, 2003). We generated hemagglutinin (HA)-DAXX constructs expressing nonphosphorylatable (S669A) and phosphomimetic (S669E) DAXX mutants. Whereas S669E DAXX migrated like hyperphosphorylated DAXX, migration of the S669A mutant corresponded to hypophosphorylated DAXX (Figure 5H). Overexpression of an active form of calcineurin led to reduced migration of wild-type (WT) DAXX but did not affect the two mutants (Figure 5H). Similarly, coexpression NVP-BGJ398 in vitro of HIPK1 promoted hyperphosphorylation of WT DAXX only (Figure 5H). These results indicate that DAXX S669 phosphorylation is modulated by calcineurin. We next explored whether the phosphorylation status of DAXX regulates its interaction with H3.3 and ATRX. As shown in Figure 3A, we found

an enrichment of endogenous hypophosphorylated DAXX in YFP-H3.3 immunoprecipitates in neurons. Similar findings were obtained with exogenously expressed WT DAXX in 293T cells (Figure 5I) as well as in neurons (Figure S5B). HIPK1 overexpression led to DAXX hyperphosphorylation, but only a small proportion of hyperphosphorylated DAXX was found to be check details associated with H3.3 (Figure 5I). This enrichment did not appear due to reduced H3.3 affinity for hyperphosphorylated DAXX, because similar levels of S669E and S669A mutants were found to be associated with H3.3 (Figure 5I). Finally, we failed

to detect any effect of DAXX phosphorylation status on its ability to interact with ATRX (Figure S5C). Because DAXX/H3.3 complexes are enriched in hypophosphorylated DAXX, we reasoned that DAXX phosphorylation status could play a role in the regulation of H3.3 deposition. To test this hypothesis, we performed rescue experiments in DAXXFlox/Flox neurons. CRE promoted efficient deletion of endogenous DAXX in cells coinfected either with a green fluorescent protein (GFP) vector or DAXX constructs ( Figures S6A–S6C). Similar expression levels of WT, S669A, and S669E DAXX were achieved in transduced neurons ( Figure 6A). Upon membrane depolarization, migration

3-mercaptopyruvate sulfurtransferase of S669A and S669E DAXX mutants was not affected, whereas levels of hyperphosphorylated WT DAXX decreased ( Figure 6A). Furthermore, no significant differences in association with Bdnf Exon IV and c-Fos regulatory regions were detected in between the constructs both at steady state and upon KCl treatment ( Figure 6B). As expected, WT DAXX rescued H3.3 loading at Bdnf Exon IV and c-Fos regulatory regions in CRE-infected DAXXFlox/Flox neurons ( Figure 6C). Notably, S669A DAXX had a more pronounced rescuing activity at most regions analyzed ( Figure 6C). Conversely, S669E DAXX failed to rescue loading at all regions ( Figure 6C). We then tested whether DAXX phosphorylation also affected its ability to regulate transcription. WT and S669A DAXX rescued expression of Bdnf Exon IV and c-Fos. In contrast, S669E DAXX was impaired in this function ( Figure 6D). Notably, S669A DAXX was more potent in rescuing c-Fos induction compared to WT DAXX ( Figure 6D).