Biology Reference
In-Depth Information
conÝguration that allows both translation and antisense RNA binding. In plasmid-bearing cells, the
truncated mRNA is rapidly bound by constitutively transcribed, unstable antisense RNA. The
resulting double-stranded RNA is instantly degraded by RNase III. In plasmid-free cells, the
unstable antisense RNA decays faster than truncation of the accumulated full-length toxin mRNA
occurs. In the absence of antisense RNA, the truncated mRNA will be translated into the toxic
peptide. The
chromosome contains at least Ýve such systems, all of which are no longer
functional. However, regulation and control through antisense RNA might reÞect an ancient and
widespread system (Wagner and Simons, 1994; Eddy, 2001; Wagner and Flrdh, 2002).
The proteins of these toxinÏantitoxin systems, most of which are only 33 to 130 amino acids
long, show such a degree of sequence variation that it is difÝcult to identify characteristic motifs.
This leaves only their small size as an indication of their presence. Experimentally, most addiction
systems are sensitive to antibiotic and heat treatment. It is assumed that a strong positive selection
is responsible for the rapid divergence of these systems. These systems presumably originated on
the bacterial chromosome. This is substantiated by alternative systems, such as the
E. coli
sop
system of
plasmid F and the
system of prophage P1, which rely on a centromere-like region to achieve
an equal and active distribution to all progeny cells more efÝcient than toxinÏantitoxin systems.
The moment a plasmid has acquired a toxinÏantitoxin system, it is under selective pressure to
become a unique toxinÏantitoxin system. The post-segregational killing is aimed at the exclusion
of competing cytoplasmic factors (Cooper and Heinemann, 2000). In a few cases, a chromosomal-
and a plasmid-based system have been shown to interact, and the plasmid-coded toxin Kid can be
neutralized by the chromosome-coded antitoxin ChpAI (Santos-Sierra et al., 1997, 1998). The
ancestral function of the proteic toxinÏantitoxin systems may lie in a stringent-relaxed response,
which occurs when bacteria face starvation. The toxin may function as a protein-synthesis inhibitor
during starvation. Continued starvation may lead to the killing of the cell, which would release
nutrients, which in turn would permit neighboring cells of the same strain to survive until conditions
improved. In the lysogenic state, the bacteriophage
par
gene, which prevents the
degradation of the host chromosomal antitoxin MazE (ChpAI) and the prophage antitoxin Phd by
speciÝcally inhibiting ClpP proteases that otherwise would inactivate the two antitoxins;
m
expresses the
rexB
acts
here as an anticell death gene. This shows that a bacteriophage can override a modiÝcationÏrescue
system and prevent CI. This, again, can be overcome either by a point mutation or the insertion of
an interposon in the
rexB
rexB
gene of the bacteriophage (Engelberg-Kulka et al., 1998; Engelberg-Kulka
and Glaser, 1999).
One can envision an evolutionary scenario in which a chromosomal host system responsible
for controlled or programmed cell death in cases of starvation evolves into a mobile, cytoplasmic,
and parasitic system. Transposons, phages, or insertion-sequence activities that move modiÝca-
tionÏrescue systems onto new plasmids may also move these systems back onto the chromosome.
Suddenly, a CI system has become a mutualistic system. This is in line with proposals that either
an already existing addiction system of the rickettsial ancestor of mitochondria or the development
of a new addiction system in the early mitochondria might have evolved into todayÔs nuclear-
encoded apoptosis systems (Kroemer, 1997; Ameisen, 2002).
In this context, it is also interesting that many of these bacterial toxinÏantitoxin systems not
only exhibit a very broad host range but are capable of functioning outside their kingdom as well.
The chromosomal RelE toxin of
E. coli
does kill yeast cells when expressed from a yeast plasmid
in
. When both toxin and antitoxin were placed on a different plasmid,
the antitoxin was to some extent able to rescue RelB toxicity in the yeast. This proves that
prokaryotic modiÝcationÏrescue systems can work in eukaryotes. Expression of
Saccharomyces cerevisiae
in a mamma-
lian cell line leads to the inhibition of cell proliferation (Kristoffersen et al., 2000). In a human
osteosarcoma cell line, RelE induces growth retardation and eventually cell death by apoptosis
(Yamamoto et al., 2002). RelE, Shiga toxin (STX), and Microcin E492, a channel-forming bacte-
riocin from
relE
, are representatives of the few but increasing number of proteins
that exert their effect on both prokaryotic and eukaryotic cells. Especially in invertebrates, where
Klebsiella pneumoniae
