Phosphorylation of Kibra at T971 regulates organ size in Drosophila
Overexpression of Kibra in the developing eye of Drosophila (using the UAS/GAL4-system with eyeless::GAL4, ey::GAL4) or in the posterior compartment of the wing (using engrailed::GAL4, en::GAL4) led to a decrease of organ size Fig. 1B-E and 16. Notably, mutation of threonine 971 to alanine (T971A), resulting in a non-phosphorylatable version of Kibra, strongly enhanced this phenotype (Fig. 1B-E), whereas overexpression of Kibra T971D, mimicking a constitutive phosphorylation, resulted in a slightly but not significantly weaker organ size reduction compared to overexpression of wild type Kibra (Fig. 1B-E). By contrast, mutation of two conserved prolines within the WW-domains of Kibra, disrupting the binding capacity of these domains abolished the ability of Kibra to reduce organ size (Fig. 1B-E), indicating that a protein-protein interaction mediated by the WW domains might be essential for this phenotype.
Phosphorylation of Kibra at T971 is not essential for viability but regulates cell proliferation and Drosophila body size under endogenous conditions
As overexpression of Kibra might lead to artificial activation of signaling pathways regulating cell proliferation and organ growth, we next tested whether Kibra T971 phosphorylation is essential under endogenous expression conditions, too. For this, we used CRISPR/Cas9 to establish a knockin of either wild-type T971, T917A or T971D Kibra proteins. Surprisingly, all fly lines expressing the different phosphorylation variants displayed similar survival rates (Fig. 2A). However, analysis of body size revealed that animals carrying two phosphorylation-deficient kibra alleles are significantly smaller compared to wild type or T971D animals (Fig. 2B). Next, we tested, whether this effect is cell-autonomous by inducing clones mutant for kibra-variants in an otherwise wild type tissue by using mosaic analysis with a repressible cell marker (MARCM). Indeed, clones of kibra T971A are significantly smaller compared to those cells expressing wild type Kibra or Kibra T971D (Fig. 2C-F), indicating a decreased cell proliferation in cells lacking Kibra T971 phosphorylation.
Organ size regulation by Kibra does not depend on Hippo signaling
Kibra is a well described upstream regulator of the Hippo signaling cascade first by forming a scaffolding platform for Salvador/Hippo together with Merlin/Expanded and second (at least in mammals) by increasing the phosphorylation and activation of Lats by binding through its WW-domains6, 16. The observation that overexpression of Kibra results in decreased organ size in wings and eyes would be in line with increased Hippo pathway activation leading to decreased Yki activity and thus downregulation of Yki target genes. Following this line, phosphorylation-deficient Kibra should show a stronger downregulation of Yki/YAP target genes. Therefore, we first tested human KIBRA in a luciferase-based YAP-reporter assay17 and found that expression of KIBRA induces a downregulation of YAP activity (as reported earlier), but not changes between wild type and phosphorylation-deficient KIBRA (KIBRA T929A) (Fig. 3A).
In order to further test our hypothesis in vivo, we generated flies overexpressing the T971A variant together with point mutations, which inactivate the WW-domains (P85A P132A = ∆WW). As expected, overexpression of Kibra∆WW T971A did not result in decreased organ size in eyes or wings but showed similar organ sizes as control flies or flies expressing Kibra∆WW alone (Fig. 3B-C).
Next, we directly assessed Yki targets in vivo by overexpression of GFP-Kibra in the posterior compartment of imaginal discs, which expressed β-Galactosidase under control of the expanded promoter (ex::lacZ) or the four-jointed promoter (fj::lacZ), which are both activated by Yki18–20. In addition, we stained for Cyclin E expression, which is also a downstream target of the Hippo pathway18. If a decreased Yki activation would be the reason for the decrease in organ size in Kibra overexpressing tissues, expression of ex::lacZ/fj::lacZ as well as Cyclin E should be decreased. Surprisingly, we found a clear upregulation of Cyclin E and if at all a minimal upregulation but no downregulation of ex::lacZ/fj::lacZ expression in the posterior compartment of the wing imaginal discs, where wild type Kibra or Kibra T971A was overexpressed (Fig. 3D-G).
Taken together, our data indicate that increased levels of KIBRA indeed enhance Hippo signaling, resulting in decreased YAP activation in cell culture. However, this effect cannot be observed in vivo, suggesting that the Kibra overexpression phenotype of reduced organ size is not the consequence of enhanced Hippo signaling leading to reduced Yki activation in vivo.
Furthermore, phosphorylation of Kibra by RSK does not affect its function in Hippo-pathway regulation, arguing against an involvement of Hippo in the strongly enhanced organ size reduction upon overexpression of phosphorylation-deficient Kibra.
Phosphorylation of Kibra regulates cell cycle progression
As the known role of Kibra in Hippo pathway regulation did not explain the reduced organ growth observed upon its overexpression, we investigated cell cycle progression in imaginal discs overexpressing wild type KIBRA or phospho-deficient KIBRA. To discriminate distinct cell cycle phases, we used the FlyFUCCI system21: In brief, we analyzed imaginal discs constitutively expressing GFP, which was fused to an E2f1 degron and RFP, fused to Cyclin-B degron. Thus, cells in G1 express only GFP, whereas in S-phase, only RFP is expressed and G2-/M-phase cells retain both fluorochromes. Kibra and Kibra T971A were overexpressed in the posterior compartment of the wing imaginal discs (marked by expression of engrailed, Fig. 4A). Cells in G1 were normalized to the total amount of cells (DAPI staining) and the quotient of posterior cells (engrailed positive and expressing Kibra wt/T971A) to anterior cells (no Kibra overexpression). As quantified in Fig. 4B, overexpression of wild type Kibra enhances the number of cells in G1 by ca. twofold. Moreover, overexpression of phosphorylation-deficient Kibra resulted in a more than four-fold increase in cells in G1-phase.
Kibra phosphorylation regulates binding to 14-3-3 and Cdk4 to control cell cycle progression
In silico analysis of the phosphorylation motif revealed a conserved consensus motif for 14-3-3- proteins (R-S-X-T-X-P). 14-3-3 proteins are adapter proteins which bind to phosphorylated motifs, thereby facilitating e.g. protein degradation, displacement from the plasma membrane or blocking of protein-protein interactions22. Furthermore, in an earlier proteomic screen for KIBRA interaction partners, we identified 14-3-3 proteins as well as the Cyclin-dependent kinase 4 (Cdk4) to co-immunoprecipitate with KIBRA (data not shown). We verified that 14-3-3 and Cdk4 associate with KIBRA using co-immunoprecipitation from WWC1/2 deficient HEK293 cells expressing GFP-KIBRA variants (Fig. 4A and Supplementary Fig. 1). Notably, KIBRA T929A displayed a strongly increased binding of Cdk4 compared to wild type KIBRA, whereas binding of 14-3-3 is decreased, suggesting that phosphorylation of KIBRA T929 affects the association with both proteins contrarily. As Cdk4 is a critical regulator of G1 phase length23, we hypothesized that binding of KIBRA to Cdk4 controls cell cycle progression and thus cell proliferation and organ size. Phosphorylation of KIBRA at T929 (T971 in Drosophila) by RSK, induced by mitogen stimuli releases Cdk4 from KIBRA, resulting in cell cycle progression and increased proliferation. Impaired phosphorylation of KIBRA by expression of T929A/T971A decreases or delays the release of Cdk4, thus inhibiting cell cycle progression and proliferation. Indeed, in Drosophila, inhibition of RSK in KIBRA overexpressing wings, which results in impaired phosphorylation of KIBRA T971, led to a decreased organ size in comparison to downregulation of RSK alone or overexpression of wild type KIBRA alone (Supplementary Fig. 2).
Finally, the overexpression phenotype of Drosophila KIBRA T971A can be diminished by simultaneous overexpression of Cdk4, while overexpression of Cdk4 alone had no effect on eye or wing size (Fig. 5B-E). Moreover, overexpression of Cdk4 during embryogenesis could rescue the increased lethality of KIBRA T971A overexpression to large extent (Fig. 5F). These data confirm that KIBRA regulates cell cycle progression and thus organ growth by binding Cdk4 and that phosphorylation of KIBRA at T971 by RSK regulates this interaction.