Stk40 deletion elevates c-JUN protein level and impairs mesoderm differentiation

Abstract

Mesoderm development is a finely tuned process initiated by the differentiation of pluripotent epiblast cells. Serine/threonine kinase 40 (STK40) controls the development of several mesoderm-derived cell types, its overexpression induces differentiation of mouse embryonic stem cells (mESCs) toward the extraembryonic endoderm, and Stk40 knockout (KO) results in multiple organ failure and is lethal at the perinatal stage in mice. However, molecular mechanisms underlying the physiological functions of STK40 in mesoderm differentiation remain elusive. Here, we report that Stk40 ablation impairs mesoderm differentiation both in vitro and in vivo. Mechanistically, STK40 interacts with both the E3 ubiquitin ligase mammalian constitutive photomorphogenesis protein 1 (COP1) and the transcriptional regulator proto-oncogene c-Jun (c-JUN), promoting c-JUN protein degradation. Consequently, Stk40 knockout leads to c-JUN protein accumulation, which, in turn, apparently suppresses WNT signaling activity and impairs the mesoderm differentiation process. Overall, this study reveals that STK40, together with COP1, represents a previously unknown regulatory axis that modulates the c-JUN protein level within an appropriate range during mesoderm differentiation from mESCs. Our findings provide critical insights into the molecular mechanisms regulating the c-JUN protein level and may have potential implications for managing cellular disorders arising from c-JUN dysfunction.

Introduction

Gastrulation is critical for early mammalian embryonic development, and one of its major tasks is to generate a mesodermal layer between the endoderm and ectoderm. Mesoderm development initiates with differentiation of pluripotent epiblast cells into the primitive streak, which then segregates into the mesoderm layer (1–3). This delicate process is coordinated by multiple key signaling pathways to ensure the correct formation of mesodermal tissues. Among them, bone morphogenesis protein (BMP)³ and WNT signaling pathways play profound roles to orchestrate mesoderm development (4–6). For instance, Wnt3-null mouse embryos lack the primitive streak and mesoderm formation (4). Moreover, WNT signaling is required for the generation of embryonic stem cell (ESC)-derived mesoderm, particularly for the expression of genes related to primitive streak formation and gastrulation in vivo (5). Furthermore, induced differentiation of pluripotent stem cells toward the mesodermal fate in monolayer cultures by manipulating WNT and BMP signaling pathways has been achieved in vitro (2, 3, 7), highlighting the decisive roles of the two pathways in mesoderm development. However, the molecular mechanisms regulating their activities in mesoderm formation remain incompletely understood.

The transcriptional regulator proto-oncogene c-Jun (c-JUN), encoded by Jun gene, acts as a subunit of the activating protein 1 (AP-1) family of transcription factors. By forming a homodimer or heterodimer with other members of the AP-1 family, c-JUN plays important roles in regulating cell proliferation as well as cell migration and oncogenic transformation (8–10). Forced expression of c-JUN in mouse ESCs (mESCs) activated endoderm lineage–related transcriptional factors (Gata4/Gata6) with the concomitant repression of Brachyury (T), a mesodermal marker gene (11), whereas the loss of c-JUN function in zebrafish inhibited ventral mesoderm induction (12), implying that the precisely balanced c-JUN level is crucial for mesoderm differentiation. A previous study showed that phosphorylated c-JUN (serine 63 and serine 73) could interact with the HMG-box transcription factor T-cell factor 4 (TCF4) to form protein complexes containing c-JUN, β-catenin, and TCF4 (13). A later study by Gan et al. (12) clearly showed that c-JUN is involved in the canonical WNT signaling, dependent on its phosphorylation. Moreover, multiple E3 ligases that add ubiquitin molecules on c-JUN have been unveiled, such as FBW7 (14), ITCH (15), and mammalian constitutive photomorphogenesis protein 1 (COP1) (16). COP1 is a RING-finger E3 ubiquitin ligase. It acts through two distinct regulatory mechanisms, either as an E3 ligase or an adaptor to recruit substrates to DET1-Cullin4 ubiquitin ligase complexes, mediating ubiquitination and degradation of target proteins, such as c-JUN, E26 transformation-specific (ETS) (17), and ETS variant (ETV) (18). c-JUN can bind COP1 directly through a conserved consensus sequence, VP(D/E), located at its C terminus (19). The mutation of the VP sequence into AA disrupts the association between c-JUN and COP1 (16, 19). Remarkably, Cop1 deficiency stimulates cell proliferation by means of elevating c-JUN protein levels in vivo during embryogenesis, and Cop1-hypomorphic mice are tumor-prone, in line with the inverse correlation between COP1 and c-JUN proteins in human prostate cancers (20). Therefore, the proper amount of c-JUN is essential for early embryo development, and the ubiquitin–proteasome system (UPS) acts as a potential pathway to reduce c-JUN protein levels.

Serine/threonine kinase 40 (Stk40) was originally identified as a direct target gene of OCT4, a pluripotency-associated transcription factor, in mESCs by our laboratory (21). Stk40-null mice die at the perinatal stage (22). Further studies unveiled that Stk40 deletion leads to disorders in adipogenesis (23), myogenesis (24), and erythropoiesis (25). Interestingly, all of these involved cell types are of mesodermal origin. Therefore, we speculated that Stk40 might participate in the control of early mesoderm development in mouse embryos. Recently, human STK40 was identified as a pseudokinase lacking the ATP-binding property (26). The same study also reported the interaction between STK40 and E3 ligase COP1 in human cells, although the functional consequence of the STK40–COP1 interaction remains unclear.

To address the question of whether STK40 plays a role in mesoderm development, we modified a previously published protocol for mesoderm induction from both wildtype (WT) and Stk40-knockout (KO) mESCs by activating BMP and WNT signaling sequentially (27) and discovered that the loss of Stk40 increased the steady-state level of c-JUN proteins and impeded mesoderm differentiation. Moreover, STK40 could facilitate COP1–c-JUN complex formation to regulate c-JUN protein levels and ensure proper mesoderm differentiation. Taken together, the current study uncovers an important role and regulatory mechanism of STK40 in the control of mesoderm differentiation from pluripotent cells.

Results

Stk40 deletion impedes mesoderm differentiation

To test the hypothesis that the deletion of Stk40 could impair mesoderm differentiation, we induced mESCs to differentiate toward the mesoderm lineage by modulating the activity of BMP4 and WNT signaling pathways sequentially, based on a previously published protocol (27). In brief, mESCs were cultured in N2B27 medium containing BMP4 (10 ng/ml) for 2 days followed by the addition of 1 μm CHIR99021 (CHIR) and 0.5% DMSO for 1 additional day (Fig. 1A). CHIR is used widely to activate WNT signaling by the inhibition of glycogen synthase kinase-3 (GSK3). To verify that ESCs were induced into the mesoderm lineage, expression levels of multiple marker genes in both WT and Stk40-null cells were examined at the indicated time points (Fig. 1, B–E). As expected, for pluripotency-associated genes, Nanog levels declined rapidly, whereas Oct4 levels decreased gradually during differentiation. Expression of Fgf5, an epiblast marker, increased dramatically at differentiation days 1 and 2 but dropped to a relatively lower level at day 3. Notably, the mesoderm marker gene T, together with Fgf8 and Cdx2, was substantially activated at differentiation day 3, suggesting that a mesoderm program was effectively induced (Fig. 1B). Moreover, the activation of WNT signaling was verified by elevated expression of Rspo3, an activator of the canonical WNT signaling pathway (28). Furthermore, immunofluorescence staining assays showed that the signal intensity of OCT4 and NANOG decreased along with the differentiation process (Fig. 1D), whereas the signal of T staining could only be detected at differentiation day 3 (Fig. 1E) rather than days 1 and 2 (data not show). Hence, the in vitro model of induced differentiation toward the mesoderm from ESCs was successfully established.

Figure 1.

Stk40 deletion impairs mesoderm differentiation from ESCs. A, schematic illustration of the in vitro mesoderm differentiation procedure. B, qRT-PCR analysis of mRNA levels of marker genes at indicated time points during mesoderm differentiation from WT and Stk40-KO mESCs. Data were normalized to the expression of Actb and are represented as -fold changes relative to those in undifferentiated ESCs. Error bars denote the means ± S.D. (n = 4 independent experiments); Student's t test: *, p < 0.05; **, p < 0.01; ***, p < 0.001. C, representative Western blot analysis of STK40 protein levels during mesoderm differentiation. β-actin was used as a loading control. Data are representative of three independent experiments. The molecular mass (kDa) is given at the right side of the blot. D, representative results of immunofluorescence staining of OCT4 (green) and NANOG (red) in WT and Stk40-KO cells at the undifferentiated stage or differentiation days 1, 2, and 3. DAPI was used to label the nuclei (blue). Scale bars represent 50 μm. E, representative results of immunofluorescence staining of T (red) in WT and Stk40-KO cells at differentiation day 3. One colony is presented in the upper panel, and the lower panel shows colonies at the lower magnification. DAPI was used to label the nuclei (blue). Scale bars represent 50 μm. Similar results were obtained in at least three independent experiments. F, the knockdown efficiency of a transduced Stk40 shRNA was evaluated by the Western blot analysis at differentiation day 3. β-actin was used as a loading control. The arrowhead indicates the specific signal of STK40. G, representative flow cytometric analysis for the percentage of T-positive cells at mesoderm differentiation day 3 from control shRNA (Ci)– and Stk40 shRNA (Stk40 i)–transduced GFP-Bry ESCs. Undifferentiated ESCs were used as a control. Similar results were obtained from four independent experiments. H, quantification of relative cell numbers during mesoderm differentiation from WT and Stk40-KO ESCs. Error bars denote the means ± S.D. (n = 3); Student's t test: *, p < 0.05. d, days.

Comparison of marker gene expression levels between WT and Stk40-KO cells revealed that the lack of Stk40 led to significant reductions of Rspo3 as well as T, Fgf8, and Cdx2 at differentiation day 3, without affecting expression profiles of Oct4, Nanog, and Fgf5, during ESC differentiation toward the mesoderm (Fig. 1, B and C). The finding suggested that STK40 might be required for the induction of the mesodermal program. Also, we found that mRNA and protein levels of STK40 increased moderately during the mesoderm differentiation process (Fig. 1, B and C), further supporting the notion that STK40 might contribute to mesoderm differentiation.

To further characterize the role of STK40 in mesoderm differentiation, immunofluorescence staining of T was conducted for cells at differentiation day 3. The signal intensity of T staining attenuated in Stk40-deleted cells (Fig. 1E). However, the expression of OCT4 and NANOG was not influenced upon Stk40 ablation (Fig. 1D), suggesting that STK40 contributes to mesoderm differentiation without disturbing the pluripotency exit of ESCs. To quantify the efficiency of mesoderm differentiation, we established an mESC line stably expressing an shRNA sequence specifically targeting Stk40 in the mESCs with the green fluorescence protein (GFP) targeted to the Brachyury locus (GFP-Bry ESCs) (29). The expression of the inserted GFP cassette is under the control of native Brachyury regulatory elements. Using the GFP-Bry ESC line, quantification of mesoderm induction becomes possible (30). We induced mesoderm differentiation from the GFP-Bry ESCs expressing either an Stk40 shRNA or control shRNA and analyzed the GFP-positive cell population at differentiation day 3 by flow cytometry, taking the percentage of GFP-positive cells as an index of mesoderm differentiation efficiencies. Knockdown of Stk40 was verified by Western blot analysis (Fig. 1F). About 72.9% GFP-positive cells were detected in the control group, whereas only 52.8% GFP-positive cells were found in cells expressing the Stk40 shRNA (Fig. 1G), supporting the proposal that STK40 is required for the proper differentiation of mESCs into the mesoderm lineage. In addition, we compared the cell number between WT and Stk40-null cells during mesoderm differentiation and found that Stk40 deletion led to a higher cell growth rate (Fig. 1H). These results indicate that the loss of Stk40 impairs mesoderm differentiation from ESCs with a concomitant increase in the cell growth rate.

To gain a global view for roles of STK40 in mesoderm differentiation, we performed RNA-Seq for both WT and Stk40-KO cells at day 3 of mesoderm differentiation. Altogether, 25 up-regulated and 85 down-regulated genes were detected in Stk40-null cells compared with WT cells (Fig. 2A). Gene ontology (GO) analyses revealed that the down-regulated genes were enriched for terms like somitogenesis, mesendoderm migration, and canonical WNT signaling pathway (Fig. 2B), consistent with impaired mesoderm differentiation observed in Stk40-KO cells. Specifically, genes down-regulated upon Stk40 ablation included an array of mesoderm-related genes, such as T and Mixl1 (31) (Fig. 2C). Moreover, the expression of Wnt3a, Wnt5a, and Rspo3, known to exert important roles in mesodermal development through the WNT signaling pathway, was also attenuated in Stk40-deleted cells as compared with WT cells (Fig. 2D), implying that the reduced WNT signaling activity might, at least partially, account for the impaired mesoderm differentiation in Stk40-KO cells.

Figure 2.

Stk40 deletion attenuates the WNT signaling activity during mesoderm differentiation from ESCs. A, the heat map of differentially expressed genes between WT and Stk40-KO cells (p value <0.05) at mesoderm differentiation day 3 with two biological replicates of both WT and Stk40-KO cells. B, GO analyses of down-regulated and up-regulated genes, respectively, in Stk40-KO cells from the data sets of RNA-Seq in A. The DAVID method was applied to GO analyses, and enrichment levels of selected GO terms (p value <0.05) are marked by −log₁₀ (p value). Missing values are shown as black boxes. C and D, the heat map of the mesoderm-related genes (C) and WNT signaling–related genes (D), which were down-regulated in Stk40-KO cells compared with WT cells at mesoderm differentiation day 3. -Fold change levels of selected genes are marked by log₂ -fold change (FC) (p value <0.05). TGF, transforming growth factor; MAPK, mitogen-activated protein kinase.

Stk40 deletion increases protein levels of c-JUN, repressing mesoderm differentiation

To explore how STK40 could be associated with the WNT signaling pathway, we employed mass spectrometric analysis to find proteins with distinct levels between the WT and Stk40-KO mouse embryonic fibroblasts (MEFs) (Fig. 3A). MEFs were chosen primarily due to their convenience in use as a cellular system. Interestingly, we found that AP-1 targets, such as DNMT1, CEBPβ, TGM2, FOSL2, APLS1, and JUNB (32), had higher protein amounts in Stk40-KO MEFs than in WT MEFs (Fig. 3B). AP-1 transcription factors, encompassing c-JUN, JUNB, JUND, c-FOS, and FRA-1, are known to control cellular proliferation, transformation, and death (8, 9, 33). Among AP-1 transcription factors, c-JUN plays critical roles in a broad range of developmental events in a dosage-dependent manner (33), and it was previously reported to participate in the modulation of canonical WNT signaling (12). Hence, we tested whether there was a difference in the c-JUN protein levels between WT and Stk40-KO MEFs. Indeed, the c-JUN protein level was obviously higher in Stk40-KO MEFs than in WT cells (Fig. 3C). To learn whether the c-JUN protein level would also be higher in Stk40-KO cells during mesoderm differentiation of ESCs, we measured both phosphorylated (Ser-63/Ser-73) and total c-JUN protein levels at 3, 6, 12, and 24 h post-CHIR addition after differentiation for 2 days. Consistently, higher total c-JUN protein levels were detected in Stk40-KO cells than in WT cells at all time points tested (Fig. 3D), although Stk40 KO did not alter the transcript level of c-Jun (Fig. 3F). In addition, the levels of phosphorylated c-JUN at serine residues 63 and 73 were also higher in Stk40-KO cells than in WT cells at most time points examined. However, there was not a discerned difference in protein levels of phosphorylated c-JUN N-terminal kinases (JNKs) (Fig. 3D), which have been reported to phosphorylate c-JUN within its N terminus (34). Despite simultaneous increases in both total and phosphorylated c-JUN protein amounts, ratios of phosphorylated c-JUN/total c-JUN proteins for their relative levels at different time points decreased in Stk40-KO cells for both Ser-63 and Ser-73 phosphorylation as compared with that in WT cells post-CHIR treatment (Fig. 3E). These results indicate that STK40 could modulate the level of c-JUN proteins at a posttranscriptional level.

Figure 3.

Stk40 knockout increases c-JUN protein levels in MEFs and during mesoderm differentiation. A, the heat map showing differences in protein levels between WT and Stk40-KO MEFs (p value <0.05) identified by mass spectrometric analysis, with three biological replicates for each cell type. B, a volcano plot representing the increased (blue) and decreased (red) proteins in Stk40-KO MEFs as compared with WT MEFs, based on data sets from A. AP-1 family–related proteins are marked with black circles in the plot. C, a representative Western blot analysis for protein levels of STK40 and c-JUN in WT and Stk40-KO MEFs. β-actin was used as a loading control. D, at the indicated time points of CHIR treatment after mesoderm differentiation for 2 days from WT and Stk40-KO ESCs, protein levels of STK40, JNK (p-Thr-183/Tyr-185), c-JUN, c-JUN (p-Ser-63), and c-JUN (p-Ser-73) were analyzed by Western blotting. Results for analyses of c-JUN and phosphorylated c-JUN levels from two independent experiments are shown. E, the gray values of phosphorylated c-JUN and total c-JUN in D were quantified using ImageJ software, and the relative protein levels were calculated by comparing the gray values from the indicated time points with that from WT cells at 3 h after CHIR induction. Then, ratios of phosphorylated c-JUN/total c-JUN proteins were obtained by comparing the values of relative protein levels. The mean values of ratios (p-Ser-63 c-JUN or p-Ser-73 c-JUN)/total c-JUN) from the two independent experiments were calculated and are presented in the plots. F, relative mRNA levels of Stk40 and Jun (c-JUN) were measured by qRT-PCR. Data were normalized to the level of Actb and are represented as -fold changes (FC) relative to those in undifferentiated ESCs. Error bars denote the means ± S.D. (n = 3); Student's t test: ***, p < 0.001. G, TOP/FOPFlash assays showing that overexpression of the WT or SS-AA mutant c-Jun attenuates the WNT signaling activity in 293FT cells. The TOPFlash or FOPFlash plasmid plus c-Jun, c-Jun(SS-AA), or an empty vector was transfected into 293FT cells, and cells were treated with CHIR at 1 μm for 6 h prior to sample collection. Luciferase activity ratios were determined by (TOP firefly/Renilla)/(FOP firefly/Renilla). H, representative Western blot analysis of Dox-induced c-Jun overexpression at mesoderm differentiation day 3 from GFP-Bry ESCs. β-actin was used as a loading control. I, representative flow cytometric analysis of GFP-positive cells at differentiation day 3 with or without c-Jun overexpression. Similar results were obtained in two independent experiments. NC, an empty vector; SSC, side scatter.

Given that WNT signaling was enriched in genes down-regulated by Stk40 KO (Fig. 2B) and that c-JUN was reported to participate in the regulation of WNT signaling (12), we conducted the TOP/FOPFlash luciferase reporter assay to determine the role of c-JUN in controlling WNT signaling activity and found that overexpression of c-JUN or its phospho-dead (SS-AA; the mutation from serine to alanine at residues Ser-63/Ser-73) mutant in 293FT cells reduced WNT signaling activity in the presence of CHIR (Fig. 3G), suggesting that ectopic expression of c-Jun could suppress the WNT signaling activity independently of c-JUN phosphorylation at Ser-63/Ser-73.

The aforementioned findings implied that STK40 might modulate WNT signaling activities through controlling c-JUN protein levels during mesoderm differentiation. To address this issue, we established a doxycycline (Dox)-induced c-Jun overexpression system in the GFP-Bry ESCs. Addition of Dox induced overexpression of c-JUN in c-Jun sequence–containing virus-infected cells but not in empty vector virus–infected cells (Fig. 3H). At differentiation day 3, the proportion of GFP-positive cells was lower in cells overexpressing exogenous c-Jun as compared with that in cells without exogenous c-Jun (Fig. 3I). Therefore, our results indicate that an abnormally high level of c-JUN has a suppressive effect on mesoderm differentiation from ESCs.

STK40 interacts with c-JUN and facilitates the formation of c-JUN and COP1 protein complexes

It is known that c-JUN proteins can be degraded through the UPS by several E3 ligases, such as COP1 (35), MEKK1 (36), FBW7 (37), and ITCH (38). Of note, STK40 was recently reported to bind to COP1 directly, possibly serving as a COP1 adaptor to recruit substrates (26). Hence, we anticipated that STK40 might control the c-JUN protein level by modulating the interaction between COP1 and c-JUN to promote c-JUN degradation. To test this hypothesis, we first validated the interaction between STK40 and COP1 as well as between STK40 and c-JUN by GST pulldown and coimmunoprecipitation (co-IP) assays, respectively. GST-STK40 specifically bound to His-c-JUN and His-COP1 in vitro, respectively (Fig. 4A). As a negative control, GST alone was not able to associate with either His-c-JUN or His-COP1. We next examined which region(s) of STK40 was responsible for its interaction with c-JUN or COP1. To this end, we generated FLAG-tagged expression vectors containing the coding sequences for the three truncated forms of STK40: the lack of the N terminus (ΔN), the lack of C terminus (ΔC), and the kinase homolog domain (PK) (Fig. 4B). Then, HA-tagged c-Jun expression vectors were cotransfected into 293FT cells in combination with FLAG-tagged WT or truncated Stk40 expression vectors. The lack of the C terminus on STK40 abrogated its interaction with c-JUN, whereas the absence of the N terminus augmented the association between STK40 and c-JUN (Fig. 4C). This observation suggested that the C terminus of STK40 was required for its interaction with c-JUN, whereas the N-terminal region might have a negative impact on the interaction between STK40 and c-JUN. Similarly, the interaction between STK40 and COP1 was abrogated by the deletion of the C-terminal region. However, their interaction was not affected by the absence of the N-terminal region of STK40 (Fig. 4D).

Figure 4.

STK40 facilitates the association between c-JUN and COP1. A, representative Western blot analysis of GST-STK40 pulldown assays. GST or GST-STK40 fusion proteins were incubated with His-c-JUN or His-COP1 fusion proteins and precipitated by GSH-Sepharose 4B beads. Precipitated His-c-JUN or His-COP1 proteins were visualized by Western blotting. Similar results were obtained from at least three independent experiments. B, a schematic representation of the modular architecture of WT and truncated STK40. C and D, representative Western blot analysis of interactions between WT or truncated STK40 and c-JUN (C) or COP1 (D). WT or truncated Stk40 expression vectors were transfected into 293FT cells in combination with HA-c-Jun or HA-Cop1 plasmids. Whole-cell lysates were immunoprecipitated with anti-FLAG M2 beads and analyzed by Western blotting with anti-HA or anti-FLAG antibodies (right panel). The expression of FLAG-tagged STK40, HA-tagged c-JUN, and HA-tagged COP1 was verified by immunoblotting with anti-FLAG and anti-HA antibodies, respectively (left panel). Data are representative of three independent experiments. E, phosphorylation sites at the N-terminal domain of STK40 were predicted by DISPHOS (39) (www.dabi.temple.edu/disphos).⁴ F, the STK40 mutants of phospho-dead (S6A) and phosphomimic (S6E) impact the interaction between STK40 and c-JUN. WT or mutated Stk40 expression vectors were cotransfected with c-Jun constructs into 293FT cells. Whole-cell lysates were immunoprecipitated with anti-FLAG M2 beads and analyzed by Western blotting with anti-HA or anti-FLAG antibodies (right panel). S6A denotes the mutation from serine (S) to alanine (A), and S6E denotes the mutation from serine to glutamic acid (E) at serine residue 6. Data are representative of three independent experiments. G, STK40 could mediate the interaction between c-JUN and COP1. The FLAG–COP1 protein complexes were immunoprecipitated from 293FT cells with anti-FLAG M2 beads and analyzed by Western blotting with the antibodies indicated. β-actin was used as a loading control. AA denotes the mutation from ²³⁵VP²³⁶ to AA. Data are representative of three independent experiments.

Posttranslational modifications could regulate protein interactions, and phosphorylation is perhaps one of the most studied modifications with respect to modulating binding affinity of proteins. To learn how the N-terminal region of STK40 affected its interaction with c-JUN, we examined phosphorylation sites within this region through prediction using the web tool DISPHOS (www.dabi.temple.edu/disphos) (39).⁴ Among the four predicted residues, the serine at residue 6 (Ser-6) had the highest scores (Fig. 4E). Hence, we investigated whether STK40 mutants of phospho-dead (S6A, the mutation from serine to alanine) and phosphomimic (S6E, the mutation from serine to glutamic acid) could impact the interaction between STK40 and c-JUN by co-IP assays. We found that the S6A STK40 mutant displayed an obviously stronger association with c-JUN than did WT STK40. In contrast, the S6E STK40 mutant had a weaker interaction with c-JUN compared with WT STK40 (Fig. 4F). The result suggests that the phosphorylation of the N-terminal Ser-6 could interfere with the interaction between STK40 and c-JUN.

To address the question of whether STK40 could modulate the formation of COP1 and c-JUN complexes to control the c-JUN protein level, co-IP assays were conducted to determine the binding intensity between COP1 and c-JUN in the presence and absence of exogenous STK40, respectively. Overexpression of HA-tagged Stk40 enhanced the interaction between COP1 and c-JUN in 293FT cells (Fig. 4G, right panel, lane 4 versus lane 3). COP1 was previously shown to bind the VP motif of c-JUN, and a mutation of VP to AA disrupted the direct interaction between COP1 and c-JUN (16, 35). Consistently, our co-IP assay results showed that c-JUN(AA) could not interact with COP1 (Fig. 4G, right panel, lane 5). Interestingly, in the presence of exogenous STK40, a relatively weak interaction between COP1 and c-JUN(AA) was detected (Fig. 4G, right panel, lane 6). Therefore, we propose that STK40 could enhance the interaction between COP1 and c-JUN.

STK40 promotes c-JUN protein degradation partially through COP1

Considering that Stk40 ablation led to elevated levels of c-JUN proteins and that STK40 facilitated the interaction between COP1 and c-JUN, we predicted that STK40 might accelerate c-JUN protein degradation. To test this idea, we blocked protein synthesis by cycloheximide (CHX) treatment to analyze c-JUN protein stability in WT and Stk40-KO MEFs. As expected, the protein stability of c-JUN was higher in Stk40-KO MEFs than in WT MEFs (Fig. 5A). Consistently, the reintroduction of Stk40 into Stk40-KO MEFs reduced steady-state levels of c-JUN protein in a Stk40 dosage–dependent manner (Fig. 5B). In contrast, the stability and amount of COP1 proteins were not affected by Stk40 expression levels (Fig. 5, A and B). These results suggest that STK40 participates in the control of c-JUN protein stability.

Figure 5.

STK40 promotes c-JUN protein turnover partially depending on COP1. A, protein stability of COP1, STK40, and c-JUN was evaluated by Western blot analysis. Prior to sample collection, WT and Stk40-KO MEFs were treated with either 50 μm CHX for the indicated length of time or DMSO for 180 min. The total cell lysate was extracted and subjected to Western blot analysis. The gray density of the c-JUN blot was measured by ImageJ software and is indicated. B, representative Western blot analysis of c-JUN, STK40, and COP1 protein levels in Stk40-KO MEFs with or without overexpression of Stk40 at different dosages after viral infection for 48 h. C, STK40-mediated attenuation of c-JUN protein level is partially dependent on COP1. Viral shRNAs of control (Ci) or Cop1 (a mixture of Cop1 i-1 and Cop1 i-2 shRNAs) together with viral GFP or Stk40-GFP were delivered in WT and Stk40-KO MEFs, respectively. Protein levels of GFP, STK40-GFP, COP1, and c-JUN were analyzed by Western blotting at 48 h postinfection. β-actin was used as a loading control. D, viral shRNAs of control (Ci), Cop1 i-1, or Cop1 i-2 were delivered in GFP-Bry ESCs. The expression levels of COP1 and c-JUN were analyzed by Western blotting. E, a representative result for percentages of GFP-positive cells from control and Cop1-knockdown cells analyzed by flow cytometry at day 3 of mesoderm differentiation from GFP-Bry ESCs. SSC, side scatter.

To gain experimental evidence that COP1 is involved in STK40-mediated control of c-JUN protein levels, we silenced the expression of Cop1 in MEFs using an shRNA interference approach. Western blotting results showed that knockdown of Cop1 increased c-JUN protein levels in both WT and Stk40-KO MEFs and that the c-JUN protein level was highest in Stk40-KO cells expressing the Cop1 shRNA, although there was a lower -fold change in c-JUN protein levels between Cop1-knockdown and control MEFs in the absence of Stk40 (Fig. 5C, lane 6 versus lane 2) than that in the presence of Stk40 (Fig. 5C, lane 5 versus lane 1). Of note, overexpression of Stk40-GFP fusion proteins decreased the amount of c-JUN proteins in Stk40-KO MEFs (Fig. 5C, lane 4 versus lane 2). However, the rescue effect of Stk40 overexpression was attenuated when Cop1 was silenced (Fig. 5C, lane 8 versus lane 6 and lane 8 versus lane 4), suggesting that STK40-mediated control of c-JUN protein level is, at least partially, dependent on COP1.

To determine the role of COP1 in the control of c-JUN protein amounts and mesoderm differentiation from ESCs, we silenced Cop1 expression by two sets of shRNAs specifically targeting Cop1, respectively, in the GFP-Bry ESCs, which were induced to mesoderm differentiation as described in Fig. 1A. At day 3 of differentiation, GFP-positive cells were analyzed by flow cytometry. The high efficiency of Cop1 knockdown was validated by Western blot analysis. Cells expressing Cop1 shRNAs had markedly higher c-JUN protein levels than control cells (Fig. 5D). Compared with cells expressing the control shRNA, the percentage of GFP-positive cells decreased dramatically in cells expressing Cop1 shRNAs (Fig. 5E). Therefore, COP1 plays an important role in tightly controlling c-JUN protein levels during mesoderm differentiation.

Stk40 deletion impairs mesoderm differentiation and leads to c-JUN protein accumulation in vivo

To determine whether STK40 participates in mesoderm development in vivo, we conducted whole-mount immunofluorescence staining of mouse embryos at the gastrulation stage (E7.0) when T-positive cells migrate from the posterior to anterior region to establish mesodermal and endodermal layers (30, 40). Antibodies against T and SOX2 were employed for the whole-mount staining to mark the mesodermal and ectodermal cells, respectively. Altogether, 10 WT and 17 Stk40-null embryos were subjected to immunostaining assays. Confocal imaging analyses demonstrated that T-positive cells distributed similarly in WT and Stk40 KO embryos, but the signal intensity of T was drastically reduced in Stk40-KO embryos. Statistically, eight of 10 WT and five of 17 Stk40-KO embryos displayed normal T signal intensities, whereas two of 10 WT and 12 of 17 Stk40-null embryos showed weak T signals. These observations suggest that STK40 is required for the proper differentiation of pluripotent cells into T-positive cells at the gastrulation stage. As a control, Stk40 ablation did not affect signal intensities of SOX2 staining significantly (Fig. 6A).

Figure 6.

Stk40 deletion impairs mesoderm development and enhances c-JUN abundance in vivo. A, representative whole-mount immunofluorescence staining images of T (green) and SOX2 (red) in WT and Stk40-KO embryos at E7.0. A normal T–staining image in a WT embryo and a weak T–staining image in a Stk40-KO embryo are shown. DAPI was used to label the nuclei (blue). Scale bars represent 250 μm. B, protein levels of COP1, STK40, and c-JUN in the limb, liver, and head of WT and Stk40-KO embryos at E13.5 were analyzed by Western blotting. The arrowhead indicates the specific signal of STK40. β-actin was used as a loading control. Samples were collected from three WT (WT-1, WT-2, and WT-3) and three Stk40-KO embryos (KO-1, KO-2, and KO-3), respectively.

Due to the ubiquitous function of c-JUN in many cell types, we posited that STK40 might modulate c-JUN protein levels in a variety of tissues. We compared c-JUN protein levels in the limb, liver, and head between WT and Stk40-KO embryos at E13.5 and found that c-JUN protein levels were higher in the limb, liver, and head from Stk40-KO embryos than in WT counterparts, whereas there was no difference in the level of COP1 proteins in the limb, liver, and head between WT and Stk40-KO embryos (Fig. 6B). Therefore, STK40 might control the c-JUN protein level in multiple organs/tissues of mouse embryos.

Discussion

c-JUN is expressed in a wide range of cell types in developmental processes and is essential for normal mouse development (41). Moreover, high c-JUN abundance shows a tight correlation with the tumorigenesis of several cell lines (9, 42). Thus, it is of significance to address the question of how c-JUN levels are modulated. Here, we report that STK40 could be an important new regulator of c-JUN protein amount. Additionally, our study reveals that STK40 might control c-JUN protein turnover through modulating the formation of COP1 and c-JUN complexes. Functionally, we show that this STK40–COP1–c-JUN axis plays a role in mesoderm differentiation (Fig. 7).

Figure 7.

The schematic illustration of how STK40 contributes to mesoderm induction from mESCs. STK40 facilitates the association between COP1 and c-JUN, likely leading to proteasomal degradation of c-JUN, and the proper amount of c-JUN ensures mesoderm differentiation. In the absence of Stk40, mesoderm induction from mESCs is impaired at least partially due to c-JUN protein accumulation and WNT signaling inhibition. Ub, ubiquitin.

Our previous studies showed that Stk40 ablation in mice leads to developmental disorders in multiple mesoderm-derived cell types (22–25). Thus, we speculated that STK40 might be involved in mesoderm development. Utilizing an in vitro model of mesoderm induction from mESCs, we explored the function of Stk40 in mesoderm differentiation at both the molecular and cellular levels. By analyzing the RNA-Seq data from WT and Stk40-KO cells at mesoderm differentiation day 3, we found that expression levels of some WNT family members were lower in Stk40-ablated cells than in WT cells, and some of these WNT family members are known to play critical roles in mesoderm development (4, 6, 43). Therefore, attenuated WNT signaling might be responsible for the impaired mesoderm differentiation caused by Stk40 KO. c-JUN has previously been reported to bridge TCF and Dvl, and its N-terminal phosphorylation is likely required in the canonical WNT signaling pathway. Functionally, c-JUN contributes to the induction of ventral mesoderm in zebrafish (12). In this study, we revealed that both phosphorylated and total c-JUN protein levels were elevated upon Stk40 ablation. However, ratios of the phosphorylated (Ser-63 or Ser-73)/total c-JUN levels were lower in Stk40 KO cells than in WT cells during mesoderm differentiation from ESCs, providing a possible explanation for the reduced WNT signaling activity in Stk40-deleted cells. Moreover, overexpression of c-Jun suppressed WNT signaling activity and decreased the percentage of T-positive mesoderm-like cell populations at day 3 of ESC differentiation. In line with our finding, Liu et al. (11) reported that overexpression of c-Jun in mESCs leads to the activation of certain endodermal marker genes with the concomitant repression of T. Collectively, we propose that the appropriate level of c-JUN, especially its phosphorylation state, participates in the control of mesoderm differentiation, probably through modulating the WNT signaling pathway.

COP1 was previously shown to recruit c-JUN to a multisubunit ubiquitin ligase complex containing DNA damage–binding protein-1 (DDB1), Cullin 4A, and regulator of Cullins-1 (ROC1) rather than acting as an E3 ligase to catalyze c-JUN ubiquitination directly (16). Here, we unveiled that overexpression of Stk40 enhanced the interaction between COP1 and c-JUN. The mutation of the VP motif of c-JUN abrogated the interaction between COP1 and c-JUN(AA). However, in the presence of exogenous STK40, the interaction between COP1 and c-JUN(AA) could be detected. Subsequently, c-JUN might be ubiquitinated by the multisubunit ubiquitin ligase complex (DET1–DDB1–Cullin 4A–RBX1) and degraded through the UPS. Moreover, we found one component of the ubiquitin ligase complex, DET1, in proteins potentially interacting with STK40 from a public database (BioPlex 2.0 data set) (44), supporting the possibility that STK40 might recruit the ubiquitin ligase complex to down-regulate c-JUN protein levels. Nevertheless, we could not exclude the possibility that COP1 might serve as an E3 ligase to add ubiquitin molecules on c-JUN directly in our context. More studies should be conducted to address the question of how STK40 precisely modulates c-JUN protein stability through COP1.

Taken together, our study suggests that STK40 promotes COP1 and c-JUN to form complexes, hence maintaining c-JUN proteins at an appropriate level and taking part in the regulation of mesoderm development (Fig. 7). The molecular mechanism described here not only extends our understanding of how COP1 regulates c-JUN protein stability but also uncovers STK40 and COP1 as new regulators for mesoderm differentiation. Additionally, given the well-established roles of c-JUN in cancers, our finding may provide potential therapeutic targets to treat diseases.

Experimental procedures

Animals

Animals were raised under the condition described previously and performed according to the guidelines approved by the Shanghai Jiao Tong University School of Medicine (22). The genotype of WT (Stk40^+/+) and Stk40-KO (Stk40^−/−) mice was determined as described previously (22).

Cell culture and differentiation

mESCs were derived from WT and Stk40-KO blastocysts (22) as described previously (45). The ESCs were maintained on inactivated MEF feeders in N2B27 medium with 2 mm GlutaMAX (Gibco), 0.1 mm nonessential amino acids, 0.1 mm β-mercaptoethanol (Sigma), 1000 units/ml leukemia inhibitory factor (Millipore), 3 μm CHIR99021, and 1 μm PD0325901 (STEMCELL Technologies) (46). Cell colonies were digested into single cells for regular passaging using Accutase (STEMCELL Technologies). For induced mesoderm differentiation from mESCs, cells were passaged onto gelatin-coated cell culture plates to remove feeder cells prior to the initiation of differentiation. After one passage, ESCs were dissociated with Accutase, and 1.5 × 10⁵ cells were seeded in 1 well of 6-well plates precoated with gelatin and maintained in N2B27 medium with addition of 1% KnockOut Serum Replacement (Gibco), 0.1% BSA (Gibco), and 10 ng/ml BMP4 (R&D Systems) for 2 days. Then, the medium was replaced by a DMEM-based medium with 15% KnockOut Serum Replacement, 1 μm CHIR99021 (STEMCELL Technologies), and 0.5% DMSO (Sigma) for 1 additional day (27).

MEFs were generated from E13.5 mouse embryos and cultured in medium consisting of DMEM (Gibco) supplemented with 10% fetal bovine serum (Gibco). Plat-E (Cell Biolabs) and 293FT (Invitrogen) cell lines were cultured in 293FT medium consisting of DMEM (Gibco) supplemented with 10% fetal bovine serum (Gibco), 2 mm GlutaMAX (Gibco), 1 mm pyruvate sodium (Gibco), and 0.1 mm nonessential amino acid (Gibco).

Virus preparation

For Stk40 and Cop1 knockdown assays, the DNA encoding sequences of shRNA specific to Stk40 and Cop1 were cloned into the pLKO.1 plasmid, respectively. Sense sequences for Stk40, Cop1, and control RNAi were as follows: Stk40 shRNA-1, 5′-GGACCCATCGGATAACTAT-3′; Stk40 shRNA-2, 5′-TGCATACCGAGTACTCTCT-3′; Cop1 shRNA-1, 5′-CCTTGGTATAACAGCACATTA-3′; Cop1 shRNA-2, 5′-GACAAATGGGCATGGCTAGAA-3′; control shRNA, 5′-GTGCGCTGCTGGTGCCAAC-3′.

For overexpression assays, cDNA sequences of Stk40 or Stk40-GFP were inserted into pMXs plasmid. For inducible overexpression assays, the cDNA sequence of c-Jun was inserted into the pLVX-Tight-puro vector, and the Tet-On advanced inducible gene expression system was established according to the manufacturer's user manual (Clontech). Viral packaging and transduction were performed as described previously (24).

RNA extraction, cDNA synthesis, and quantitative real-time PCR (qRT-PCR)

Total RNA was extracted using TRIzol reagent (Invitrogen) and reverse transcribed into cDNA utilizing oligo(dT)₁₈ and ReverTra Ace reverse transcriptase (Toyobo). qRT-PCR was performed using the ABI PRISM 7900 Fast Real-Time PCR system (Applied Biosystems) and SYBR Premix Ex Taq (Takara). The primer sequences for qRT-PCR are provided in Table S1.

RNA-Seq

Total RNA was extracted from WT and Stk40-0KO cells collected at mesoderm differentiation day 3 using TRIzol reagent. Sequencing libraries were prepared according to Illumina's instructions. Paired-end RNA-Seq of 2 × 150-bp reads were sequenced on the Illumina HiSeq X Ten. Salmon software was used to calculate the samples' transcripts per million and raw counts (47). DESeq2 (48) was then used to identify differentially expressed genes with the following setting: adjusted p value <0.05. RNA-Seq raw data were submitted to GEO (accession number GSE125877), and processed data are provided in Table S2.

Protein extraction and Western blot analysis

Total protein extracted from cells or mouse tissues was prepared with radioimmune precipitation assay lysis buffer (1 mm EDTA, 1% Nonidet P-40, 50 mm Tris (pH 7.5), 150 mm NaCl, and 10% glycerol), and protein concentrations were determined using the Pierce BCA Protein Assay kit (Thermo Fisher) following the manufacturer's instructions. Proteins were separated by SDS-PAGE and transferred to nitrocellulose membranes (GE Healthcare). Membranes were incubated with specific primary antibodies, and the antibody–protein complexes were visualized by horseradish peroxidase–conjugated secondary antibodies (Jackson ImmunoResearch Laboratories) and Pierce ECL Western Blotting Substrate (Thermo Fisher). Primary antibodies used for Western blotting are provided in Table S3.

Mass spectrometry

The mass spectrometric analysis was performed as described previously (49).

Digestion and TMT labeling

WT and Stk40-KO MEF samples were lysed using urea lysis buffer (8 m urea, 75 mm NaCl, 50 mm Tris (pH 8.2), 1% (v/v) EDTA-free protease inhibitor, 1 mm NaF, 1 mm β-glycerophosphate, 1 mm sodium orthovanadate, 10 mm sodium pyrophosphate, and 1 mm phenylmethylsulfonyl fluoride), measured by the Bradford assay (50), reduced with 5 mm DTT (56 °C, 25 min), alkylated in 14 mm iodoacetamide for 30 min in the dark, and quenched by DTT. The protein mixtures were diluted in 50 mm Tris-HCl (pH 8.8), digested by Lys-C overnight at 25 °C, desalted using an OASIS HLB 1-ml Vac cartridge (Waters), and subjected to TMT six-plex labeling. After TMT labeling, all samples were combined and lyophilized.

Strong cation-exchange fractionation

The peptide mixture was fractionated using a cation ion-exchange column (2.1-mm internal diameter × 20 cm packed with Poros 10 S, Dionex, Sunnyvale, CA) with an UltiMate® 3000 HPLC system at a flow rate of 200 μl/min using buffer A (7 mm KH₂PO₄ (pH 2.65) and 30% ACN) and buffer B (7 mm KH₂PO₄, 350 mm KCl (pH 2.65), and 30% ACN) in the following gradient: 0- 8% B in 0.1 min, 8–20% B for 15.9 min, 20–40% B for 20 min, 40–100% B for 1 min, 100% B for 5 min, 100–0% B in 1 min, 0% B for 25 min. Fractions were collected every 1.5 min.

LC-MS³ analysis

29 fractions were separated by a reverse-phase microcapillary column (0.075 × 150 mm, Acclaim® PepMap100 C₁₈ column, 3 μm, 100 Å; Dionex) using 2% ACN and 0.5% acetic acid (buffer A) and 80% ACN and 0.5% acetic acid (buffer B) under a 222-min gradient (3% buffer B for 10 min, 3–5% buffer B for 3 min, 5–20% buffer B for 167 min, 20–36% buffer B for 15 min, 36–100% buffer B for 1 min, 100% buffer B for 7 min, 100–3% buffer B for 1 min, 3% buffer B for 18 min) and analyzed using an LTQ Orbitrap Velos (Thermo Finnigan, San Jose, CA) with a higher-energy collisional dissociation MS³ method.

Protein identification and quantification

The raw files were processed with MaxQuant software (version 1.2.2.5) using the UniProt mouse protein database (55,269 sequences). A common contaminants database was also included for quality control. The reverse strategy was used to estimate the false discovery rate. Except for TMT labels, carbamidomethyl (Cys) was set as fixed modification. Variable modifications were oxidation (Met) and acetyl (protein N terminus). The site, peptide, and protein false discovery rates were all set to 0.01. Protein quantification was calculated by combining MaxQuant identification results with a local modified Libra algorithm (49). For the identification of differentiated expressed proteins among groups, the cutoffs of -fold change and p value (Student's t test) were set to 1.5 and 0.05, respectively.

The MS proteomics data have been deposited to the ProteomeXchange Consortium (www.proteomexchange.org) via the PRIDE partner repository with the data set identifier PXD012579. The analyzed data are provided in Table S4.

Immunoprecipitation and GST pulldown assays

The full-length coding sequence (CDS) of Stk40, c-Jun, and Cop1 were cloned into the pcDNA 3.0 vector, respectively. The truncated forms and mutants of Stk40 and mutated c-Jun were also cloned into the pcDNA 3.0 vector, respectively. 293FT cells were seeded at 2 × 10⁶ cells in 6-cm dishes on day 1, and 2 μg of expression vectors each or an empty vector were transfected with Lipofectamine 2000 reagent according to the manufacturer's instructions (Thermo Fisher Scientific) on day 2. After 24 h, cells were washed in the culture dish with cold PBS three times. Whole-cell lysates were prepared in co-IP buffer (50 mm Tris-HCl (pH 7.4) with 150 mm NaCl, 1 mm EDTA, 1% Triton X-100, and protease inhibitor mixture (Selleck)) and incubated with anti-FLAG M2 beads (Sigma) overnight at 4 °C. Then, beads were washed three times and eluted by 2× loading buffer for 10 min at 95 °C. For GST pulldown assays, the full-length Stk40 CDS was cloned into the pGEX-4T-1 plasmid, and the full-length c-Jun and Cop1 CDSs were cloned into the pET-30a(+) vector, respectively. GST and His fusion proteins were expressed in the BL21 strain and purified according to the manufacturers' instructions (GE Healthcare and Novagen). GST pulldown assays were performed as described previously (21). The samples from immunoprecipitation and GST pulldown were analyzed by Western blotting.

Flow cytometric analysis

GFP-Bry mESCs were infected with lentiviral particles that express shRNA. After puromycin selection (1 μg/ml), the remaining cells were subjected to induced differentiation toward mesoderm, and both GFP-positive and -negative cells were analyzed by flow cytometry.

Immunofluorescence staining

ESCs and differentiated cells were fixed with 4% paraformaldehyde at room temperature for 10 min, and immunostaining was performed as described previously (24). The mouse embryos were collected at E7.0, and whole-mount immunostaining was performed according to the iDISCO protocol (51). Images were captured using a confocal microscope (Leica TCS SP8). Primary antibodies used for immunostaining are provided in Table S3.

Protein stability assays with CHX

CHX (Sigma-Aldrich) was dissolved in DMSO at a concentration of 100 mm. Cells were cultured with CHX at a concentration of 50 μm for the indicated time. Then, cells were collected and subjected to Western blotting to visualize protein levels.

TOP/FOPFlash assays

8×TOPFlash or 8×FOPFlash and pRL-TK vectors (Promega) in combination with or without the WT or mutant (SS-AA) c-Jun expression vector were transfected into 293FT cells with Lipofectamine 2000 reagent according to the manufacturer's instructions (Thermo Fisher Scientific). pRL-TK serves as a control to normalize the transfection efficiency. Twenty-four hours later, cells were treated with CHIR99021 (1 μm) for an additional 6 h and then lysed for luciferase assays. Luciferase activities were evaluated using a Dual-Luciferase Assay kit (Promega) according to the manufacturer's instructions. TOP/FOPFlash ratios are represented as mean ± S.D. (n = 3).

Statistical analysis

Data are presented as the mean ± S.D. from at least three independent experiments or samples. Statistical significance was analyzed by two-tailed Student's t test and is shown as follows: *, p < 0.05; **, p < 0.01; ***, p < 0.001.

Author contributions

J. H., B. L., and Y. J. conceptualization; J. H. resources; J. H. and Z. F. data curation; J. H., Z. F., M. S., Y. G., and X. G. software; J. H. formal analysis; J. H., H. Y., and B. L. investigation; J. H., S. L., F. X., L. G., F. T., J. G., H. Y., Y. G., X. G., and B. L. methodology; J. H., B. L., and Y. J. writing-original draft; J. H., S. L., X. S., L. W., F. X., and Y. J. project administration; J. H., B. L., and Y. J. writing-review and editing; H. Y. and Y. J. funding acquisition; Y. J. supervision.