TI - Inactivation through PHOSphorylation of Runx2 by GSK-3beta . AB - We next examined the molecular mechanism underlying the GSK-3beta inhibition of bone formation . In the two major osteogenic signalings in which GSK-3beta is known to be involved , i.e. , the canonical Wnt/beta-catenin and the PI3K/Akt signalings [5] . A recent in vivo study showed that beta-catenin hardly affected osteoblasts through a cell -autonomous mechanism [16] . Considering that the other signaling PI3K/Akt is related to Runx2 transactivation in its osteogenic action [17] , we examined the involvement of Runx2 in the GSK-3beta regulation of bone formation . We initially confirmed both GSK-3beta and Runx2 expressions in the calvaria , tibia , and cultured osteoblasts (Fig 3A) . Bone formation determined by von Kossa staining and the osteocalcin mRNA level was enhanced by the Runx2 overexpression in both Gsk-3beta+/+ and Gsk-3beta+/- calvarial osteoblast cultures (Fig 3B) . To examine the regulation of transcriptional activity of Runx2 by GSK-3beta , a luciferase reporter gene construct containing a 1,050 bp osteocalcin gene fragment ( 1,050 OC-Luc ) including the Runx2 binding sites was transfected into human hepatoma HuH-7 cells . The luciferase reporter analysis revealed that the Runx2 -dependent transcription was suppressed by the co-expression of wild-type GSK-3beta and CA-GSK-3beta , but not by that of KI-GSK-3beta (Fig 3C) , whereas it was enhanced by lithium chloride and SB216763 (Fig 3D) . Collectively , these data demonstrate that the kinase activity of GSK-3beta suppresses the Runx2 transcriptional activity . To further investigate how GSK-3beta is involved in the Runx2 activity , we examined the effects of CA-GSK-3beta overexpression , lithium chloride treatment , and the genetic GSK-3beta insufficiency on the expression and subcellular localization of Runx2 , and found that none altered either of them ( Fig 4A ) . We then transfected Gsk-3beta+/+ and Gsk-3beta+/- osteoblasts with Runx2 , and compared the binding of the nuclear extracts with the oligonucleotide probe of the Runx2 binding sequence , osteoblast-specific cis-acting element 2 ( OSE2 ) of the mouse osteocalcin gene promoter [18] , by electrophoretic mobility shift assay ( EMSA ) . We found a complex that was confirmed to represent the Runx2-OSE2 binding , since it diappeared by the addition of 50-fold excess of unlabeled wild-type OSE2 probe , but not by the mutated probe lacking the Runx2 binding sequence , and was undetectable when the nuclear extract from cells without Runx2 transfection was used ( Fig 4B ) . The specific complex was augmented by the Gsk-3beta+/- nuclear extracts as compared to that by the Gsk-3beta+/+ extracts , indicating that GSK-3beta attenuates the DNA binding activity of Runx2 . We then investigated biochemical interactions between Runx2 and GSK-3beta by co-immunoprecipitation assay , which showed the possible direct interaction between these two molecules ( Fig 4C , D ) . To learn the contribution of the phosphorylaTION of Runx2 by GSK-3beta to the attenuation of Runx2 transcriptional activity , we generated phosphorylaTION-deficient mutants of Runx2 by creating three to four amino acid replacements at the five consensus sites for the PHOSphorylation by GSK-3beta [5] : S92A-S96A-S100A , S369A-S373A-S377A , S389A-T393A-S397A , T394A-S398A-T402A and T476A-T480A-S484A-S488A . The luciferase reporter analysis using the 1,050 OC-Luc -transfected HuH-7 cells revealed that the phosphorylaTION-deficient mutant at S369-S373-S377 enhanced the transcriptional activity , while mutations at the other four phosphorylaTION sites showed comparable activity to the wild-type Runx2 , indicating that the specific PHOSphorylation at S369-S373-S377 suppresses the Runx2 activity (Fig 4E) . In vitro kinase assay confirmed that the Runx2 PHOSphorylation by GSK-3beta was reduced by the S369-S373-S377 mutation (Fig 4F) . When we compared the DNA binding of nuclear extracts from HeLa cells transfected with wild-type and the S369-S373-S377 mutant Runx2 by EMSA , the mutation enhanced the specific Runx2-DNA binding (Fig 4G) . Finally , the luciferase reporter analysis disclosed that the regulations of Runx2 -dependent transcription by gain - and loss-of-functions of GSK-3beta , i.e. , suppression by CA-GSK-3beta overexpression and enhancement by lithium chloride , were cancelled by the S369-S373-S377 mutation (Fig 4H) . These lines of results demonstrate that the PHOSphorylation of Runx2 at S369-S373-S377 by GSK-3beta attenuates the transcriptional activity of Runx2 , leading to the suppression of bone formation .