Cells compartmentalize activities not only with membranes, but also by employing physical principles of phase separation. Weak interactions between relatively unstructured proteins and nucleic acids in particular can induce a phase separation that leads to the formation of discrete accumulations in the cytosplasm that are enriched in proteins and other molecules with affinity for this unmixed phase. We recently discovered a novel principle of coupling liquid unmixing to signal transduction by studying the dual specificity kinase DYRK3 .

Liquid unmixing by DYRK3

GFP-DYRK3 condenses P-granule-like speckles at a certain expression threshold. A total of 241 cells were automatically segmented using computational image analysis, mean fluorescent intensity was measured per cell, and cells containing a homogeneous distribution [H] or granules [G] of GFP-DYRK3 were classified by SVM training. Switching between a homogeneous to granular distribution occurs suddenly at a narrow expression threshold.


Activity-dependent dynamic cycling of DYRK3 between liquid-umixed stress granules (SGs) and the cytosol and mechanism of mTORC1 reactivation by DYRK3. During stressful conditions, stress-induced translational silencing induces the condensation of SGs. DYRK3 will partition in SGs via its N-terminal domain, as well as mTORC1 components, which prevents mTORC1 signaling. When stress signals are gone, the kinase activity of DYRK3 is required for the dissolution of SGs and mTORC1 relocation to the cytosol, and for phosphorylating PRAS40, which attenuates the binding of PRAS40 to the mTORC1 complex. Combined, this allows reactivation of mTORC1 signaling.


Current questions:

  • To which extent are the other DYRK family members and related kinases in the human genome involved in liquid unmixing? And do these other family members regulate and sense liquid unmixing of different structures, relevant for regulation of the cytoskeleton or splicing?
  • Does DYRK3 phosphorylate its own N-terminus and at which site? Does that control its cyclic partitioning between RNA granules and the cytosol?
  • Does the effect of DYRK3 on the phosphorylation of RNA-binding proteins participate in RNA granule dissolution?
  • Does DYRK3 couple liquid unmixing to other signal transduction pathways?


Current lab members involved:

Arpan Rai

Doris Berchtold


Some relevant publications from the lab: