Deciphering the genetic instructions that establish neuronal differentiation and neuronal connection during development is one of the keys to understand the complexity of the adult brain. For the nervous system, as for the entire body, our understanding of patterning process has advanced rapidly through the elucidation of functions of novel genes that control development. Thus, my laboratory has adopted developmental approaches and molecular biology tools to gain more insight about how mammalian nervous system organizes and functions by revealing functions of uncharacterized neural genes.

Two intertwined research directions in my laboratory

The first direction of my research interest is to analyze functions of uncharacterized protein kinases whose expression is preferentially present in the developing nervous system. Previously we identified a novel nuclear CDC2-like, arginine/serine (RS)-rich protein kinase from embryonic day 14 (E14) rat cortex. We found the protein encoded by this gene is collaborated with Cyclin L1 and Cyclin L2. Thus, we named this kinase cyclin-dependent kinase 12 (Cdk12). In the cellular level, Cdk12 is colocalized with the splicing factor, SC35, in nuclear speckles. We further demonstrated that Cdk12 is a novel type of alternative splicing regulator by counteracting SC35 and SF2/ASF. Through sequence comparison and collaboration with Dr. Anne-Marie Geneviere in Université Pierre et Marie Curie-Paris 6, France, we demonstrated that CDC2L5, a protein with sequences very similar to those of Cdk12, also interacts with Cyclin L1 and Cyclin L2, and alters splicing pattern in a manner similar to Cdk12. We renamed CDC2L5 as Cdk13. Our first efforts on characterization of Cdk12 and Cdk13 are recognized by a recent correspondence on defining the cyclin-dependent kinases. The developmentally regulated expression pattern of Cdk12 and Cdk13 in the developing nervous system implicates that they may be involved in the neural development. We thus established P19 neuronal differentiation cell model and primary cortical neuronal culture and demonstrated that Cdk12 and Cdk13 are required for the axonal elongation. In the absence of Cdk12 or Cdk13 a substantial decrease of neurons with long axons was detected. A common signaling pathway mediates effects for Cdk12 and Cdk13. Overexpression of Cdk13 in Cdk12-depleted cells partially restores the defects of Cdk12 knock-down, and vice versa. We further provided evidence that Cdk5 is a common downstream effector by showing that Cdk12 and Cdk13 regulate Cdk5 expression at the RNA level in vitro and in vivo. In addition, overexpression of Cdk5 protein in P19 cells is able to partially rescue the axonal extension defect observed when Cdk12 or Cdk13 is depleted. Our data thus revealed a novel way to control axonal elongation and regulate expression of Cdk5 whose function is critical in cortical lamination, neuronal cell migration and axon outgrowth.

We have also generated conditional Cdk12^-/- mice as an in vivo model to examine functions of Cdk12 during mouse development. Constitutive deletion of Cdk12 leads to embryonic lethality at E5.5~E6.5. Phenotypes of conditional Cdk12^-/- mice after crossing with Nestin-Cre include many defects in the developing cerebral cortex. The detailed phenotypes and underlying mechanisms are the topics of the current research.

The second research direction is to search surface molecules that may function in early neural development and could be used as stem cell markers. By subtractive hybridization, we had found an uncharacterized member of immunoglobulin family, protogenin (Prtg), whose expression is very abundant in the E7~E9 mouse embryo but disappears after E10.5. Using the P19 cell line as a neural differentiation model, we demonstrated that the expression of Prtg is quickly suppressed when cells are committed to differentiation under influence of retinoic acid. We reasoned that this gene is expressed in primitive neural progenitor cells, and is down-regulated when cells enter into differentiation mode. Using yeast two-hybrid assay, we identified one molecule, ERdj3, as a ligand that binds Prtg. ERdj3 is a DnaJ-domain containing protein present in ER and is also secreted out of cells. Interestingly, activation of Prtg through ERdj3 prevents precocious neuronal differentiation in P19 cells and in chick neural tube. We had also prepared conventional Prtg knockout mice. Prtg^-/- mice show abnormal craniofacial bones and display anterior transformation in axial skeletons. We demonstrated that Prtg is required for the survival of the rostral cephalic neural crest cells during E9 and E10. In the absence of Prtg, increased death of these cells is detected. As a result, fewer rostral cephalic neural crest cells populate branchial arch 1 and frontonasal primordium during the period. Later in development, the loss of these rostral cephalic neural crest cells indirectly delays the development of mesoderm-derived mesenchymal cells. A combination of these direct and indirect effects results in several skeletal defects in upper jaw and skull vault in Prtg^-/- mutant mice at birth. We further elucidated the molecular signaling mechanism that mediates Prtg effects. We demonstrated that Radil is its downstream effector. Upon activation of Prtg by its ligand ERdj3, Radil translocates to cell membrane and conducts an inside-out signaling that triggers conformational changes of the α5β1 integrin. Our data thus revealed a novel way to regulate integrin activity, which is important for survival of rostral cephalic neural crest cells.