In pluripotent cells phosphorylation has

In pluripotent cells, phosphorylation has a central role in directing cell identity by relaying growth-factor signaling through key pathways (i.e., fibroblast growth factor [FGF] and transforming growth factor β [TGF-β]) (Chen et al., 2011; Singh et al., 2012a; Vallier et al., 2005; Yu et al., 2011). The ultimate targets of these phosphorylation cascades are largely unknown, although recent works have provided some direction by mapping phosphorylation on the pluripotency factors OCT4 and SOX2 (Brumbaugh et al., 2012; Jeong et al., 2010; Phanstiel et al., 2011). Conspicuously absent in these studies is NANOG, a divergent homeobox transcription factor that promotes pluripotency by binding to DNA and regulating the expression of genes related to cell fate (Boyer et al., 2005; Chambers et al., 2003; Mitsui et al., 2003; Pan and Thomson, 2007). In mice, overexpression of NANOG permits extended culture of undifferentiated embryonic stem order N6022 (ESCs) in the absence of otherwise obligatory extrinsic signaling factors such as LIF and BMP4 (Chambers et al., 2003; Pan and Thomson, 2007). Correspondingly, NANOG overexpression in human ESCs obviates the requirement for exogenous signaling through feeder cells in basal media or FGF in defined culture systems (Darr et al., 2006; Xu et al., 2008). Hence, NANOG has a conserved role in mediating growth-factor signals that are critical for pluripotency, and, intriguingly, its overexpression is sufficient to bypass these signaling pathways to maintain the ESC state. Still, a direct link between NANOG and the signaling molecules that determine cell state remains elusive. In mouse, NANOG protein levels are dynamic (Chambers et al., 2007), and it was recently proposed that NANOG stability is tied to phosphorylation (Moretto-Zita et al., 2010). Several studies suggested that mouse NANOG exists as a phosphoprotein (Li et al., 2011; Moretto-Zita et al., 2010; Yates and Chambers, 2005); however, its unique primary sequence and relatively low abundance make it difficult to purify and detect in a physiologically relevant context (i.e., without overexpression and in pluripotent cell types). As a result, there are currently no known phosphorylation sites for NANOG from human pluripotent stem cells and only a single site for endogenous mouse NANOG (Li et al., 2011). To address this gap, we applied high-resolution mass spectrometry to show that NANOG is heavily phosphorylated at proline-directed sites on its N terminus. To place these modifications in the context of signaling molecules, we developed the multiplexed assay for kinase specificity (MAKS), and found that ERK2 and CDK1 differentially phosphorylate NANOG in vitro.
Results
Discussion Together with OCT4 and SOX2, NANOG functions as a key transcriptional regulator in pluripotency. To date, almost no information is available regarding posttranslational regulation of NANOG. Our data show that NANOG is multiply phosphorylated in human ESCs. Nine of 11 sites were observed in a serine/threonine-rich patch from residues 50–79, suggesting that this region may serve as a regulatory cluster. In support of this notion, recent computational work identified a putative PEST sequence (residues 47–72) on NANOG that overlaps greatly with the hyperphosphorylated region (Figure 1D; Ramakrishna et al., 2011). Deletion of the PEST sequence altered NANOG ubiquitination and degradation (Ramakrishna et al., 2011). Further, mutational analysis in mouse for sites homologous to Ser52, Ser65, and Ser71 suggested that phosphorylation of these residues stabilizes NANOG (Moretto-Zita et al., 2010). All of these results point to a role for this region in regulating NANOG protein levels. Our data provide direct evidence that these regulatory sites are phosphorylated on human NANOG and in a physiological context. The function of phosphorylation at Ser22/23 and Ser258 is currently unclear; however, Ser258 is located in the CD2 activation domain and may modulate NANOG’s ability to activate transcription (Do et al., 2009). Further work will be needed to clarify the exact role of each modification identified here.