In yeast, activation of TOR in a nutrient-rich environment is responsible for directing a large part of cellular resources toward ribosome synthesis (38
). The data obtained in this study suggest that inactivation of TOR not only stops the synthesis of new ribosomes, as previous studies have demonstrated (26
), but also triggers extensive turnover of the existing ribosomes, resulting in a dramatic reduction of the cellular ribosome content. Our observations indicate that more than one half of all ribosomes present in a yeast culture grown in rich medium can be turned over within the first 5 to 6 h after exposure to rapamycin before the remaining ribosome pool becomes stabilized (). These data essentially imply that TOR-mediated control of the ribosome content is dual in nature: TOR signaling can affect both ribosome synthesis and the degradation of mature ribosomes. From a theoretical perspective, such dual regulation may provide a more flexible strategy than the regulation of synthesis alone for fine tuning of the ribosome content in a rapidly changing environment. Previous studies of yeast indeed suggested a contributing role for ribosome degradation during transitions to a lower growth rate, which occur when the nutrient quality of growth medium begins to decline (28
). Interestingly, ribosome degradation was also recently reported to take place in bacteria when culture conditions become limiting for exponential growth (43
), suggesting an adaptive role for turnover as part of the dynamic control of the ribosome content across a broad range of microorganisms.
Given than more than 105
ribosomes are present in each cell of S. cerevisiae
during growth in rich medium (62
), the extent of the decrease in rRNA upon rapamycin treatment (D) implies turnover of tens of thousands of ribosomes per hour. Thus, yeast cells have a highly efficient system in place to degrade their ribosomes. In previous studies, ribosomes and ribosomal proteins were observed in the yeast vacuole when autophagy was activated by starvation conditions (32
). Data presented above, however, indicate that active autophagy is not a prerequisite for the rapid ribosome degradation in rapamycin-treated cells (). In addition, the finding that turnover of rRNA decay intermediates is influenced by the exosome together with its cytoplasmic cofactors ( and ) argues that at least the initial steps of this process take place in the cytoplasm. These observations may not necessarily conflict with the findings in previous studies, as yeast cells may carry out ribosome degradation for different purposes by utilizing distinct mechanisms. For instance, targeting of ribosomes to the vacuole through autophagy may be important for recycling of the ribosomal material during long-term starvation or in preparation for sporulation as suggested previously (57
), whereas cytoplasmic degradation may be used primarily for adjusting the size of the translation machinery in response to rapidly occurring changes in the nutrient environment.
One important question raised by our findings is what mechanism is responsible for switching from the stable state of ribosomes during steady growth to their degradation upon TOR inactivation. There is very little observable lag in the onset of rRNA degradation after rapamycin treatment (D), suggesting that the machinery for the rapid ribosome degradation may already exist in growing cells even though during exponential growth ribosomes appear to undergo little turnover (E). Inhibition of translation initiation is a well-documented effect of rapamycin (6
) and is also characteristic of conditions of nutrient downshifts (5
). Inhibition of initiation would be expected to create a pool of ribosomal subunits that have finished previous rounds of translation but cannot start another translation job. Hence, one attractive hypothesis is that yeast cells may have a system that recognizes such idle subunits and targets them for degradation. Stabilization of rRNA in rapamycin-treated cells after treatment with cycloheximide (C) lends support to this idea, as cycloheximide prevents the runoff of translating ribosomes from mRNA, and this might protect them from degradation by keeping subunits sequestered away from the idle pool. Future studies will be needed to test this hypothesis.
The evidence of the cytoplasmic rRNA turnover in a eukaryotic organism suggests intriguing parallels in mechanisms of nutrient-dependent ribosome degradation between eukaryotes and prokaryotes, as the latter use a variety of nucleases in this process (18
). Our data show that the exosome, Ski7, and apparently the entire SKI complex of exosome cofactors function by promoting rRNA decay ( and ). Interestingly, the exosome core contains subunits homologous to the prokaryotic RNase PH (48
), which was recently shown to play an important role in rRNA degradation in Escherichia coli
). The Ski-exosome system in yeast appears to function largely at later stages of 25S rRNA degradation than RNase PH, suggesting that additional nucleases may work upstream to initiate this process in eukaryotes. Our analysis of other exonucleases involved in rRNA decay showed that the 5′ exonuclease Xrn1 can also play a role in the degradation of rRNA, although its activity appears to be highly dependent on the strain background (D. G. Pestov and N. Shcherbik, unpublished observations), suggesting that pathways to degrade rRNA fragments in the cytoplasm may be flexible and utilize multiple nucleases.
Defective cytoplasmic ribosomes containing mutations in rRNA were recently found to be selectively degraded through a process termed nonfunctional rRNA decay (17
). Our analysis of strains deficient in known genes involved in NRD (A and B) indicates that nutrient-dependent turnover is controlled separately, although we cannot exclude the possibility that the two processes may intersect at some point. Both NRD and nutrient-dependent turnover are likely to involve additional, early-acting nucleases that remain to be identified. Endonucleases represent one possible group of candidate factors that could initiate RNA decay; however, such nucleases have to date remained elusive. Another possibility is that proteins that do not possess nucleolytic activity themselves, such as helicases, could initiate ribosome turnover by stripping components that protect rRNA in mature ribosomes, thereby making rRNA susceptible to attack by nonspecifically acting nucleases. The mutants unable to efficiently process rRNA decay intermediates, such as ski
Δ strains deficient in 25S rRNA decay (A and A), should facilitate future searches for factors that act at early steps of eukaryotic ribosome degradation. Because ribosomes are ribonucleoprotein particles, we also anticipate that additional systems exist in cells to process ribosomal proteins removed from disassembled cytoplasmic ribosomes, and the factors involved in such activities also await identification and analysis.
Finally, it will be interesting to extend analysis of rRNA degradation to metazoan cells. Increased rRNA turnover was observed previously in contact-inhibited mammalian cells (64
) and cells undergoing differentiation (11
). In addition, recent studies of human cells revealed the presence of diverse rRNA fragments in the cytoplasm and the increase of such fragments after knockdowns of scavenging nucleases (56
). Thus, it is possible that a process analogous to the cytoplasmic ribosome turnover described herein contributes to the control of the ribosome content in mammalian cells.