Lysosomes are degradation and signaling organelles that adapt their biogenesis to meet many different cellular demands; however, it is unknown how lysosomes change their numbers for cell division. Here, we report that the cyclin-dependent kinases CDK4/6 regulate lysosome biogenesis during the cell cycle. Chemical or genetic inactivation of CDK4/6 increases lysosomal numbers by activating the lysosome and autophagy transcription factors TFEB and TFE3. CDK4/6 interact with and phosphorylate TFEB/TFE3 in the nucleus, thereby inactivating them by promoting their shuttling to the cytoplasm. During the cell cycle, lysosome numbers increase in S and G2/M phases when cyclin D turnover diminishes CDK4/6 activity. These findings not only uncover the molecular events that direct the nuclear export of TFEB/TFE3, but also suggest a mechanism that controls lysosome biogenesis in the cell cycle. CDK4/6 inhibitors promote autophagy and lysosome-dependent degradation, which has important implications for the therapy of cancer and lysosome-related disorders.
Lysosomes are the major digestive organelles that degrade both extra- and intracellular materials generated by endocytosis, phagocytosis, and autophagy; thus, they play important roles in many physiological processes such as the immune response, plasma membrane repair, bone resorption, and cell death (Luzio et al., 2007; Saftig and Klumperman, 2009; Xu and Ren, 2015; Yang and Wang, 2017). Lysosomes also serve as signaling hubs that sense cellular energy and amino acid levels and mediate signal transduction (Efeyan et al., 2015; Ferguson, 2015; Settembre et al., 2013). Because of their essential roles in cell homeostasis, the biogenesis and functions of lysosomes are tightly regulated. This is mainly achieved by regulating the subcellular localization and activities of TFEB and TFE3, two transcription factors of lysosome biogenesis and autophagy (Martina et al., 2014; Mills and Taghert, 2012; Raben and Puertollano, 2016; Sardiello et al., 2009; Settembre et al., 2011). For example, in cells with sufficient nutrients, the lysosome-localized mammalian target of rapamycin (mTOR) phosphorylates TFEB (at Ser142 and Ser211) and TFE3 (at Ser321), leading to their release from lysosomes and subsequent interaction with 14–3-3 proteins (Martina et al., 2012, 2014; Martina and Puertollano, 2013; Roczniak-Ferguson et al., 2012; Settembre et al., 2012). This keeps TFEB and TFE3 in the cytosol, where they are inactive. When mTOR activity is inhibited by starvation or other conditions, no further phosphorylation of TFEB/TFE3 occurs; instead, they are dephosphorylated by the phosphatase calcineurin, leading to their nuclear translocation and activation (Medina et al., 2015; Wang et al., 2015). Other signals may converge on mTOR to regulate TFEB/TFE3 activity (Puertollano et al., 2018). In addition, PKC-GSK3β signaling regulates TFEB phosphorylation at Ser138 and Ser134 to affect its subcellular localization in an mTOR-independent manner (Li et al., 2016). More recently, it was found that the export of TFEB/TFE3 from the nucleus is mediated by the nuclear exportin CRM1 (Li et al., 2018; Napolitano et al., 2018). However, the signaling mechanism that directs TFEB/TFE3 nuclear export is unclear.
Although lysosomes are known to respond to many different signals by controlling their own biogenesis through TFEB and TFE3 (Raben and Puertollano, 2016; Settembre et al., 2013), it is not known whether lysosomes change their numbers in a mother cell for dispensation to daughter cells at mitotic cell division. Successful cell division involves G1 (the first gap), S (DNA synthesis), G2 (the second gap), and M (mitosis) phases, which are driven by cyclin-dependent kinases (CDKs; Asghar et al., 2015; Lim and Kaldis, 2013; Sherr et al., 2016); however, the link between cell cycle progression and lysosome biogenesis remains to be uncovered. Here, we reveal the essential role of CDK4 and CDK6 in the nuclear export of TFEB and TFE3. We found that CDK4 and CDK6 interact with and phosphorylate nuclear TFEB and TFE3, thereby promoting their shuttling to the cytoplasm. We further found that lysosome biogenesis is elevated at the S and G2/M phases when the levels of cyclin D1, the activator of CDK4 and CDK6, decline. These results thus reveal not only a mechanism that directs the nuclear export of TFEB and TFE3 but also a mechanism that regulates lysosome biogenesis in the cell cycle.
CDK4/6 inhibitors induce TFEB- and TFE3-dependent lysosome biogenesis
To explore the mechanisms that underlie lysosome biogenesis, we performed screens for both natural and commercial small-molecule compounds that increase lysosomal abundance. We previously reported that the natural compounds HEP14 and HEP15 induce lysosome biogenesis in an mTOR-independent and PKC-dependent manner (Li et al., 2016). Our screen also identified two commercial compounds, LY2835219 (abemaciclib) and PD0332991 (palbociclib), that are known to specifically inhibit CDK4/6 (Fig. 1 A and Table S1). Both LY2835219 and PD0332991 increased LysoTracker Red staining in HeLa cells in a concentration-dependent manner, similar to the mTOR inhibitor Torin 1 (Fig. 1, A and B). LY2835219 also increased LysoTracker Red staining in several other cell types (Fig. S1 A). Western blotting and immunostaining indicated that the protein levels of lysosome-associated membrane protein l (LAMP1) and lysosome-integral membrane protein II (LIMPII) were significantly increased following LY2835219 treatment (Fig. 1, C and D). Lysosome numbers, measured by counting LAMP1 foci, were also significantly elevated (Fig. 1 D). In addition, LY2835219 triggered the formation of GFP-LC3 puncta (Fig. 1 E), and the endogenous levels of LC3B-II were strongly increased (Fig. 1, E and F). PD0332991 enhanced endogenous LAMP1 levels and LC3B puncta, like Torin 1 (Fig. S1 B). Using RFP-GFP-LC3 as an indicator, we confirmed that LY2835219 strongly induced the formation of autolysosomes (foci positive for RFP and negative for GFP; Fig. S1 C). Coimmunostaining of endogenous LAMP1 with mCherry in mCherry-LC3–expressing cells further revealed that the numbers of both free lysosomes (foci positive for LAMP1 and negative for mCherry-LC3) and autolysosomes (foci positive for both LAMP1 and mCherry-LC3) were significantly increased by LY2835219 (Fig. S1 D). Collectively, the above findings suggest that these CDK4/6 inhibitors induced lysosome biogenesis and autophagy.
We next investigated if the ability of the CDK4/6 inhibitors to increase lysosome abundance requires TFEB and TFE3, the transcription factors that scale lysosome biogenesis and autophagy (Mills and Taghert, 2012; Raben and Puertollano, 2016; Settembre et al., 2013). siRNA depletion of both TFEB and TFE3 significantly reduced the LY2835219- and Torin 1–induced increases in LAMP1 and LC3B, as indicated by immunostaining (Fig. 1, G and H). This suggests that these CDK4/6 inhibitors, like Torin 1, function through TFEB and TFE3 to induce lysosome biogenesis. Supporting this conclusion, LY2835219 and PD0332991 treatment led to obvious nuclear localization of endogenous TFEB and TFE3, as well as ectopically expressed TFEB-EGFP, mCherry-TFE3, and EGFP-TFE3 (Fig. 1 I and Fig. S1 E). Consistent with this, several TFEB/TFE3 target genes were significantly up-regulated following exposure of cells to LY2835219 (Fig. S1 F; Sardiello et al., 2009). Pharmacological inhibitors of other CDKs—except for R547 and SU9516, which are weakly active against CDK4—did not induce obvious nuclear translocation of TFEB-EGFP (Table S1). These results indicate that specific inhibition of CDK4/6 leads to TFEB and TFE3 activation and consequently to increased lysosome biogenesis and autophagy.
Genetic depletion of CDK4 or CDK6 leads to TFEB and TFE3 activation
To consolidate the conclusion that CDK4/6 inhibition led to TFEB/TFE3 activation and lysosome biogenesis, we knocked down both CDK4 and CDK6 in HeLa cells using siRNA. CDK4/6 knockdown strongly increased LysoTracker Red staining, LAMP1 protein levels, and the number of LAMP1 foci, as revealed by immunostaining and Western blotting (Fig. 2, A and B). To corroborate these results, we further generated CDK4 knockout (KO) and CDK6 KO HeLa cells using CRISPR/Cas9. Cells with KO of CDK4 or CDK6 had strongly increased LysoTracker Red staining, LAMP1 levels, and numbers of LAMP1 foci compared with the control HeLa cells (Fig. 2, C and D). In addition, LC3 puncta and LC3B-II levels were also significantly increased in these KO cells compared with the control (Fig. 2, E and F). These results suggest that loss of CDK4 or CDK6 led to lysosome biogenesis and autophagy.
We then examined the subcellular localizations of TFEB and TFE3. Compared with control cells, a higher percentage of CDK4 KO and CDK6 KO cells had nucleus-localized endogenous TFEB and TFE3 (Fig. 2 G). Similarly, ectopically expressed TFEB-EGFP and EGFP-TFE3 also localized to the nucleus in a high percentage of CDK4 and CDK6 KO cells compared with their predominantly cytosolic localization in control cells (Fig. 2 G). LY2835219 further increased the nuclear localization of endogenous TFEB and TFE3 in CDK4 KO cells and CDK6 KO cells (Fig. S1 G), which suggests that CDK4 and CDK6 probably act redundantly on these transcription factors. siRNA knockdown of TFEB and TFE3 significantly suppressed the increase in LAMP1 levels, LAMP1 foci, and LC3 puncta in CDK4 KO cells and CDK6 KO cells (Fig. 2, H and I). Taken together, these findings confirmed that inhibition or loss of CDK4/6 led to TFEB and TFE3 activation, and hence, lysosome biogenesis and autophagy.
CDK4 and CDK6 interact with TFEB and TFE3 in the nucleus
We next investigated how CDK4 and CDK6 activate TFEB and TFE3. In HeLa cells, LY2835219 treatment did not obviously change the phosphorylation of the mTOR substrates S6 kinase or ULK1, in contrast to the strong inhibition of phosphorylation by Torin 1 (Fig. S2 A). Similarly, LY2835219 did not change the phosphorylation of ERK1/2 (extracellular regulated protein kinase 1/2) or AKT, in contrast to the inhibitors of these kinases (Fig. S2, B and C). In addition, LY2835219 did not change GSK3β phosphorylation or induce membrane translocation of PKCα or PKCδ, unlike HEP14 (Li et al., 2016; Fig. S2, D and E), which indicates that CDK4/6 inhibition did not affect the PKC-GSK3β signaling axis. Consistent with these findings, phosphorylation of ULK1, ERK1/2, and GSK3β was not changed in CDK4 or CDK6 KO cells (Fig. S2 F). CDK4 or CDK6 KO did not change the subcellular localization of either PKCα or PKCδ (Fig. S2 G). Thus, inhibition of CDK4/6 did not cause TFEB/TFE3 activation by directly affecting mTOR, ERK1/2, AKT, GSK3β, or PKC.
We thus investigated if CDK4 and CDK6 directly act on TFEB and TFE3. In cells under normal conditions, immunostaining assays revealed that CDK4 and CDK6 and their cofactors cyclin D1 and cyclin D3 predominantly localized to the nucleus, while TFEB-EGFP and EGFP-TFE3 were cytosolic (Fig. 3, A–D; and Fig. S3, A–D). When cells were treated with Torin 1, TFEB-EGFP and EGFP-TFE3 translocated into the nucleus and showed increased colocalization with CDK4 and CDK6, as well as with cyclin D1 and cyclin D3 (Fig. 3, A–D; and Fig. S3, A–D). Using coimmunoprecipitation (Co-IP) assays, we found that ectopically expressed EGFP-CDK4 and EGFP-CDK6 were coprecipitated with Flag-TFEB and Flag-TFE3 (Fig. 3, E and F). Similarly, ectopically expressed cyclin D1–EGFP and cyclin D3–EGFP interacted with Flag-TFEB and Flag-TFE3 in Co-IP assays (Fig. S3, E–H).
We next investigated where the interactions of TFEB and TFE3 with CDK4 and CDK6 occur. Subcellular fractionation indicated that endogenous TFEB and TFE3 are predominantly cytosolic, while CDK4, CDK6, cyclin D1, and cyclin D3 are enriched in the nucleus (Fig. 3 G). However, in Co-IP assays, cytosolic TFEB and TFE3 did not coprecipitate with cytosolic CDK4, CDK6, cyclin D1, or cyclin D3. In contrast, nuclear TFEB and TFE3 were coimmunoprecipitated with nuclear CDK4, CDK6, cyclin D1, and cyclin D3 (Fig. 3, H–K). Taken together, these results indicated that TFEB and TFE3 interact with CDK4 and CDK6 in the nucleus.
CDK4 and CDK6 phosphorylate TFEB and TFE3
We then investigated whether CDK4 and CDK6 phosphorylate TFEB and TFE3. We purified recombinant proteins of His6- and SMT-fused TFEB(1–130), TFEB(105–300), and TFEB(295–476) and incubated them with recombinant CDK4/cyclin D1 or CDK6/cyclin D3 complex in the presence of 32γ-ATP. Both CDK4 and CDK6 phosphorylated these TFEB fragments (Fig. 4 A). The strongest phosphorylation signals were observed in the reactions containing CDK4 + TFEB(295–476) and CDK6 + TFEB(105–300). The phosphorylation reactions were inhibited by LY2835219 (Fig. 4 A), indicating that the phosphorylation was dependent on CDK4/6.
We next performed mass spectrometry to determine which amino acid residues in these TFEB fragments were phosphorylated. Our results indicated that Ser114, Ser142, Thr331, and Ser467 were most frequently phosphorylated by both CDK4 and CDK6 (Fig. 4 B and Fig. S4). We thus mutated these residues individually in TFEB-EGFP and examined the subcellular localizations of the mutant proteins. Neither S114A nor T331A caused obvious nuclear localization of TFEB-EGFP; however, TFEB(S142A)–EGFP was observed in the nucleus in ∼70% of cells (Fig. 4 C), as reported previously (Settembre et al., 2011, 2012). The S467A mutation also led TFEB-EGFP to localize to the nucleus in a substantial population of cells (Fig. 4 C), as reported (Palmieri et al., 2017). Nevertheless, TFEB-EGFP containing S114A/S142A/T331A/S467A had a similar nuclear distribution to the S142A mutant (Fig. 4 C). Thus, although CDK4 and CDK6 phosphorylate TFEB at multiple sites, the phosphorylation at Ser142 plays a major role in preventing TFEB from localizing to the nucleus. Supporting this notion, in vitro phosphorylation by CDK4 or CDK6 of TFEB(105–300) containing S142A, but not S211A, was strongly reduced compared with the WT (Fig. 4 D). In addition, LY2835219 strongly inhibited the phosphorylation of TFEB at Ser142 in cells expressing TFEB-EGFP, as detected with an antibody that recognizes phospho-Ser142 (p-Ser142) of TFEB, p-Ser246 of TFE3, and p-Ser180 of microphthalmia-associated transcription factor (MITF; Fig. 4 E). In CDK4 and CDK6 KO cells, the p-Ser142 signal was decreased compared with that in control HeLa cells (Fig. 4 F). Collectively, these findings indicate that CDK4/6 phosphorylate TFEB on Ser142 in cells.
Because TFE3 shares high similarity with TFEB (Fig. 4 G), we performed in vitro phosphorylation of recombinant TFE3(213–411), which contains Ser246, the residue corresponding to Ser142 in TFEB (Fig. 4 G). Both CDK4 and CDK6 phosphorylated TFE3(213–411) in an LY2835219-dependent manner (Fig. 4 H). When Ser246 was mutated to Ala, the phosphorylation signals were strongly reduced (Fig. 4 H). Using the same antibody that recognizes p-Ser142 in TFEB, which also recognizes p-Ser246, we found that LY2835219 strongly inhibited EGFP-TFE3 phosphorylation at Ser246 in cells (Fig. 4 I). The p-Ser246 signals were also decreased in CDK4 or CDK6 KO cells (Fig. 4 J). Consistent with this, EGFP-TFE3(S246A) strongly localized to the nucleus (Fig. 4 K). The TFE3(S568A) mutation, like TFEB(S467A), increased the degree of nuclear localization (Fig. 4 K). However, the TFE3 S246A/S568A double mutation did not significantly increase the nuclear localization compared with the S246A single mutation (Fig. 4 K), which suggests that Ser246 is the determinant for TFE3 to localize to the nucleus. Altogether, these results suggest that TFE3 phosphorylation at Ser246 by CDK4/6, like TFEB phosphorylation at Ser142, prevents it from localizing to the nucleus.
Phosphorylation of TFEB and TFE3 by CDK4/6 is essential for their nuclear export
TFEB and TFE3 shuttle between the cytosol and lysosomes and between the cytosol and the nucleus (Napolitano et al., 2018; Raben and Puertollano, 2016). Under normal conditions, TFEB and TFE3 imported into the nucleus probably move quickly back to the cytosol, and thus they are mainly observed in the cytosol. Supporting this notion, inactivation of the nuclear pore protein CRM1 causes nuclear retention of TFEB (Li et al., 2018; Napolitano et al., 2018). Because CDK4/6 mainly localize in the nucleus, where they interact with and phosphorylate TFEB and TFE3 (Fig. 3 and Fig. 4), we reasoned that they should play a role in the nuclear export of TFEB and TFE3. To test this hypothesis, we examined the effect of LY2835219 on the localization of TFEB and TFE3 in cells with hyperactive mTOR caused by ectopic expression of RagAGTP-RagCGDP or RagBGTP-RagDGDP (Martina et al., 2014; Martina and Puertollano, 2013). Whereas Torin 1, which inhibits mTOR directly, induced nuclear translocation of TFEB-EGFP and EGFP-TFE3 in cells expressing RagAGTP-RagCGDP or RagBGTP-RagDGDP, LY2835219 failed to induce nuclear localization of TFEB and TFE3 in the same cells (Fig. 5 A). These results suggest that CDK4/6 act on TFEB and TFE3 only after they enter the nucleus. Consistent with this, LY2835219 failed to inhibit phosphorylation of TFEB at Ser142 and of TFE3 at Ser246 (Fig. 5, B and C). In contrast, Torin 1 inhibited the phosphorylation of TFEB at Ser142 and of TFE3 at Ser246 in cells expressing RagAGTP-RagCGDP or RagBGTP-RagDGDP (Fig. 5, B and C). One explanation for this is that inhibition of mTOR not only directly suppressed mTOR-mediated phosphorylation of TFEB Ser142 and TFE3 Ser246 (Puertollano et al., 2018) but also suppressed CDK4/6 activities by down-regulating cyclin D (Alao, 2007). Importantly, the interaction of Flag-tagged TFEB(S142A) with EGFP-tagged CRM1, which is required for TFEB/TFE3 nuclear export (Li et al., 2018; Napolitano et al., 2018), was strongly reduced compared with Flag-tagged WT TFEB in Co-IP assays (Fig. 5 D). These findings, together with the fact that TFEB(S142A) and TFE3(S246A) localize to the nucleus (Fig. 4), suggest that CDK4/6 phosphorylate TFEB and TFE3 in the nucleus to promote their CRM1-dependant nuclear export.
To consolidate this conclusion, we treated HeLa cells expressing TFEB-EGFP or EGFP-TFE3 with Torin 1 to induce their nuclear translocation. We then removed Torin 1 and monitored their export from the nucleus over time. TFEB-EGFP and EGFP-TFE3 rapidly lost their nuclear localization following removal of Torin 1 (Fig. 5, E and F). For example, only ∼20% of cells had obvious nuclear TFEB and TFE3 3 h after removal of Torin 1, compared with ∼100% of cells at the starting time point (Fig. 5, E and F). In the presence of LY2835219, however, ∼90% of cells still contained nuclear TFEB and TFE3 3 h after Torin 1 removal (Fig. 5, E and F). In CDK4 KO and CDK6 KO cells, nuclear export of TFEB-EGFP and EGFP-TFE3 induced by similar removal of Torin 1 was strongly inhibited (Fig. S5, A and B). These findings suggest that inhibition of CDK4/6 prevents TFEB and TFE3 from leaving the nucleus.
In addition, we examined fluorescence loss in photobleaching (FLIP) of the cytosolic TFEB-EGFP signals following removal of Torin 1. Continuous bleaching of cytosolic TFEB-EGFP led to a sharp decrease in nuclear TFEB-EGFP signals, indicating that nuclear TFEB-EGFP was exported from the nucleus to the cytosol, where it was bleached (Fig. 5, G and H). In contrast, nuclear TFEB-EGFP signals decreased at a much slower rate in the same photobleaching assay in the presence of LY2835219 (Fig. 5, G and H).
To determine the specific effect of CDK4/6 on the nuclear export of TFEB and TFE3, we examined the subcellular localization of EGFP-tagged FOXO3 and nuclear factor κB (NF-κB), two transcription factors that undergo nucleo-cytoplasmic shuttling and require CRM1 for nuclear export (Latré de Laté et al., 2010; Zhang et al., 2012). Unlike TFEB-EGFP and EGFP-TFE3, FOXO3-EGFP and NF-κB–EGFP were cytosolic in CDK4 KO and CDK6 KO cells (Fig. S5 C). In addition, we treated HeLa cells expressing FOXO3-EGFP and NF-κB–EGFP with LY2835219 or the CRM1 inhibitor leptomycin B (LMB). While LMB caused strong nuclear retention of all four transcription factors, LY2835219 did not change the cytosolic localization of FOXO3 and NF-κB (Fig. S5 D). Altogether, these results demonstrate that CDK4/6 phosphorylate nuclear TFEB and TFE3 to recycle them back to the cytosol.
CDK4/6 regulate cell cycle–dependent lysosome biogenesis
The findings that CDK4/6 inhibition led to nuclear retention and activation of TFEB and TFE3 prompted us to examine lysosome biogenesis during the cell cycle. We simultaneously stained HeLa, HepG2, and HCT116 cells with Hoechst 33342, which labels chromosomal DNA, and LysoTracker Red, which stains lysosomes, and performed flow cytometry to analyze lysosomal abundance in each phase of the cell cycle. In all cell types examined, cells at the S and G2/M phases had a significantly higher intensity of LysoTracker Red staining than the G1 phase cells (Fig. 6, A and D). In another assay, we first fed cells with FITC-Dextran and allowed it to reach lysosomes, and then we fixed the cells and stained them with propidium iodide (PI), which labels nuclear DNA. We then performed flow cytometry to determine lysosome numbers in cells at distinct cell cycle phases. The results indicated that cells at the S and G2/M phases had a significantly higher intensity of FITC-Dextran labeling than cells at the G1 phase (Fig. 6, B and D). Furthermore, we immunostained endogenous LAMP1 in HeLa, HepG2, and HCT116 cells and performed flow cytometry. We found that LAMP1 staining was significantly increased in cells at the S and G2/M phases compared with cells at the G1 phase (Fig. 6, C and D). Collectively, these findings indicated that lysosome biogenesis is increased in the S and G2/M phases of the cell cycle.
CDK4 and CDK6 are known to drive G1 phase progression, and their activities are dependent on D-type cyclins (Fig. 6 E; Asghar et al., 2015; Lim and Kaldis, 2013; Sherr et al., 2016). Cyclin D1 is quickly degraded at the G1/S boundary of the cell cycle by the ubiquitin–26S proteasome pathway, thus switching off CDK4/6 activities (Fig. 6 E; Alt et al., 2000; Diehl et al., 1998, 1997; Qie and Diehl, 2016; Sherr et al., 2016). To determine whether the increase in lysosome biogenesis corresponds to this inactivation of CDK4/6, we synchronized cells to distinct cell cycle stages using standard assays (Ma and Poon, 2011). In HeLa, HepG2, and HCT116 cells, the S, G2, and M phases had strongly increased LAMP1 levels compared with the G1 phase, while the cyclin D1 protein levels were greatly decreased (Fig. 6, F and G). The levels of cyclin E1, which activates CDK2 at the late G1 stage (Goel et al., 2018; Koff et al., 1992; Sherr et al., 2016), were also down-regulated like cyclin D1 (Fig. 6 G), as reported previously (Baldin et al., 1993; Koff et al., 1992). In contrast, cyclin B1, which activates CDK1 in the M phase (Goel et al., 2018; Lim and Kaldis, 2013; Sherr et al., 2016), was increased at the G2 and M phases (Fig. 6 G). Pharmacological inhibition of CDK1, CDK2, and other CDKs did not activate TFEB (Table S1), while CDK4/6 inhibition led to increased lysosome biogenesis in a TFEB- and TFE3-dependent manner. These findings collectively suggest that the CDK4/6–TFEB/TFE3 axis contributes to the elevated lysosome biogenesis at the S–M phases.
To further investigate the role of the CDK4/6–TFEB/TFE3 axis in lysosome biogenesis in the cell cycle, we first determined the lysosome abundance in CDK4 KO and CDK6 KO cells. Compared with the control cells at the G1 phase, CDK4 KO and CDK6 KO cells at G1 had significant increases in the intensity of LysoTracker Red staining (Fig. 7 A). This indicates that reducing CDK4/6 activity indeed led to elevated lysosome biogenesis. CDK4 KO and CDK6 KO cells at the S–M phases also had a higher intensity of LysoTracker Red staining, which likely resulted from a stronger effect of CDK4 or CDK6 KO compared with cyclin D turnover-induced decline in CDK4/6 activity in control cells at the S–M phases (Fig. 7 A).
We further treated HeLa cells with LY2835219 for 12 h, which more thoroughly inhibited CDK4/6 activity (Jansen et al., 2017), and examined lysosomal abundance in each cell cycle phase. LY2835219 significantly increased LysoTracker Red staining in G1 cells, and the increase was comparable with that in S and G2/M cells not treated with LY2835219 (Fig. 7 B). Thus, inhibition of CDK4/6 at G1 enhanced lysosome biogenesis, similar to the lysosomal increase at the S and G2/M phases when CDK4/6 activity was diminished by cyclin D turnover. Intriguingly, LY2835219 further enhanced LysoTracker Red staining in S and G2/M cells (Fig. 7 B). A possible explanation for this is that S–M–phase cells have residual CDK4/6 activities; thus, inhibition of CDK4/6 further increased lysosomal abundance.
We next performed siRNA to knock down TFEB and TFE3 together, and we analyzed lysosomal abundance in each phase of the cell cycle in HeLa cells. TFEB + TFE3 siRNA significantly reduced lysosomal abundance, measured with LysoTracker Red staining or FITC-Dextran labeling (Fig. 7 C). This suggests that TFEB and TFE3 are required for the lysosomal increase in the S–M phases. We then examined TFEB Ser142 phosphorylation in TFEB-EGFP–expressing HeLa cells sorted with flow cytometry and found that TFEB Ser142 phosphorylation was strongly reduced in the S–M phases compared with G1 (Fig. 7 D). In addition, in HeLa cells synchronized to individual cell cycle phases, the nuclear localization of TFEB and TFE3 markedly increased at the S–M phases compared with G1 (Fig. 7, E–G). Thus, the reduction of TFEB Ser142 phosphorylation and the increase in TFEB and TFE3 nuclear localization correlated well with the inactivation of CDK4 and CDK6 in the S–M phases. Taken together, these results provide further evidence that the CDK4/6–TFEB/TFE3 axis regulates lysosome biogenesis in the cell cycle (Fig. 7 H).
Inhibition of CDK4/6 promotes cellular clearance of various substrates
Finally, we examined if lysosomes induced by CDK4/6 inhibition are functional. LY2835219-induced lysosomes were positive for Magic Red, an indicator of active lysosomal protease cathepsin B (Fig. 8 A; Boonacker and Van Noorden, 2001; Van Noorden et al., 1997). This effect was inhibited by bafilomycin A1 (BFA1), an inhibitor of lysosomal acidification (Fig. 8 A). Similarly, LY2835219-induced LysoTracker Red–positive lysosomes were also positive for BODIPY–pepstatin A, which indicates the activation of the lysosomal protease cathepsin D (Fig. 8 B; Chen et al., 2000). Furthermore, LY2835219, like Torin 1, increased lysosomal protease activities measured with β-N-acetylglucosaminidase (NAG) assays (Fig. 8 C; Li et al., 2016). In HepG2 cells that were overloaded with oleic acid, LY2835219 treatment strongly reduced the number of lipid droplets, as did Torin 1 (Fig. 8 D). In the presence of the lysosomal inhibitor BFA1, neither LY2835219 nor Torin 1 reduced the number of lipid droplets (Fig. 8 D), indicating that they acted in a lysosome-dependent manner. In HeLa cells that express Htt97Q-GFP, LY2835219 and Torin 1 similarly reduced the accumulation of polyglutamine (polyQ) aggregates (Fig. 8 E). Altogether, these results suggest that inhibition of CDK4/6 promotes lysosomal activity and cellular clearance of a variety of substrates.
It is not understood how intracellular organelles replicate themselves in order to be dispensed to daughter cells at mitotic cell division. Here, our findings define a mechanism that regulates lysosome biogenesis during the cell cycle. In cells with sufficient nutrients, pharmacological inhibition or genetic inactivation of CDK4/6 leads to nuclear retention and activation of TFEB and TFE3, consequently resulting in TFEB- and TFE3-dependent lysosome biogenesis and autophagy. Our results further revealed that CDK4 and CDK6 interact with and phosphorylate TFEB and TFE3 solely in the nucleus, because only nuclear TFEB and TFE3 coprecipitated with nuclear CDK4 and CDK6 as well as with their activators cyclin D1 and cyclin D3. CDK4/6 directly phosphorylate TFEB at Ser142 and TFE3 at Ser246 in the nucleus, as evidenced by the finding that LY2835219 failed to inhibit such phosphorylation when TFEB and TFE3 were prevented from translocating to the nucleus. A TFEB mutant that cannot be phosphorylated by CDK4/6 has a much weaker interaction with the CRM1 exportin than WT TFEB does, which suggests that CDK4/6–mediated phosphorylation is essential for CRM1-dependent nuclear export of TFEB/TFE3. In addition, it is possible that nuclear TFEB and TFE3, if not phosphorylated, have higher transcriptional activities. Thus, inhibition of CDK4/6 reduces the nucleo-cytoplasmic shuttling of TFEB and TFE3 and consequently enhances their activation.
Importantly, inhibitors of other major CDKs do not affect the subcellular localization of TFEB. These findings collectively suggest that CDK4 and CDK6, rather than other CDKs, regulate lysosome biogenesis through TFEB and TFE3 in the cell cycle. Supporting this conclusion, Ser142 phosphorylation of TFEB is markedly reduced and nuclear retention of TFEB and lysosome biogenesis are greatly increased in the S–M phases, concomitant with the decline of cyclin D1, the essential activator of CDK4/6. Thus, the high activities of CDK4/6 suppress lysosome biogenesis at the G1 phase, while their low activities enhance lysosome biogenesis in the S–M phases by relieving the inhibition of TFEB and TFE3 (Fig. 7 H).
Phosphorylation of cytoplasmic TFEB and TFE3 keeps them in the cytoplasm, where they are inactive. On the lysosome, mTOR phosphorylates TFEB (Ser211 and Ser142) and TFE3 (Ser321 and Ser246), leading to their lysosomal release and subsequent interaction with 14–3-3 proteins in the cytoplasm (Martina et al., 2012, 2014; Roczniak-Ferguson et al., 2012; Settembre et al., 2012; Vega-Rubin-de-Celis et al., 2017). TFEB Ser142, and probably TFE3 Ser246, can also be phosphorylated by ERK (Settembre et al., 2011, 2012). Under some circumstances, TFEB might be phosphorylated at Ser3 by MAP4K3 or at Ser122 by mTOR (Hsu et al., 2018; Vega-Rubin-de-Celis et al., 2017). When cells are devoid of nutrients, mTOR is inactivated and fails to phosphorylate TFEB and TFE3. In the meantime, lysosomal calcium efflux is induced, which activates calcineurin to dephosphorylate existing phosphorylated TFEB and TFE3 (Medina et al., 2015; Wang et al., 2015). This leads to their nuclear translocation and activation. In addition to nutrient signaling, GSK3β phosphorylates TFEB (Ser134 and Ser138) to promote its lysosomal localization (Li et al., 2016). It remains to be determined whether phosphorylation of TFEB at Ser142 by mTOR or ERK primes GSK3β-dependent Ser138 and Ser134 phosphorylation. However, activation of PKC leads to GSK3β phosphorylation and inactivation without compromising mTOR, resulting in TFEB dephosphorylation and nuclear translocation (Li et al., 2016). Thus, cytoplasmic phosphorylation by mTOR, ERK, GSK3β, and other kinases maintains TFEB and TFE3 in a cytoplasmic and inactive state (Puertollano et al., 2018). Interestingly, recent studies suggest that mTOR and probably GSK3β also phosphorylate TFEB (Ser142 and Ser138) in the nucleus, promoting its CRM1-mediated nuclear export and inactivation (Li et al., 2018; Napolitano et al., 2018). Nevertheless, further investigation is required to determine whether mTOR and GSK3β indeed localize and function in the nucleus (Napolitano et al., 2018).
Our findings clearly demonstrated that in cells with sufficient nutrients, CDK4/6 are the kinases that phosphorylate TFEB and TFE3 in the nucleus, especially at the G1 phase of the cell cycle. It is likely that after being exported into the cytoplasm, Ser142-phosporylated TFEB is further phosphorylated at Ser138 and Ser134 by GSK3β, promoting its cytoplasmic localization. Alternatively, CDK4/6 phosphorylation of TFEB at Ser142 may prime the phosphorylation at Ser138 and Ser134 by GSK3β in the nucleus at the G1 phase, if GSK3β indeed has a nuclear role. In the S–M phases, cyclin D1 is exported through the CRM1 exportin to the cytoplasm and subsequently degraded by the 26S proteasome (Alt et al., 2000; Diehl et al., 1998, 1997). This switches off CDK4/6 activity and consequently TFEB Ser142 phosphorylation, thus activating TFEB and TFE3 by retaining them in the nucleus.
CDK4/6 inhibitors including LY2835219 and PD0332991 are currently being used or are under clinical trial for cancer therapy (Asghar et al., 2015; Goel et al., 2018; Lim and Kaldis, 2013; Qie and Diehl, 2016; Sherr et al., 2016). Our findings that CDK4/6 inhibitors induce lysosome biogenesis and autophagy have implications for the use of CDK4/6 inhibitors in cancer treatment, because autophagy has differential impacts on distinct phases in tumorigenesis (Acevedo et al., 2016; Galluzzi et al., 2015; Valenzuela et al., 2017; Vijayaraghavan et al., 2017). Moreover, it will be valuable to evaluate the potential use of CDK4/6 inhibitors for treatment of lysosome-related disorders based on our findings that CDK4/6 inhibitors promote lysosome functions and cellular clearance.
Materials and methods
Cell culture, transfection, and reagents
All cell lines used in this work were obtained from ATCC. Cells were cultured at 37°C with 5% CO2 in DMEM (Gibco) with 10% heat-inactivated FBS (BioInd), 100 U/ml penicillin, and 100 µg/ml streptomycin. Cells were tested for mycoplasma contamination using DAPI staining, and all test results were negative. Transient transfections were performed with Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. Other reagents involved in this study are listed in Table S2.
Generation of CDK4 and CDK6 KO HeLa cells
CDK4 and CDK6 KO cells were generated using CRISPR/Cas9 technology. The CDK4 guide RNA, CDK4sgRNA1 (5′-CACCGATCTCGGTGAACGATGCAAT-3′), was cloned into the pSpCas9 (BB)–2A-GFP (PX458) vector (Addgene; plasmid 48138) and transfected into HeLa cells. 72 h later, GFP-positive cells were sorted by FACS and cultured for 10 d. 48 colonies were picked and examined for deletion in the CDK4 gene by PCR and Western blotting. CDK4 deletion was further confirmed by sequencing. Similar strategies were applied to generate CDK6-KO cells using the CDK6 guide RNA, CDK6sgRNA1 (5′-CACCGTTAGATCGCGATGCACTACT-3′).
RNA oligonucleotides used for siRNA are listed in Table S3. Cells grown in 6-well plates or confocal dishes were transfected with 100 pmol RNA oligonucleotides using Lipofectamine 2000 according to the manufacturer’s instructions. Cells were harvested for further experiments 48 h after transfection. The efficiency of siRNAs was evaluated by Western blotting.
The antibodies used in this study are listed in Table S4.
Examination of lysosomes and lysosomal activities
Cells grown on glass-bottom dishes (Cellvis) were treated with individual compounds for 3 h and further cultured in fresh medium containing LysoTracker Red DND-99 (0.3 µM) for 30 min. Cells were then changed to LysoTracker-free medium and observed by confocal microscopy. To assess cathepsin B and cathepsin D activities, cells were treated with individual compounds for 3 h, followed by staining with Magic Red substrates (ImmunoChemistry Technologies) or BODIPY FL–pepstatin A (1 µM) according to the manufacturer’s instruction.
NAG assays were performed using a kit from Sigma-Aldrich (CS0780). HeLa cells were treated with different concentrations of LY2835219 or Torin 1 (1 µM) for 3 h and lysed with RIPA buffer. Total protein concentration was determined with the BCA Protein Assay Kit (Thermo Fisher Scientific). NAG activity was measured in 20 µg cell lysate from each sample.
Immunostaining and confocal microscopy
For immunostaining, cells grown on glass-bottom dishes or coverslips were fixed with 4% PFA in PBS at room temperature for 15 min and permeabilized with 0.5% saponin (Sigma-Aldrich) in PBS at room temperature for another 15 min. Cells were incubated with primary antibodies in the staining buffer (1% BSA and 0.05% saponin in PBS) at 4°C overnight. In colocalization experiments, cells were permeabilized with 0.1% Triton X-100 for 5 min and then incubated with blocking buffer (1% BSA, 0.3 M glycine, and 0.1% Tween 20 in PBS) for 1 h. Cells were then incubated with primary antibodies in the staining buffer (1% BSA and 0.1% Tween 20 in PBS) at 4°C overnight. Dilution of individual antibodies is listed in Table S4. Cells were washed three times in PBS and incubated with Alexa Fluor 488– or Alexa Fluor 568–conjugated secondary antibodies for 1 h at room temperature. Cells were washed three times, then stained with DAPI (Sigma-Aldrich). All samples were visualized using a Zeiss LSM 880 inverted Confocal Laser Scanning Microscope with Airyscan and an inverted Olympus FV1000 confocal microscope. Images in colocalization experiments were obtained with Airyscan super-resolution mode using an α Plan-Apochromat 100×/1.46 oil immersion objective lens and were processed and analyzed with ZEN 2 (blue edition), FV10-ASW 4.0a Viewer, or ImageJ (National Institutes of Health).
HeLa cells expressing TFEB-EGFP were grown on 35-mm glass-bottom dishes and treated with 25 nM Torin 1 for 1 h, which was replaced with fresh DMEM medium containing either DMSO or LY2835219 (1 µM) for 1.5 h. The FLIP assays were then performed on a Zeiss LSM 880 Confocal Microscope using a 63× oil immersion objective lens with a 488-nm laser excitation. The cytosol of cells was designated as the photobleaching region, and the fluorescence intensities of the nucleus were acquired for analysis. After scanning for five frames, the cytosol was bleached with three iterations at 5% power, and the nuclear fluorescence was monitored every 10 s at 2–3% power within 30 min. A fluorescent region from an adjoining cell in the same field was used to correct for general photobleaching. The data were generated by fitting in a double exponential decay model using Zen software (Zeiss). Nuclear fluorescence intensity values at each time point were normalized with the start time point. Quantification of normalized nuclear intensity was generated by GraphPad Prism (8.0.1).
Western blotting and IPs
For Western blotting, cells were lysed in ice-cold RIPA buffer (20 mM Tris-HCl, pH 7.5, 100 mM NaCl, 0.1% SDS, 0.5% sodium deoxycholate, and 1 mM PMSF) containing Complete Protease Inhibitor Cocktail and Phosphatase Inhibitor Cocktail (Roche). Cell lysates were centrifuged at 13,000 rpm for 15 min, and 20 µg of total proteins was resolved by SDS-PAGE and probed with the indicated antibodies (Table S4). The amount of α-tubulin was used as the loading control. Blots were developed with chemiluminescent detection reagent (ECL; GE Healthcare and Thermo Fisher Scientific) and imaged with a Smartchemi machine (Sage Creation). Quantification of Western blots was performed using ImageJ software.
For IPs of GFP- or FLAG-tagged proteins, cells were lysed in IP buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% NP-40, 1 mM PMSF, and 1% glycerol) containing Complete Protease Inhibitor Cocktail and Phosphatase Inhibitor Cocktail. Cell lysates were centrifuged at 13,000 rpm for 15 min at 4°C, and the clear supernatants were incubated with either FLAG antibody (M2)–conjugated beads (Sigma-Aldrich) or GFP beads (ChromoTek) overnight at 4°C. The beads were centrifuged and extensively washed in wash buffer. Precipitated proteins were resolved by SDS-PAGE, then blotted and detected with different antibodies.
Co-IPs of endogenous CDK4, CDK6, cyclin D1, or cyclin D3 with TFEB or TFE3 were performed with cytosolic and nuclear fractions. For fractionation, cells were first lysed in cytosol lysis buffer (10 mM Hepes, pH 7.8, 15 mM KCl, 1 mM MgCl2, 0.1 mM EDTA, 1 mM DTT, 1 mM PMSF, and 10% glycerol) containing Complete Protease Inhibitor Cocktail and Phosphatase Inhibitor Cocktail and ground with a tissue homogenizer. Lysates were centrifuged at 1,100× g for 10 min at 4°C. The supernatant was collected and adjusted to 100 mM NaCl and 0.1% NP-40 and used as the cytosolic fraction for IP. The pellets were washed once with the cytosol lysis buffer and lysed in the nucleus lysis buffer (50 mM Hepes, pH 7.8, 3 mM MgCl2, 300 mM NaCl, 1 mM DTT, 0.1 mM PMSF, 10% NP-40, and 10% glycerol) containing Benzonase Nuclease (Sigma-Aldrich) for 1 h at 4°C. The nuclear lysate was adjusted to 100 mM NaCl and 0.1% NP-40 and cleared by centrifugation at 10,000× g for 10 min. The cleared nuclear lysate was further used for IP. The cytosolic and nuclear protein lysates were incubated with primary antibodies of CDK4, CDK6, cyclin D1, and cyclin D3 overnight at 4°C and further incubated with Protein A agarose beads (Thermo Fisher Scientific) for an additional 4 h at 4°C. After extensive washing, precipitated proteins were resolved by SDS-PAGE, then blotted and probed with different antibodies.
In vitro kinase assays
Active GST-CDK4/cyclin D1 complex (Product Number 0142–0143-1) and GST-CDK6/cyclin D3 complex (Product Number 0051–0373-1) used in this study were purchased from ProQinase GmbH. Purified His6-SMT3-TFEB or His6-SMT3-TFE3 proteins (2 µg) immobilized on Ni-chelating Sepharose beads were incubated with CDK4/cyclin D1 complex or CDK6/cyclin D3 complex (1 µg of each) for 2 h at 30°C in 30 µl kinase buffer (25 mM Hepes, pH 7.4, 50 mM NaCl, 5 mM MgCl2, 1 mM CaCl2, 1 mM DTT, and 30 µM ATP) containing 1.0 µCi of [γ-32P]ATP (Perkin Elmer). After extensive washing, proteins bound on beads were resolved by SDS-PAGE, and phosphorylation signals were detected using autoradiography. The same gels were subsequently stained with Coomassie blue to visualize the total proteins.
For mass spectrometry analysis, phosphorylation reactions were performed without adding [γ-32P] ATP and resolved on SDS-PAGE. His6-SMT3-TFEB protein bands were excised and subjected to an in-gel trypsin digest. The tryptic peptides were then analyzed by liquid chromatography–tandem mass spectrometry using an LTQ Orbitrap Elite mass spectrometer (Thermo Fisher Scientific) coupled online to an Easy-nLC 1000 in the data-dependent mode. Data were analyzed by Thermo Scientific Proteome Discoverer software version 1.4 to compare the phosphorylation of TFEB proteins.
Plasmids and site-directed mutagenesis
Detection of the TFEB phospho-142 or TFE3 phospho-246
Cells expressing TFEB-EGFP and EGFP-TFE3 were lysed in lysis buffer (25 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.5 mM EDTA, and 0.1% NP-40) containing Complete Protease Inhibitor Cocktail and Phosphatase Inhibitor Cocktail for 1 h on ice. Lysates were centrifuged at 13,000 rpm for 15 min, and then the lysates were incubated with 10 µl of GFP-Trap agarose beads (Chromotek) for 2 h at 4°C. After extensive washing, proteins were precipitated and resolved by SDS-PAGE, blotted, and probed with GFP antibody and an antibody that recognizes p-Ser142 of TFEB, p-Ser246 of TFE3, and p-Ser180 of MITF (phospho–[TFEB(Ser142)/TFE3(Ser246)/MITF(Ser180)]). This antibody was generated by using a synthetic Ser180 phosphopeptide derived from human MITF, which is conserved with Ser142 in TFEB and Ser246 in TFE3 (PA5-36755; Invitrogen).
Analysis of lysosomes during the cell cycle
HeLa, HepG2, and HCT116 cells were costained with LysoTracker Red (0.3 µM) and Hoechst 33342 for 30 min. After extensive washing, cells were suspended in cold PBS and analyzed by flow cytometry. For LAMP1 immunostaining, cells were first fixed with 4% PFA for 15 min and permeabilized with permeabilization buffer (1× PBS, 0.3% Triton X-100, and 0.5% BSA) for 15 min and incubated with LAMP1 antibody in the incubation buffer (0.5% BSA in PBS) for 1.5 h at room temperature. After extensive washing, cells were resuspended in incubation buffer containing Alexa Fluor 488 secondary antibody for 30 min. Cells were then collected and incubated in PI/RNase staining buffer (4087; Cell Signaling Technology) for an additional 30 min and analyzed by flow cytometry. For the FITC-Dextran experiment, cells were fed with 0.25 mg/ml FITC-Dextran for 12 h followed by 2-h recovery in fresh DMEM medium. Cells were then fixed with 70% ethanol (2 h, 4°C) and stained with PI for 30 min and analyzed by flow cytometry. All samples were analyzed on a FACSAria SORP machine (BD Biosciences), and the fluorescence intensity profiles of each cell cycle stage were analyzed using BD FACSDiva 8.0.1. Cell cycle distribution was analyzed by PI or Hoechst 33342 content and fitted to the Dean-Jett-Fox cell cycle model using FlowJo software.
Synchronization of cells
Synchronization of HeLa, HepG2, and HCT116 cells was performed as described (Ma and Poon, 2011). Briefly, cells were cultured in 100-mm dishes to 40–50% confluency. G1-phase synchronization was performed by lovastatin (20 µM) treatment for 24 h. A double thymidine block procedure was used to synchronize cells in S phase. Cells were first incubated with thymidine (2 mM) for 14 h (first thymidine block). After extensive washing, cells were then incubated with deoxycytidine (24 µM) for 9 h. Thymidine (2 mM) was added again, and the incubation was continued for another 14 h (second thymidine block). Cells were extensively washed again with PBS, then supplemented with deoxycytidine (24 µM) and incubated at 37°C. S-phase cells were immediately collected and analyzed. G2-phase cells were synchronized by treatment with the CDK inhibitor RO3306. After release from the second thymidine block, cells were allowed to recover for 2 h and then incubated with RO3306 (10 µM) for 10 h. M-phase cells were enriched by treating with nocodazole and mechanical shake-off. After release from the second thymidine block, cells were first grown for 2 h, and then incubated with nocodazole (0.1 µg/ml) for 10 h. M-phase cells, which were rounded and less attached to the dishes, were collected by mechanical shake-off. Cells synchronized at different stages were harvested and immunostained with LAMP1 antibody. Western blotting was performed to examine the levels of LAMP1, cyclin D1, cyclin E, and cyclin B. The reagents used in cell synchronization are listed in Table S2.
Quantitative RT-PCR (qRT-PCR)
Total RNA was extracted from cells using TRIzol Reagent (Invitrogen) and chloroform. 2 µg RNA was used as a template to generate cDNAs using the GoScript Reverse Transcription System (Promega). qRT-PCR reactions were performed on a CFX96 Real Time System C1000 Touch Thermal Cycler (BioRad). GAPDH was used as the endogenous reference gene. The primers for TFEB-targeted genes are listed in Table S7.
Lipid droplet clearance assay
HepG2 cells seeded in 35-mm glass-bottom dishes (Cellvis) were fed with oleic acid (100 µM, 12 h) to induce lipid droplet formation. After brief washing, cells were cultured in fresh DMEM medium containing LY2835219 (0.5 µM) or Torin 1 (0.5 µM) without or with BFA1 (0.1 µM) to different time points. Cells were stained with BODIPY (1 µg/ml) for 30 min at 37°C before examination with confocal microscopy.
Htt polyQ clearance assay
HeLa cells stably expressing Tet–on Htt97Q–GFP were incubated with doxycycline (1 µg/ml, 12 h) to induce Htt97Q-GFP expression (Li et al., 2016). Cells were briefly washed to remove doxycycline and further cultured in DMEM medium containing LY2835219 (0.5 µM) or Torin 1 (0.5 µM) to different time points. PolyQ foci were observed under a confocal microscope.
Data were analyzed with GraphPad Prism Software (8.0.1) to generate curves and bar charts. Statistically significant differences between two groups were determined by using the unpaired two-tailed t test. Statistically significant differences in experiments with more than two conditions were determined by one-way ANOVA with the post hoc Holm-Sidak test. *, P < 0.05; **, P < 0.01; ***, P < 0.001; P > 0.05 was considered not significant.
Online supplemental material
Fig. S1 shows that CDK4/6 inhibitors induce lysosome biogenesis. Fig. S2 illustrates that the inactivation of CDK4/6 has no direct effect on mTOR, ERK2, AKT, GSK3β, and PKC activities. Fig. S3 shows that cyclin D1 and cyclin D3 interact with TFEB and TFE3. Fig. S4 shows the results of mass spectrometry analysis of the phosphorylated sites in TFEB(105–300) after treatment with CDK4/cyclin D1 or CDK6/cyclin D3 complex. Fig. S5 shows that CDK4 and CDK6 are essential for TFEB and TFE3 nuclear export. Table S1 shows the effects of other CDK inhibitors on TFEB translocation. Table S2 lists chemical compounds and reagents used in the study. Table S3 lists siRNA oligos used in the study. Table S4 lists all antibodies used in the study. Table S5 summarizes the expression vectors used in this study. Table S6 lists the oligos for site-directed mutagenesis. Table S7 lists the primers for qRT-PCR.
We thank Dr. X. Ye for suggestions and Dr. I. Hanson for proofreading the manuscript.
This research was supported by grants from the National Natural Science Foundation of China (31730053 to C. Yang, 31801136 to Q. Yin, and 31671480 to Y. Jian) and the National Basic Research Program of China (2017YFA0503403 to C. Yang).
The authors declare no competing financial interests.
Author contributions: C. Yang and Q. Yin conceived and designed the research. C. Yang supervised the research. Q. Yin did most of the experiments and analyzed the data. Y. Jian performed in vitro kinase assays and IPs. M. Xu performed compound screening and generated KO cells and some plasmids. Y. Wang and X. Huang performed mass spectrometry and analysis. Z. Liu contributed to flow cytometry and data analysis. L. Xu, N. Wang, Q. Li, J. Li, and H. Zhou contributed materials. C. Yang and Q. Yin wrote the manuscript with feedback from all authors.
Q. Yin and Y. Jian contributed equally to this paper.