Fig 1.
Role of the NDR/LATS kinase–Mob complex in hippo signaling pathways.
Left: A metazoan hippo pathway in which a LATS–Mob1 complex is activated by MST/hippo kinases and inhibits the Yki/YAP transcriptional coactivator, based on research from Drosophila and mammalian cells. Right: Budding yeast have two distinct hippo pathways, the mitotic exit network (MEN), in which the LATS-related Dbf2 or Dbf20 kinase in complex with Mob1 controls mitotic exit and cytokinesis, and the RAM network, in which the NDR-related Cbk1 kinase in complex with Mob2 controls cell separation and morphogenesis.
Fig 2.
Structure of the Cbk1–Mob2 complex.
(A) An overview of the Cbk1–Mob2 complex. Mob2 (orange) binds to the N-terminal region of Cbk1 (blue) through a large surface formed by the H2 and H7 helix and the H4–H5 loop. “N” and “C” denote the N- and C-terminal ends of the protein constructs. Flexible regions that could not be located in the electron density maps are shown with dashed lines. HM is the AGC kinase HM, and “αINH” denotes the inhibitory alpha helix that is part of the activation loop (shown in cyan). (B) The N-terminal region of the Cbk1 forms a bipartite Mob2-binding surface comprised of a long αMob helix and a highly conserved arginine rich N-linker region. (C) An activation loop segment (cyan) of Cbk1 containing the major autoregulatory site (S570) blocks access to the active site (D475) (left panel). An inhibitory helix (αINH in [A]) is anchored in the substrate-binding groove by several salt bridges (shown with black dotted lines). The substrate binds in the negatively charged “open” binding pocket that replaces αINH (shown in cyan), as seen in the state with the bound αINH (right panel). The model for the Cbk1–substrate complex was generated by superimposing the crystallographic model of the “open” state of Cbk1 with PKA binding to a phosphorylated substrate (Protein Data Bank [PDB] ID: 1JLU). The black cartoon with gray spheres (from P-1 to P-5 Cα positions) shows the PKA substrate superimposed in the Cbk1 substrate-binding pocket, while the activation loop of Cbk1, which displays the characteristics of a pseudo-substrate region, is shown with the semitransparent cartoon representation in cyan. The surface of Cbk1 is colored according to its electrostatic potential, where red indicates negatively charged surface. The nucleotide cofactor (AMPPNP) is shown in stick representation.
Fig 3.
Molecular dynamics simulation indicates phosphorylation-induced rearrangement of the NDR/LATS kinase HM-binding slot to favor enzyme activation.
(A) Comparison of the Cbk1 HM with the HM of PKB. The panel shows the Cbk1–Mob2 complex superimposed with PKB, where the αC helix and the HM of the superimposed PKB are shown in yellow (PDB ID: 1O6K). Side chains of important HM residues are shown with sticks. Superimposition of PKB and Cbk1–Mob2 reveals that the main chains of the HM motifs are >7 Å apart. The Cbk1 HM motif is shifted upwards to bind to the upper part of the HM-binding slot, while the PKB HM binds close to the N-terminal PKB kinase lobe. (B) Cbk1–Mob2 complexes in which T743 in the C-terminal HM is either dephosphorylated (HM-T) or phosphorylated (HM-P) were subjected to 125-ns MD simulations. Here we represent indicated regions of Cbk1 as centers of mass of the following: HM, amino acids 742–749; N-linker, amino acids 341–346; αC, amino acids 393–406; αMob, amino acids 297–317. We started distances of both HM forms to the N-linker region, αC, and αMob at [8.7 Å; 18.0 Å; 14.9 Å], corresponding to the conformational state captured in the Cbk1–Mob2 complex crystal structure, and recalculated the distances between HM forms, N-linker, αC, and αMob after each simulation step. (C) Three-dimensional scatter plots show 100 ns of the simulations, with each dot indicating N-linker–HM, αC–HM, and αMob–HM distances (x-, y-, and z-axes) for each nanosecond. For clarity, dots are colored from blue to red according to their αMob–HM distances (z-axis). The complex with HM-P exhibits shorter distances and smaller changes in position, indicating that T743 phosphorylation compresses the HM-binding slot and constrains the HM region association. This occurs only in Cbk1 bound to Mob coactivator (see S5 Fig). (D) HM phosphorylation promotes αMob bending and compresses the HM binding slot. Secondary structure analysis of αMob illustrates that this region remains helical (blue) when HM-T is bound, but acquires a short turn (yellow) when HM-P is bound. (E) An enlarged view of the final MD-simulated Cbk1(HM-P)–Mob2 complex indicates the HM-P’s interactions that bend αMob, as well as van der Waals contacts with αC. (F) Ionic interactions suggested by the MD model (E314/R746 and R307/pT743) (left) were tested by charge swapping and analysis of cell separation. Single mutants (grey) displayed larger group sizes, indicating defective RAM network function, but were partially recovered when both charged sites were swapped (red). Data for (C) and (F) can be found in S1 Data.
Fig 4.
Molecular analysis of Cbk1 docking peptides in Ace2 and Ssd1 highlights importance of a [YF]xFP motif.
(A) Protein schematics with points of interest highlighted. Orange lines denote Cbk1 consensus sites, and blue boxes denote docking motifs. (B) Pulldown of Cbk1 kinase domain by truncated Ssd1 N-terminal (1–5) and C-terminal (6–11) docking motifs suggests that only the FKFP motif is required for interaction. (C) Alanine scan of the Ace2270–290 docking motif suggests that residues N-terminal to the YQFP motif aid in Cbk1 binding. (D) Mutational analysis of Ssd1 N-terminal docking motif highlights the sequence stringency of the core motif and suggests a consensus docking motif of [YF][KR]FP. (E) Cbk1 in vitro kinase assay with Ace2 or Ssd1 fragments containing a phosphorylation site (HxRxx[ST]) and either a WT docking motif (left) or mutated docking motif (dock*, right). Phosphorylation is enhanced in the presence of the WT docking motif.
Fig 5.
Structural model of the Cbk1–docking peptide complex.
(A) AutoDock binding simulation of Ssd1 docking fragment (DFKFP) identifies an interaction surface on Cbk1 and suggests important aromatic interactions with residues W444, F447, and Y687. (B) Y687A but not nearby F699A abrogate Cbk1 pulldown by Ssd1208–214 as predicted by the model. (C) FP analysis of putative docking motif binding site residues. The binding of all mutants to pepHM-P was similar to that of WT Cbk1, indicating that the mutants are folded properly (see first column in the table summarizing peptide binding affinities). W444A, F447A, and Y687A were reduced in their binding affinity for Ssd1 and Ace2 docking peptides, while F699A was not. (D) Peptide kinase assay of WT and Y687A mutant of Cbk1251–756–Mob2. Peptides are based on Ssd1 (210–229), which contains the N-terminal docking motif (Docktide: FKFP, AKAPtide: AKAP) and a Cbk1 consensus site. The presence of the WT docking motif enhances phosphorylation by WT Cbk1 relative to a peptide containing a mutant docking motif (similar to Fig 4F). Cbk1(Y687A) loses this enhancement when unable to bind WT docking motif. Data is plotted as mean ± standard error of the mean. Student’s unpaired t-test: **p < 0.01; ns, not significant. Data for (C) can be found S2 Data, and data for (D) can be found in S3 Data.
Fig 6.
Docking sites increase robustness of Cbk1 control of in vivo substrates.
(A) Mutation of the Cbk1 docking motif in Ace2 (ace2dock*) confers a modest defect in cell separation in cells with WT CBK1. (B) The ace2dock* allele has a marked cell separation defect in cells carrying the cbk1-as2 allele, which is catalytically weakened. Note that cbk1-as2 cells exhibit no cell separation defect when the WT ACE2 allele is present. (C and D) CBK1 WT cells carrying ace2dock* have a slight reduction in CTS1 transcript levels (C), while cbk1-as2 cells carrying ace2dock* exhibit strongly reduced CTS1 transcription (D). (E) Overexpression of the ssd1dock* allele, carrying mutations that eliminate docking interaction with Cbk1, affects viability of WT cells, and this is far more dramatic in cbk1-as2 cells. Cells were 10-fold serially diluted from left to right and plated on galactose (inducing) or glucose (repressing) media: reduced colony formation in serial dilutions reflects impaired viability. Data for (A) and (B) can be found in S4 Data, and data for (C) and (D) can be found in S5 Data.
Fig 7.
Co-conservation of docking and phosphorylation sites identifies known and likely Cbk1 regulatory targets.
Conservation of Cbk1 docking motifs and phosphorylation consensus sites in individual proteins, calculated using ConDens [48]. For each motif type, the most significant score among the matches in each protein is assigned as the score for that protein. The x-axis plots conservation scores of [YF]xFP sequences (Cbk1 docking), and the y-axis plots conservation scores of Hx[RK]xx[ST] sequences (Cbk1 phosphorylation consensus). The known Cbk1 substrates Ssd1 and Ace2 are highlighted in red. Proteins without docking motifs that score very highly for phosphorylation consensus conservation are noted. ConDens scores can found in S6 Data.
Fig 8.
Model for role of Mob binding and HM phosphorylation in NDR/LATS kinase activation.
Mob supports the NDR/LATS kinase N-terminal extension in a conformation that creates a binding slot for both HM-T and HM-P. (A) HM-T remains associated with the N-linker region in a manner that does not promote an ordered conformation of the αC helix. (B) HM-P promotes reorganization of αMob, compressing the HM binding slot and holding aromatic side chains of the HM in a configuration that favors ordering of αC.
Fig 9.
A probable network of proteins controlled by the yeast RAM network, and sequential evolution of docking in Cbk1 substrates.
(A) An outline of known and likely Cbk1 substrates in which the core docking motif is conserved (green lines) and likely substrates that lack a docking motif (blue lines). Ace2 and Ssd1 (red type) are known Cbk1 phosphorylation targets, and black lines indicate regulatory interactions inferred from molecular genetic analysis. Orange and green arcs indicate localization of target proteins to either the nucleus or sites of cell growth and cortical remodeling (bud neck/cytokinesis site/cell cortex): Cbk1 displays both of these localization patterns. (B) Distribution of phosphorylation consensus sites and docking motifs in known and likely Cbk1 substrates. Orthologs of Ace2, Boi1, and Tao3 are present in the lineages drawn. Blue indicates lineages in which the proteins have conserved phosphorylation consensus sites, while in those drawn in black, consensus sites are absent. Inferred acquisition of consensus sites is represented by a blue star. Lineages in which the proteins have conserved Cbk1 docking motifs are highlighted green, with inferred acquisition of this motif labeled by a yellow star. (C) Sequential addition model for acquisition of regulatory motifs during signaling system evolution. Substrates first become phosphorylation targets of a given upstream kinase. Docking motifs are subsequently acquired in unstructured, rapidly evolving sequence. The robustness of substrate phosphorylation increases upon docking motif acquisition. A. nid, Aspergillus nidulans; C. alb, Candida albicans; K. lac, Kluyveromyces lactis; N. cra, Neurospora crassa; S. cer, Saccharomyces cerevisiae; S. pom, Schizosaccharomyces pombe.