Review for "Cell-specific modulation of nuclear pore complexes controls cell cycle entry during asymmetric division"

Completed on 7 Dec 2017 by Neil Adames. Sourced from

Login to endorse this review.

Comments to author

Kumar et al. describe how a yeast histone deacetylase (HDAC), Hos3, is involved in delaying the START transition specifically in daughter cells. This manuscript is very thorough and utilizes some elegant experiments. Based on its length and content, I suspect it is intended for Cell or Molecular Cell.

Based on previous work on Hos3 localization and its role in G1 phase timing, Kumar et al. hypothesized that Hos3 specifically inhibits START in daughter cells. The authors demonstrate that Hos3 becomes associated with the nucleoplasmic side of nuclear pores of daughter nuclei during anaphase. Interestingly, Hos3 association with daughter NPCs requires that Hos3 localize to the mother-bud neck and that the nucleus traverse the neck. Daughter NPC association of Hos3 also depends on the karyopherin Mtr10.

Kumar et al. propose that Hos3 maintains a high level of the START inhibitor Whi5 in daughter nuclei by altering Whi5 import dynamics and show that altering Hos3 localization and activity also alters Whi5 nuclear distributions and the length of G1. However, the authors importantly show genetically that Hos3 and Whi5 act synergistically, implying that Hos3 also affects another process important for START. Besides Whi5, Hos3 affects the localization of a number of proteins involved in nuclear transport of proteins and mRNA. Because the repressor of CLN3, Ace2, has been shown to asymmetrically accumulate in daughter nuclei, the authors checked if Cln3 abundance was affected by Hos3, but it was not. However, CLN2 expression is affected by Hos3, with expression occurring early in the absence of Hos3 activity.

Since HDACs like Hos3 are instrumental in gene silencing, and CLN2 is relatively close to the telomere, the authors examined if CLN2 is silenced by Hos3. As one would expect for a silencing mechanism, CLN2 is associated (along with its telomere) with the nuclear periphery during G1, but not S, and more so in daughter nuclei. Moreover, this peripheral localization depends on Hos3 activity. Artificial tethering of CLN2 to the periphery resulted in delayed START in both mother and daughter cells independent of Hos3 activity, suggesting that this artificial tethering bypasses Hos3 function.

Finally, Kumar et al. start to dig into the Hos3 silencing mechanism by showing that Hos3 deacetylates components of the NPC and their deacetylation is important for Whi5 asymmetry, CLN2 tethering to the nuclear periphery, and the G1 delay in daughter cells.


1) The manuscript lacks a lot of background information needed for a reader not versed in yeast cell cycle regulation or nuclear pore function.

The START transition in yeast is regulated by cyclin-dependent kinase (CDK) associated with the G1 cyclins Cln3, and Cln1/Cln2 (Cln1 and Cln2 are partially redundant). Cln3-CDK phosphorylates and inhibits Whi5, which binds to a major S-phase transcription factor complex called SBF. The balance between Cln3 and Whi5 is a major determinant of the critical cell size necessary to enter S-phase. Once Cln3 is able to overcome Whi5, SBF transcribes CLN2. Cln2-CDK is also able to phosphorylate and inhibit Whi5 in a positive feedback loop that results in rapid Whi5 phosphorylation and consequent export from the nucleus. Whi5 nuclear import is constitutive and dependent on the karyopherins Kap95 and Cse1, whereas Whi5 export is dependent on the karyopherin Msn5, which interacts only with phosphorylated forms of its substrates. Re-import of Whi5 occurs in anaphase when the mitotic exit network phosphatase, Cdc14, dephosphorylates Whi5.

Yeast nuclear pore complexes (NPCs) have been shown to regulate many disparate cellular processes beyond that of nucleocytoplasmic shuttling. Relevant to this manuscript, association of genes with the nuclear pore is an antecedent for transcriptional silencing and activation (or memory), depending on the proteins that bind to the gene’s promoter (SAGA complex for activation, YKU70/80 and Rap1 for silencing). In the case of transcriptional activation (or memory), association of the gene with the NPC is thought to facilitate handing off of the transcript to the mRNA export machinery (of which Mtr2 and Mex67 are a part) and maintenance in a poised open conformational state. In the case of transcriptional silencing, the NPC facilitates association of the silenced gene with the inner nuclear membrane protein Esc1 (yeast don’t have an equivalent to the nuclear lamina, which is important for silencing in vertebrates). In general, gene silencing eventually involves histone deacetylation by recruitment of HDACs and heterochromatin formation.

2) Presumably, Hos3 moves from the neck pool to the NPC pool by their close proximity, but the authors do not measure if there is a directly inverse correlation between the amount of neck-localized Hos3 and daughter NPC-localized Hos3 during transit of the nucleus through the neck. Alternatively, Hos3 could be moving from cytosolic pools to the NPCs.

3) The authors don’t explain how cytoplasmic dynein (Dyn1) is involved in movement of the anaphase spindle into the neck. Dynein associates with the bud cortex and pulls on cytoplasmic (astral) microtubules emanating from the daughter-oriented spindle pole body, pulling the spindle into the mother-bud neck.

4) The authors show that in cdc12-1 septin mutants, Hos3 is symmetrically localized to mother and daughter NPCs, while in hsl7∆ mutants Hos3 is still daughter-specific, but only accumulates to the spindle pole body, and in cdc15-1 mitotic exit mutants, Hos3 is localized to the neck and daughter spindle pole body. These observations beg for a better explanation than that provided – “We conclude that Hos3 recruitment to the bud neck is dispensable for its localization to the dSPB but is essential for its enrichment at the daughter cell nuclear periphery.” Clearly, close proximity of the NPCs to the neck isn’t important for Hos3-NPC localization because Hos3 is symmetric in the cdc12-1 mutants but the mother NPCs don’t traverse the neck (or do they?). The septin ring has been shown to be important for the asymmetry of numerous daughter or mother-specific molecules, so Hos3 may normally be prevented from diffusing into the mother by the septins. Moreover, it seems that Hos3 NPC localization to NPCs might depend on its phosphorylation state since loss of two different kinases maintain Hos3 asymmetry but prevent its daughter NPC association (SPB localization may be a weak one normally masked by NPC localization).

5) The authors don’t explain why they tethered Hos3 to the pore interior protein Nup49 rather than the nuclear basket protein Nup60.

6) The authors should have measured CLN2 expression in the experiments in which they artificially tethered CLN2 to the nuclear periphery (Fig. 6) and in which they used constitutive acetylation mimicry NPC mutants (Fig. 7).