Cells can adopt divergent fates upon division by unequally distributing molecules or structures that direct distinct gene expression programs. The genesis of this asymmetry rests on the cell's underlying architecture, and can involve segregation of mRNAs, transcription factors, and cell surface receptors [1
]. Unquestionably critical for metazoan development, asymmetric gene expression is also important in unicellular eukaryotes. In the budding yeast Saccharomyces cerevisiae
, for example, unequal partitioning of specific transcription factors causes mother and daughter cells to express different genes late in division [3
Asymmetry of intracellular cell fate determinants requires their physical segregation as well as a mechanism to ensure that they do not act before the differentiating cells are functionally separated. In a number of well-characterized cases, transcriptional regulators are directly partitioned by cytoskeleton-associated machinery.
In Drosophila melanogaster
, differentiation of neuroblasts and ganglion mother cells (GMCs) is achieved through asymmetric segregation of the transcription factor Prospero's protein and mRNA, in association with the adaptor proteins Miranda and Staufen [8
]. Miranda's segregation to the cortex of the presumptive GMC involves the actin cytoskeleton and the opposing activities of myosin VI and myosin II; this is mitotically regulated by the anaphase-promoting complex/cyclosome [11
]. In the next cell cycle, Prospero translocates to the GMC nucleus, where it regulates transcription of GMC-specific genes. In budding yeast, daughter cells are prevented from switching mating types by the asymmetrically segregated transcriptional repressor Ash1. This partitioning also depends on the actin cytoskeleton: ASH1
mRNA is transported by a class V myosin to the bud tip during mitosis and tethered to the daughter cell cortex. It remains there until the beginning of the next cell cycle, whereupon it is translated to produce the Ash1 repressor protein [4
Asymmetric gene expression is also important in the last step of budding yeast cell division, but is generated by a different mechanism. Final separation of mother and daughter yeast cells requires removal of a chitin-rich septum constructed between the cells during cytokinesis [12
]. Destruction of this septum occurs from the daughter side. This asymmetry is due to a daughter-specific transcriptional program driven by the transcriptional activator Ace2, which accumulates specifically in the daughter cell nucleus and induces expression of enzymes involved in septum degradation [3
]. Partitioning of this transcription factor is independent of mechanisms required for ASH1
], and remains incompletely understood.
Ace2′s activation and daughter nucleus accumulation are coordinated with mitotic exit [7
]. The transcription factor first localizes faintly to both mother and daughter nuclei, then accumulates to high levels exclusively in the daughter nucleus at the end of mitosis. The timing of this accumulation relative to cytokinesis is uncertain. Ace2′s nuclear import is likely blocked by mitotic cyclin-dependent kinase (CDK) phosphorylation of sites near its nuclear localization sequence (NLS) [15
]: this inhibition is presumably reversed when CDK phosphorylations are removed by the phosphatase Cdc14 during mitotic exit. Ace2 nuclear export depends on the exportin Crm1/Xpo1 [7
]. Loss of nuclear export results in symmetric Ace2 accumulation in both mother and daughter nuclei, indicating that the transcription factor is isotropically distributed in mother and daughter cells and that its asymmetry is probably not due to selective import in the daughter cell nucleus or degradation in the mother cell nucleus.
Ace2 is controlled by a conserved signaling pathway termed the Regulation of Ace2 and Morphogenesis (RAM) network [3
]. Cells lacking RAM network function fail to separate, growing as large clusters of cells connected by the primary septum between mother and daughter cells. This separation defect is the result of failure to segregate and activate Ace2 in daughter cells and thus a lack of expression of the Ace2 target genes required for septum destruction. It is unclear how the RAM network promotes the daughter-specific segregation and activation of Ace2.
Cbk1, a protein kinase of the broadly conserved NDR/LATS family [19
], is a critical component of the RAM network. Cbk1 localizes to the bud neck and daughter cell nuclei during mitosis; the kinase's nuclear localization requires Ace2 [3
]. Cbk1 kinase activity is critical for Ace2 localization and activation: in cells lacking Cbk1, Ace2 localizes faintly to both mother and daughter nuclei and cannot activate transcription of its target genes [7
]. The kinase phosphorylates an N-terminal fragment of Ace2 in vitro, suggesting a direct regulatory connection [20
]. However, the identity and functional significance of Cbk1 phosphorylation sites within Ace2 are unknown.
In this study, we sought to understand how Cbk1 controls Ace2. We used an unbiased approach to elucidate the kinase's phosphorylation consensus motif and find a distinctive specificity that is likely conserved in related kinases across large evolutionary distances. This motif identified three Cbk1 phosphorylation sites within Ace2 that are crucial for the transcription factor's asymmetric distribution and function. Our in vivo and in vitro analyses of the functional significance of these sites indicate that Cbk1 phosphorylation controls Ace2 in two distinct ways: by directly blocking its interaction with nuclear export machinery and by enhancing its activity as a transcription factor. We also found that Cbk1 promotes Ace2 segregation well before cytokinesis and that the kinase is functionally partitioned to the daughter cell, allowing it to phosphorylate Ace2 and generate asymmetry from an initially isotropically distributed pool of the transcription factor.