Messenger RNAs transcribed from yeast linear cytoplasmic plasmids possess unconventional 5’ and 3’ UTRs and suggest a novel mechanism of translation

Václav Vopálenský; Michal Sýkora; Tomáš Mašek; Martin Pospíšek

We prepared the M1 cDNA, cloned a gene coding for the prepro-K1 killer toxin into a yeast shuttle expression plasmid (106) under the control of the strong constitutive TPI promoter (plasmid pYX212::M1) and introduced this expression plasmid into the S. cerevisiae CWO4 ρ 0 strains carrying all three CDC33 alleles described above. The resulting strains display K1 killer and immunity phenotypes at the permissive temperature (24°C) (Fig. 6B) . The experiment depicted in Fig. 6 shows a comparison of the production of the K. lactis toxin expressed naturally from the pGKL1 plasmid and the K1 toxin expressed artificially from the nuclear plasmid pYX212::M1. The K1 toxin gene was transcribed by RNA polymerase II in the nucleus and thus provided a 5' m 7 G-capped and 3' polyadenylated mRNA control. We cultivated all the CWO4 ρ 0 [pGKL1/2] and CWO4 ρ 0 [pYX212::M1] strains in parallel to the exponential phase at the permissive temperature (24°C). All the strains were then harvested and briefly washed three times with an excess of the fresh medium prewarmed to 24°C. Half of each culture was further cultivated at 24°C. The second half was rapidly transferred to fresh medium prewarmed to 37°C and further cultivated at this elevated temperature. At this point, initial samples marked zero were taken for biomass measurements and the determination of toxin activity in the culture medium. Toxin activities were undetected in all samples at time zero and provided clear evidence that all cultures were washed sufficiently and that the toxin activities present in samples taken after 3, 6 and 12 hours of cultivation corresponded to proteins produced only in the course of the experiment (Fig. 6) . The production of the K. lactis killer toxin in S. cerevisiae strains bearing pGKL plasmids was almost comparable under permissive and restrictive conditions, regardless of whether the wild-type or temperature-sensitive CDC33 allele was present ( Fig. 6A; 37°C) . Production of . CC-BY-NC-ND 4.0 International license author/funder. It is made available under a The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/325316 doi: bioRxiv preprint the K1 toxin in the cdc33-1 and cdc33-42 yeast strains containing the pYX212::M1 plasmid vanished rapidly at 37°C, whereas the wild-type Cdc33 cap-binding protein permitted toxin production at both 24 and 37°C (Fig. 6B) . The presence of the temperature-sensitive eIF4E proteins in cdc33-1 and cdc33-42 yeast strains was documented by their restricted growth at 37°C. Both strains containing the wild-type allele of the CDC33 gene grew well at both 24 and 37°C (Fig. 6) . The growth of CWO4 CDC33, ρ 0 [pYX212::M1] was slightly slower at 24°C than at 37°C. The most likely reason is that K1 killer toxin is more active at 24°C than at 37°C, even though we cloned the coding sequence of the K1 super-killer variant that shows enhanced thermal stability (107). We previously showed that K1 preprotoxin produced artificially from the nuclear plasmid does not fully support the immunity phenotype of the host cells (106). Higher K1 toxin stability at 24°C thus probably led to the slightly slower growth. However, both cultures eventually reached comparable biomass, in contrast to the strains bearing conditional cdc33 ts alleles and cultured at a restrictive temperature (37°C), which completed the initiated cell division and ceased growth (Fig. 6B ). The activity of the TPI promoter used to produce K1 tox