single apoptosis inducer, already allow us to grasp the complexity of the signaling
events in apoptosis.
2.3
Screening strategyAnother approach to clone apoptosis-relevant genes in cell culture was established in
the laboratory of Philip Leder (Grimm and Leder, 1997). As many gene activities are
redundant, knockouts (loss-of-function) cannot always be expected to yield a cellular
effect. In addition, as the selection strategy uses cell death as the final readout, some
genes that are part of a larger network of apoptosis regulators, and therefore are less
effective mediators, might not get detected. Moreover, the investigators were
motivated by a conspicuous correlation: almost every gene that transmits an apoptosis
signal also has the dominant capacity to induce cell death upon overexpression. This
activity is even conserved across species boundaries. It might be explained by the
observation that apoptosis is mediated by specific protein-protein interactions that
are induced when one of the partners is over expressed. At first, it seemed impractical
to use this dominant effect for isolating apoptosis genes because their activity leads
to the degenerative process of cell death, instead of cell survival, and therefore
amplification of the responsible genes, as in the case of the selection procedure.
However, a genetic screen to isolate such dominant, apoptosis-inducing genes could
be developed (Figure 1c). It is based on iterative transfections of very small expression
plasmid pools into human embryonic kidney cells (HEK). In order to decrease the
number of transfections, the cDNAs are taken from an unamplified, normalized gene
library. The subsequent microscopic inspection of the apoptotic phenotype facilitates
detection of cell death.
Since cell death can be induced by many different stimuli, the specificity of the
signal that can lead to apoptosis in HEK cells has been characterized. Several control
experiments with dominant-negative mutants of known genes, constitutively active
oncogenes, and various cytostatic drugs could not generate apoptotic cells, indicating
that induction of cell death needs a defined signal; an unspecific insult of the cells
does not suffice (Albayrak et al., submitted). This is also supported by a theoretical
consideration: given the severe consequences, committing suicide cannot be a minor
decision for a cell. Therefore, defined checks have to be overcome in a cell to induce
apoptosis. The specificity of the proapoptotic signal is also underscored by the finding
that adenine nucleotide translocase (ANT)-1, a component of the ‘permeability
transition’ (PT-) pore, can potently induce apoptosis. In contrast, ANT-2, a gene
that is 90% identical to ANT-1, is completely inactive in cell death induction (Bauer
et al., 1999).
After carrying out the screen for about 40% of a gene library, more than 70 positive
genes were found (about 0.07% of all cDNAs). This high count correlates with the
expected, and already experimentally shown, complex regulation of apoptosis
induction, and it reflects the number of sensors that exist in the cell for this program.
Among the isolated genes are several genes already known as apoptosis inducers,
CELL-CULTURE SYSTEMS IN APOPTOSIS 207