Escrever um comentário

Review of House et al.: The reticulon homology domain of Pex30 generates membrane curvature at ER subdomains for lipid droplet biogenesis

In this manuscript, House et al. investigate the role of the Pex30 reticulon homology domain (RHD) for Pex30 localisation and function in Saccharomyces cerevisiae. Using light and electron microscopy approaches, they show that Pex30 mutants with elongated transmembrane domains are no longer targeted to the cortical Endoplasmic Reticulum (ER) and lack the ability to tubulate the ER to the same extent as the wildtype protein upon overexpression. Furthermore, the authors demonstrate that the Pex30 mutants cannot rescue lipid droplet morphology nor delayed lipid droplet biogenesis in pex30∆ backgrounds. In addition, molecular dynamics simulations support a role for the Pex30 RHD in stabilising local membrane curvature. Overall, the study identifies Pex30 localisation to curved ER regions as an important factor for lipid droplet biogenesis.

While the core finding of the paper, that the Pex30 RHD is required for proper Pex30 localisation to exert its function as a curvature inducing protein especially in lipid droplet biogenesis, is convincing, the authors did not to address the direct link between membrane protein localisation and function when mutating the transmembrane domain (see detailed comments below). Additionally, the clarity of the manuscript could be improved in order to help the reader follow the story line. We particularly noticed a lot of repetitions in the text that could be removed for better readability.

Major points:

· The authors convincingly show that mutating the transmembrane domain of the RHD impacts Pex30 localisation. However, it is unclear if these mutations directly affect the function of the protein or whether they lack the ability to rescue ER or lipid droplet morphology due to mislocalisation. The authors should at least address the duality of RHD localisation and function in the discussion.

· Similarly to the work of Zurek et al., 2010 on the RHD of Reticulon 4 in mammalian cells, which the authors also quote, House et al. should confirm that the Pex30 RHD mutants retain the correct topology in the ER bilayer and can therefore still be compared to the wildtype protein.

· The authors should provide a reference for the lipid composition they used in the MD simulations. Could the authors specifically comment on the absence of PI from their model membranes, as it is the most abundant anionic lipid in the yeast ER?

Minor points:

· Figure 1:

o “This extension enables the TMD to span both leaflets of the ER bilayer”

It would help the reader if the authors could first introduce the mechanism by which RHD containing proteins are suggested to induce membrane curvature.

o “GFP tagged Pex30 and Pex30 TMD“

Please specify whether these fusion proteins were endogenously tagged or expressed from a plasmid.

o 1.A: Could the authors discuss the impact of the charge of added residues?

o 1.B: Could the authors describe in the legend where the dysferlin domain and the N and C terminus are?

o 1.C: The authors should provide quantification of wildtype and mutant Pex30 abundance by immunoblot to confirm that the inability to restore ER morphology is not due to instability of the mutant protein.

o Legend of 1.C and 1.D: Could the authors comment on the relevance of imaging in logarithmic phase for figure 1.D and not for 1C?

o 1.E: Did the authors try to use an automated image analysis tool for quantification? Additionally, as the wildtype phenotype is not always homogeneous, could they provide supplementary information on which phenotypes are considered wildtype-like cortical ER morphology during manual quantification?

o “In S. cerevisiae, reticulons and reticulon-like proteins such as Rtn1, Rtn2, and Yop1 mainly regulate the ER shape. Deletion of all three proteins leads to more ER sheets than ER tubules suggesting they contribute to tubulating the ER membrane. Cells devoid of Pex30 do not exhibit change in cellular ER morphology as it is a less abundant protein and localized to ER subdomains.” Could the authors provide numbers for the relative abundance of the above proteins? Which exact subdomains do the authors refer to?

· Figure 2:

o 2.A: “WT cells show reticulated cortical ER”

The term “reticulated” usually refers to a net-like structure which cannot be seen here. Maybe the authors could use a more fitting description such as “short interspersed stretches of cortical ER”.

· Figure 3:

o “Thus, the MD simulations show that modifying the first short hairpin results in a more pronounced reduction in local membrane curvature and thickness than modifying the second short hairpin, which has only a modest effect.” In the previous panel, the mutants seem to be equally ineffective in rescuing tubular ER morphology. Could the authors comment on why this in-silico observation is not reflected for the TMD2 mutant in the in vivo experiments?

o Could the authors comment on the confidence of the ColabFold predictions for the WT and TMD mutants of Pex30?

o 3.A and 3.B visually show a difference in the curvature of the membrane around the RHD. Together with the quantification in 3.C the authors could comment on the biological relevance of this change in membrane curvature. It would also be helpful if the authors could provide a frame of reference for the mean curvature of the membrane surface around a transmembrane protein known to be inserted in a “flat” membrane region

· Figure 4:

o “When we overexpressed Pex30-GFP in sei1pex30Δ cells, it rescued the growth defect and accumulated at ER-LD contact sites as in sei1Δ” Could the authors introduce the function of Sei1? Also, could the authors introduce the sei1Δ and the sei1pex30Δ phenotype in Fig 4.A?

o “When Pex30 (TMD1)-GFP and Pex30 (TMD1.TMD2)-GFP mutants are overexpressed in sei1pex30Δ, cells still retain the LD phenotype as in sei1pex30Δ cells and these Pex30 TMD mutants do not accumulate into large puncta at ER-LD contact sites as WT Pex30-GFP (Figure 4A-4D)” In the introduction, the authors state that this accumulation is happening via the DysF domain. Could the authors comment on why the TMD mutants cannot rescue the phenotype despite having the DysF domain?

o Legend of 4.E: The original experiment from Joshi et al. 2018 were performed at 23°C, could the authors comment on why they used 37°C here?

· Figure 5:

o “we overexpressed Pex30-GFP, Pex30 (TMD1)-GFP, Pex30 (TMD2)-GFP and Pex30 (TMD1.TMD2)-GFP in are1are2dga1pex30Δ with LRO1 expressed under GAL1 promoter and labelled with Erg6-mCherry.” Could the authors introduce the functions of the deleted genes and explain the lipid droplet biogenesis assay in more detail?

o “Moreover, unlike the cells devoid of Pex30, cells expressing Pex30 TMD mutants exhibit a dominant negative effect on LD biogenesis as they show decreased LD formation at 5-hour timepoint (Figure 5A)” The term “dominant negative” describes the negative interference of a mutant when it is co-expressed with a wildtype protein. The authors should find a more fitting term here.

o 5.A: Could the authors provide error bars as well as representative microscopy images?

· Discussion:

o “ As the lens grows within the ER bilayer, the local membrane curvature generated by Pex30 lowers the energy barrier needed for LD nucleation while also reducing the surface tension in the ER for LD growth (Figure 5B).” The proposed mechanism contradict the experimental finding that many small LDs accumulate in pex30Δ, which shows that they are able to nucleate but not to grow.

Competing interests

The authors declare that they have no competing interests.

Use of Artificial Intelligence (AI)

The authors declare that they did not use generative AI to come up with new ideas for their review.