on the membrane, is
the gene most proximal to the signal created by interferon-γ. From there, the signal
spreads to different organelles, including lysosomes, and activates caspases.
Eventually, we will learn how all the genes from this screen contribute to apoptosis
induction and how they can be integrated into a signaling pathway leading to cell
death.
In an independent experiment, Gabig and colleagues determined an antisense
construct to an endogenous gene by a similar expression-cloning assay, using IL-3
deprivation as apoptosis inducer in myloid cells. The corresponding gene, called
Requiem (Gabig et al., 1994), encodes a zinc-finger protein. When downregulated,
it protects cells against IL-3 deprivation. From its primary sequence, Requiem is
supposed to encode a transcription factor, but little else is known about this gene.
Using a retroviral insertion mutagenesis, Sarit Larisch and colleagues discovered
ARTS, a septin-like protein whose downregulation protects cells against TGF-β
apoptosis (Larisch et al., 2000). Until then, septins were only known to be part of
the cytoskeleton and, possibly, to aid in the assembly of signaling complexes involving
protein kinases. ARTS, in contrast, is an alternatively spliced isoform that lacks the
septin-like C-terminus. Interestingly, this protein is localized to mitochondria and
translocates to the nucleus upon apoptosis induction. While its exact function in
apoptosis in this organelle remains to be determined, it is an additional protein that
is released from mitochondria after apoptosis induction (see Chapter 7).
In a variation of the original antisense approach, Luciano D’Adamio and colleagues
(Vito et al., 1996) transfected a random primed cDNA library into mouse T-cell
hybridoma cells, and selected those that survived T-cell receptor cross-linking. This
approach is based on preceding experiments that led to the isolation of growth-
suppressor genes (Gudkov et al., 1994). Since the library in this experiment was
generated in a random way, it could be expected to contain inhibitors of apoptosis
that might function not only via antisense but also by truncated proteins that could
fortuitously constitute dominant-negative variants. Of the six isolated genes, three
have been characterized to date. ALG-3 and ALG-4 do encode proteins and act in a
dominant-negative manner by interfering with their endogenous proteins. ALG-2,
however, was expressed in an antisense orientation and downregulates the endogenous
protein. It encodes a Ca2+-binding protein that is also involved in glucocorticoid-
mediated apoptosis, as its antisense can repress cell death by this hormone. ALG-3
encodes a truncated version of presenilin-2. Mutations in this gene are responsible
for some familial cases of Alzheimer’s disease, which is characterized by an aberrant
cell death in neurons (Mattson, 2000). It is a conserved membrane protein that
contributes to the γ-secretase activity necessary to liberate β-amyloid from its
precursor protein. The localization of its protein to the endoplasmic reticulum (ER)
—which is a Ca2+ store for apoptosis induction (see Chapter 6)—and the Ca2+-
binding properties of ALG-2 suggest that these two genes regulate the concentration
of this proapoptotic ion. Both genes also mediated the proapoptotic effects of the Fas
receptor.
CELL-CULTURE SYSTEMS IN APOPTOSIS 205