Comment on “Principles of ER cotranslational translocation revealed by proximity-specific ribosome profiling”

Abstract

Jan et al. propose that ribosomes translating secretome mRNAs are recruited to the ER upon emergence of the signal peptide and return to the cytosol following termination. A thorough accounting of their controls demonstrates, however, that ribosomes are retained on the ER through many full cycles of translation.

Jan et al. (1) report that ribosomes are recruited from the cytosol to the ER membrane upon the emergence of an ER signal peptide and that these ribosomes are released back into the cytosol soon after completion of protein synthesis. While these conclusions are consistent with the canonical signal recognition particle (SRP) pathway model (2), recent work has demonstrated that ribosomes associate stably with the ER and as bound ribosomes translate both signal sequence- and cytosolic protein-encoding mRNAs (3–6), findings which cannot be accounted for by the SRP model. In an attempt to reconcile these fundamental discrepancies, we re-analyzed the authors’ data, including control data that were not integrated into their conclusions. In so doing, we found that the data from Jan et al. support a model where ribosomes stably associate with the ER.

To examine ribosome trafficking and translation dynamics, Jan et al. developed proximity-specific ribosome profiling, where chimeras of ER membrane proteins and the biotin ligase BirA were used to label ribosomes bearing BirA acceptor sequence-tagged ribosomal protein RPL16 (7). In one iteration, BirA-Ssh1, where Ssh1 is paralogous to the protein-conducting channel Sec61, was used to label translocon-proximal ribosomes. Jan et al. then determined the location of tagged ribosomes on mRNAs by ribosome profiling. The authors observed that very few ribosomes were labeled by BirA-Ssh1 prior to emergence of a signal sequence, leading them to conclude that ribosomes translating secretome mRNAs are co-translationally targeted to the ER after the signal sequence is translated, in agreement with the SRP model.

We were able to reproduce the authors’ primary observation using their data, where BirA-Ssh1 labelled ribosomes were depleted on mRNAs encoding predicted signal sequences (8) until ~30 codons after the signal sequence (Fig. 1A). However, when we analyzed their data from control experiments with a BirA-Ubc6 tail anchor chimera (Ubc6_TA), designed to be a general reporter for ER-associated translation, this translational pattern was not observed (Fig. 1B). Instead, the BirA-Ubc6_TA labeled ribosomes identically, and prior to emergence of a signal sequence. The presence of translating ribosomes on the ER prior to emergence of the signal sequence indicates that ribosomes can be ER-associated during translation initiation and thus that ribosome recruitment to the ER is not obligatorily dependent on SRP-mediated recruitment. The paucity of ribosomes in the first ~30 codons observed with the BirA-Ssh1 reporter then likely reflects recruitment of ER-bound ribosomes that had undergone initiation prior to their accessing BirA-Ssh1 translocons, rather than the trafficking of cytosolic ribosomes to BirA-Ssh1 translocons. Critically, because ER-bound ribosomes are well represented on secretome mRNAs prior to emergence of the signal sequence, the data from Jan et al. demonstrate that co-translational targeting to the ER is unlikely to be the primary mechanism by which mRNAs or ribosomes are ER-localized.

Figure 1. ER-bound ribosomes prior to signal sequence emergence.

The enrichment of ribosomes in the biotin-labeled pulldown relative to input was calculated at each codon for mRNAs that encode signal sequences. Plots were generated for (A) BirA-Ssh1 and (B) BirA-Ubc6_TA. Shaded area represents SD across all analyzed genes. Each sample was analyzed after a 7 min biotin pulse without cycloheximide.

If ribosomes are not recruited to the ER in a manner coupled to emergence of the signal sequence, how might the authors’ conclusions regarding ribosome exchange rates between the cytosol and ER be viewed? In Jan et al., in vivo ribosome exchange rates were determined by tracking the mRNA composition of BirA-Sec63-labelled ribosomes as a function of time. At very short labeling periods, BirA-Sec63 labeled ribosomes were enriched in secretome mRNAs. This enrichment dissipated at longer (7 min) labeling periods, at which point BirA-Sec63 labeled ribosomes were about equally associated with cytosolic and secretome mRNAs. This change was interpreted by the authors as reflecting ribosome release to the cytosol and re-initiation on cytosolic protein-encoding mRNAs, and thus that “ribosomes at the yeast ER are highly dynamic, freely exchanging into the cytosol within at most a few rounds of translation.” For this approach to be valid, one must make the assumption that ER-bound ribosomes translate only secretome mRNAs. However, when analyzing the composition of labeled ribosomes during this time course, we found that ribosomes translating cytosolic proteins comprise the vast majority of labelled ribosomes even after 1 minute of labelling (Fig. 2), invalidating the assumption required to infer ribosome half life from these data. This abundance of mRNAs encoding cytosolic proteins being translated on the ER has been widely reported (3, 9), but was not discussed in this manuscript. We suggest that the changes in mRNA composition occurring as a function of labeling time reflect a modest kinetic labeling advantage for ribosomes in close proximity to translocon-associated BirA-Sec63, rather then ribosome exchange. Although the precise half life of a ribosome on the ER therefore remains unknown, the bulk of evidence that we have discussed above and elsewhere (10) indicates that it is substantially longer than a single cycle of translation.

Figure 2. Stable ribosome binding to the ER.

(A) The fraction of BirA-Sec63 labeled ribosomes engaged in the translation of cytosolic or ER proteins was calculated for various labeling times. (B) Stacked bar plot indicating the proportion of ER-associated mRNAs that encode cytosolic proteins from various studies.

After additional analysis, the data from Jan et al. best support a model where ribosomes bind stably to the ER rather than one where ribosome localization is coupled to protein targeting. In the system that Jan et al. propose, where ribosome exchange between the cytosol and ER is rapid, one would expect prominent co-translational ribosome targeting to the ER and robust release of ribosomes following completion of protein synthesis. However, neither of these predictions is supported by the provided data. Rather, in challenging the role of SRP and co-translational protein targeting in the subcellular organization of translation, these data suggest provocative new questions regarding the in vivo mechanisms of ribosome exchange and mRNA traffic on the ER.