Except for pGKL1, there are three nonautonomous linear plasmids from different yeast species known to encode killer toxins, i.e., pPac1-2 of P. acaciae, pPin1-3 of P. inositovora, and pWR1A of D. robertsiae (Gunge et al. 1981; Wor-sham and Bolen 1990; Hayman and Bolen 1991; Klassen and Meinhardt 2002). Each of the aforementioned elements encodes a chitin binding protein, structurally akin to the zymocin a-subunit (Klassen and Meinhardt 2002, 2003; Klassen et al. 2004). Additionally, hydrophobic C-terminal regions very similar to the zymocin ß-subunit were identified in such deduced chitin binding polypeptides, suggesting chitin binding and subsequent import of a cell cycle arresting or lethal subunit as for zymocin. Consistently, mutations affecting the target cells major chitin synthase (Chs3) conferred toxin resistance to all of the toxins (Klassen and Meinhardt 2003; Klassen et al. 2004). Moreover, reminiscent of the y-subunit, separate structural genes encoding potentially secreted proteins were identified in either case (Fig. 3).
For the P. inositovora plasmid pPin1-3, such protein does indeed compare to zymocin y and—not astonishingly—as for zymocin, toxicity is impaired in an Elongator (elp3) mutant (Klassen and Meinhardt 2003). The functionally related P. inositovora toxin together with K. lactis zymocin were assigned to a corporate class of linear plasmid encoded toxins (type I), which are characterized by their dependency on cell wall chitin and, more importantly, on a functional Elongator complex (Schaffrath and Meinhardt 2004; Jeske et al. 2006b). Distinct from the K. lactis system, however, immunity is not encoded by the linear plasmids in P. inositovora (Hayman and Bolen 1991). Though there is a gene similar to the K. lactis pGKL1 ORF3 (see above) in the toxin encoding pPin1-3 (Klassen and Meinhardt 2003), its function has not been studied.
The killer plasmids pPac1-2 and pWR1A from P. acaciae and D. robertsiae, respectively, are structurally closely related, but their resemblance to type I toxin encoding elements (pGKL1 and pPin1-3) stretches only across the ORF encoding the chitin binding polypeptide, whereas the separately encoded and potentially secreted proteins (pPac1-2 Orf2p and pWR1A Orf3p) hardly bear any likeness to zymocin y and/or the P. inositovora equivalent (Klassen et al. 2004). However, as for zymocin y, intracellular expression of either protein phenocopies the impact of the extracellularly applied toxin on target cells; thus, their function as intracellularly acting toxin subunits immediately comes to mind (Klassen et al. 2004).
In accordance with functions different from zymocin y, toxins of P. acaciae (PaT) and D. robertsiae do not require Elongator (Klassen et al. 2004). Hence, tRNAs with an Elongator-dependent mcm5-modified wobble uridine are unlikely to be their targets (see above). Interestingly, a mutation in the TRM9 gene (encoding a tRNA methyltransferase known to be involved in completion of the mcm5 modification) (Kalhor and Clarke 2003; Lu et al.
2005; Jablonowski et al. 2006) confers resistance to the type II toxin of P. aca-ciae to a certain degree (McCracken et al. 1994; our unpublished results). Taking also into consideration that other than for zymocin, overexpression of tRNA§UC has no protective consequences in type II toxicity (our unpublished results), Trm9 may have an additional cellular role besides mcm5 modification of tRNA.
Type II toxins arrest target cells in the S phase of the cell cycle and activate the intra-S-phase DNA damage checkpoint, a scenario that agrees with a replication inhibitory and/or DNA damaging function (Klassen et al. 2004). The P. acaciae toxin (PaT) induces death in a two-step fashion; during the first 3-4 h in toxin, target cells lose their viability to approximately 30% that of control levels. Such a phase is characterized by hyperphosphorylation of the DNA-damage checkpoint kinase Rad53 and mutation induction, indicating DNA damage to be involved in lethality at this stage (Klassen et al. 2004; Klassen and Meinhardt 2005). Our recent work has revealed that DNA damage induced by PaT occurs during the S phase of the cell cycle and probably involves formation of broken replication forks. Congruently, mutants impaired in stalled replication fork stabilization and recovery, as well as in double-strand break (DSB) repair, react extremely sensitively to PaT (our unpublished results).
A period comprising several hours of constant viability in toxin (~30%) follows until the final decline to less than 1% occurs. Final cell death is characterized by the appearance of typical apoptotic markers, including abnormal nuclear morphology, reactive oxygen species, DNA fragmentation, and phosphatidylserine flipping (Klassen and Meinhardt 2005). Since Rheovirus encoded killer toxins with clearly distinct modes of action (such as K1 and K28 from S. cerevisiae and zygocin from Zygosaccharomyces bailii) also activate programmed cell death in target cells (Reiter et al. 2005), apoptosis is apparently an aftereffect and not confined to individual toxins.
Interestingly, type I and type II toxins are both influenced by the mating-type status of the target cell. Diploid cells are not only significantly more resistant to both of the toxins compared to either of the haploid parents, but also overexpression of the MATa locus in a MATa strain, as well as mutations in genes required for silencing of the cryptic mating-type loci (SIR genes), render target cells resistant to toxins of both types (Butler et al. 1994; Klassen et al. 2006). Such an effect is likely due to the repression of haploid-specific genes required for the action of either toxin and, thus, may be due to the cellular uptake mechanism, which is akin in either case (see above; Table 3). Moreover, both zymocin and PaT interfere with the mating competence of target cells, as these are refractory (Klassen et al. 2006). Thus, both killers act efficiently on haploid cells only and prevent them from mating, thereby ensuring most effective target cell killing under unfavorable environmental conditions, and causing sporulation of target cells which compete for limited resources (Klassen et al. 2006).
Table 3 Genes essential for killer toxin action, arranged according to their function in target cells. Toxin resistance and susceptibility of individual mutants to killer toxins from K. lactis (Kl), P. acaciae (Pa), P. inositovora (Pi), and D. robertsiae (Wr) are indicated. Pa, Pi, and Wr resistance/sensitivity is not proven for each gene in a given category
Target cell Toxin relevant Toxin Toxin Refs.
modification process resistance suscepti-
(relevant genes) bility
Chitin synthesis defect CHS3, CHS4, CHS5, CHS6, CHS7
Plasma membrane sphingolipid synthesis defect
IPT1, LAG1, LAC1
Cell wall Kl/Pa/Pi/Wr receptor; uptake
Uptake; membrane receptor (?)
Takita and Castilho-Valavivius 1993; Kawamoto et al. 1993; Butler et al. 1991a; Jablonowski et al. 2001a; Klassen et al. 2004; McCracken et al. 1994
Zink et al. 2005; McCracken et al. 1994;
Plasma membrane Intracellular
H+ATPase defect activation
Simultaneous expression Uptake (?) of MATa and MATa SIR1-4; HMR, HML
High copy glutaredoxin Uptake (?) GRX3
Elongator subunit defect tRNA ELP1/IKI3, ELP2, ELP3, modification, ELP4, ELP5/IKI1, ELP6 transcription, exocytosis
Elongator relevant Elongator factor defect modification/
KTI11, KTI12, interaction
ATS1/KTI13, HRR25/KTI14, URM1, UBA4, SIT4, SAP185, SAP190; overexpression of SAP155; KTI12
Kl/Pa Mehlgarten and
Schaffrath 2004; McCracken et al. 1994; unpublished results
Kl/Pa Butler et al. 1994;
Klassen et al. 2006
Kl Jablonowski et al.
Kl/Pi Pa/Wr Yajima et al. 1997;
Frohloff et al. 2001; Klassen et al. 2004; McCracken et al. 1994
2001; Fichtner and Schaffrath 2002; Mehlgarten and Schaffrath 2003; Fichtner et al. 2003; Jablonowski et al. 2001b,c; Klassen et al. 2004; McCracken et al. 1994
Table 3 (continued)
Target cell modification (relevant genes)
Toxin relevant Toxin Toxin Refs.
process resistance suscepti bility tRNA overexpresion Target tPMAGlu ; tRNAuUu + overexpression
Butler et al. 1994; Lu et al. 2005; Jablonowski et al. 2006;
unpublished results Lu et al. 2005; Jablonowski et al.
2006; McCracken et al. 1994;
unpublished results Fichtner et al. 2003
Diphthamide synthesis ?
As for type I, immunity to type II toxins is procured by genes located on the nonautonomous, toxin encoding elements pPac1-2 and pWR1A (Wor-sham and Bolen 1991; our unpublished results). Predicted immunity factors encoded by pPac1-2 ORF4 and pWR1A ORF5 (our unpublished results) display striking similarities to each other but not to type I immunity encoded by pGKL1 ORF3 (see above). As for the latter, however, PaT immunity acts by in-tracellular interference with Orf2p function, rather than preventing its uptake (unpublished results).
In accordance with fundamental differences concerning both toxicity and immunity of type I and II toxins, pGKL1 Orf3p protects exclusively from the action of type I toxin, whereas at least partial cross-immunity was detected between P. acaciae and D. robertsiae toxin/immunity systems (our unpublished results).
Localization of an immunity conferring gene on the toxin encoding plas-mid provides an autoselection mechanism for the entire system, as cells of a natural population having occasionally lost such an element are counters-elected by toxin-producing neighbor cells. All known linear plasmid encoded killer toxins (except that of P. inositovora, see above) eliminate plasmid-free strains of the same species, suggesting that such toxins may indeed have autoselection purposes rather than providing an advantage to the host. Generally speaking, in killer/immunity combinations encoded by yeast linear plasmids there are (at least two) functionally distinct cargo proteins (types I and II), which exist conjointly with a respective compatible immunity factor. The cargo proteins, representing the effective principle of the toxin in either case, rely on a conserved carrier (the chitin binding/hydrophobic protein) which facilitates transport into target cells.
Was this article helpful?