Pii: S0962-8924(99)01711-0

PII: S0962-8924(99)01711-0 [PSI] was originally described in 1965 by Cox as a translation infidelity factor (Fig. 1)1. In [PSI] strains, a weak suppressor tRNA produced detectable nonsense suppression (stop-codon readthrough; see Box 1), whereas in [psi] strains the same tRNA suppressor appeared to be inactive1,2. Later experiments revealed that [PSI] was not required for suppressor tRNA function; rather, it increased the efficiency of suppression to a readily detectable level3,4. This enhancement of nonsense suppression (allosuppression) by [PSI] was not restricted to a single suppressor tRNA; suppression by other tRNAs as well as by mutations in ribosomal proteins was affected similarly. Thus, [PSI] functioned as a general or omnipotent allosuppressor2,5. Moreover, [PSI] can direct nonsense suppression on its own in strains lacking known genetic suppressors2,5. The phenomenon of allosuppression was well characterized in prokaryotes by the time [PSI] was described. However, [PSI] is distinct from these factors in that its pattern of inheritance is unusual. [PSI] is dominant in genetic crosses: if [PSI] and [psi] haploid strains are mated, the resulting diploid has the [PSI] phenotype (nonsense suppression)1. By classical genetics predictions, such diploids are presumed to be heterozygous for [PSI]. Surprisingly, however, the nonsense suppressor phenotype segregates to all four meiotic progeny1; that is, [PSI] is transmitted as a dominant, non-mendelian trait (hence the capital letters and brackets in its name). This pattern of inheritance was later explained by definitive localization of the [PSI] factor to the cytoplasm: [PSI] could be transmitted by cytoduction – that is, abortive matings in which cytoplasmic mixing occurs in the absence of nuclear fusion6. Through a series of experiments, [PSI] was distinguished from the known cytoplasmic nucleic acids, including the mitochondrial genome, killer virus and the sporulation-associated 20S RNA, as well as the extrachromosomal 2m and 3m DNAs2. To date, transmission of [PSI] has not been linked to the propagation of a novel or altered nucleic acid. In addition to its unusual mode of inheritance, the [PSI] factor is distinguished from conventional genetic elements by its metastability. The [PSI] and [psi] states are not absolute: [PSI] strains convert to [psi] at low but measurable frequencies, and new [PSI] elements appear spontaneously in [psi] strains at a similar rate1,7. Perplexingly, the frequency of [PSI] →[psi] conversion (i.e. [PSI] curing) is increased dramatically by treatment with low concentrations of agents that are non-mutagenic to nucleic acids, such as high salt60, guanidine hydrochloride and methanol8. Once lost, [PSI] can reappear spontaneously in cured strains7,9. Another peculiar aspect of [PSI] is that it can exist in a cryptic state. That is, in specific crosses, the phenotype will reproducibly disappear and then reappear in a predictable way in meiotic progeny1,10,11. The spontaneous loss and reappearance of [PSI] in yeast strains (reversible curing) as well as the ability of [PSI] to exist in a cryptic state are difficult to reconcile with the idea of a nucleic acid determinant.