To study the function of DYRK1A, we have performed knockdown of DYRK1A in multiple cell lines, including in HEK293 cells (18). We noticed that DYRK1A depleted cells were smaller in size compared to the control cells. To explore if DYRK1A has a role in maintaining cell size, we performed systematic analysis of the cell size phenotypes after transient knockdown of DYRK1A in HEK293 and observed a reduction in cell size (Figure 1A-B). Since mutations in DYRK1A are associated with both stunted growth as well as microcephaly, we wondered if neuronal cells require DYRK1A for maintaining proper cell size. Therefore, we knocked down DYRK1A in human neuroblastoma cell line SH-SY5Y and observed significant reduction in cell size (Figure 1 C-D). Further, we tested if the observed cell size phenotype is conserved in mouse. We targeted Dyrk1a in NIH-3T3 using CRISPR/Cas9 and observed a significant reduction in cell size (Figure 1E-F). Our data suggests that DYRK1A promotes cell size across cell lines derived from different models and tissues. Thus, we tested if the overexpression of DYRK1A can promote cell size. Since strong overexpression of DYRK1A leads to cell cycle exit (2, 30, 31), we used the Doxycycline-inducible system to express low levels of Flag-DYRK1A in HEK293 cells (32). We noted a change in cell size with increasing dosage of Doxycycline that reached saturation state at 100ng/ml Doxycycline in our system (Supplementary figure 1). We observed a significant increase in cell size at 40 ng/ml Doxycycline treatment for 48 hours, when the DYRK1A mRNA levels were approximately twice that of control (Figure 1G-H). Overall, our knockdown and overexpression experiments in different cell lines, including both human and mouse suggest that DYRK1A promotes cell size.

To understand how DYRK1A promotes cell growth, we investigated DYRK1A protein-interactors. Previously, we have identified protein-interactors of wild-type and kinase-dead DYRK1A (K188R) from HEK293 cells using affinity purification and mass spectrometry (18, 32). In these purifications, we have observed significant enrichment of DCAF7, ARIP4, RB1, EP300, CREBBP, TRAF2, TRAF3, FAM53C, RNF169, many of which have also been identified by many other groups as bona-fide DYRK1A interacting proteins (33–37). We also observed significant enrichment of TSC components TSC1 and TSC2 among proteins co-purified with wild-type as well as kinase-dead DYRK1A (K188R) (Figure 2A). Interestingly, a few peptides of TBC1D7, the third component of TSC were also detected in the wild-type pull-down (see dNSAF values in Figure 2 A), suggesting that DYRK1A interacts with the TSC complex. This interaction is intriguing as TSC is a negative regulator of the mTORC1, which is known to promote cell growth (38–40).

To ascertain the interaction between DYRK1A and TSC, we expressed Flag-DYRK1A in HEK293 cells through the Doxycycline-inducible system and performed Flag affinity purification. Surprisingly, when we used HEPES buffer containing either 0.5% NP40 or 1% Tween 20, we did not detect co-purification of TSC2 with Flag-DYRK1A (data not shown). However, when we prepared extracts by lysing the cells in hypotonic buffer without detergent followed by Dounce homogenization, we found that both, TSC1 and TSC2 co-purify with Flag-DYRK1A (Figure 2B). Further, we tested the interaction between endogenous DYRK1A and TSC by immunoprecipitating DYKR1A; we observed co-purification of both TSC1 and TSC2 (Figure 2C). These results confirmed that DYRK1A physically interacts with the TSC complex. We then sought to identify the domains of DYRK1A that interact with TSC1 as TSC1 is the scaffolding protein in the TSC (20). We co-overexpressed HA3-TSC1 and various truncated forms of Flag-DYRK1A in HEK293 cells (Supplementary Figure 2). We found that interaction between DYRK1A and TSC1 was abolished in all the truncated forms that lack DYRK1A kinase domain, which suggests that DYRK1A kinase domain is required for the interaction (Supplementary Figure 2). We further confirmed that DYRK1A interacts with TSC1 through its kinase domain by co-expressing the FLAG tagged DYRK1A kinase domain and HA3-TSC1, followed by immunoprecipitation using anti FLAG antibodies (Figure 2D). Taken together, our results show that DYRK1A physically interacts with the TSC.

Next, we sought to test the implications of TSC and DYRK1A interaction on mTORC1 activity. mTORC1 is known to be induced in HEK293 cells by treatment with fetal bovine serum (FBS), insulin, and many growth factors following the serum starvation (21, 22, 41, 42). We performed similar experiments, where we knocked down DYRK1A followed by serum starvation and tested for levels of pS6K, a reporter for mTOR activity. We did not observe a significant reduction in basal phosphorylation levels on S6K and S6. However, 10 and 20 min after FBS treatment we observed significantly reduced levels of Phospho-S6K and Phospho-S6 in DYRK1A KD as compared to control HEK293 cells (Figure 3A-C). Further, we tested the levels of pS6K and pS6 in NIH3T3 after CRISPR-mediated targeting of Dyrk1a. Interestingly, we found significant reductions in both pS6K and pS6 as compared to controls (Figure 3 D-F). Overall, our results indicate that DYRK1A is required for promoting mTORC1 activity.

Next, we investigated if DYRK1A phosphorylates TSC1/TSC2, which is known to regulate mTORC1 activity. For this, we first knocked down DYRK1A in HEK293 cells, co-expressed HA3-TSC1 and Flag-TSC2 and performed immunoprecipitation followed by LC/tandem MS (MS/MS) to identify phosphorylation status of TSC1 and TSC2 in these cells. We observed that DYRK1A knockdown samples have reduced TSC2 phosphorylation at T1462 (T1462) compare to the control samples. T1462 has been previously identified as a site phosphorylated by various kinases that activate the mTORC1 pathway, including AKT (26), ERK (27), p90 ribosomal S6 kinase 1 (RSK1) (28), IκB kinase β (IKKβ) (43), and MAPKAPK2 (MK2) (24) in response to insulin and other growth factors. To ascertain that indeed TSC2 T1462 is the site phosphorylated by DYRK1A, we performed in-vitro kinase assays using purified wild-type and kinase-dead DYRK1A. As TSC2 is about 200KDa protein, we purified transiently overexpressed HA3-TSC1 and Flag-TSC2 from HEK293 cells and performed in vitro kinase assay on beads. Wild-type DYRK1A autophosphorylates and runs as fuzzy bands around 90-100Kda in polyacrylamide gels, whereas kinase-dead DYRK1A runs only as a single sharp band (Figure 4A). In our assay, we did not observe an increase in phosphorylation at S1387 or at Serine 939 in presence of ATP and DYRK1A (Figure 4A). However, immunopurified TSC2 is heavily phosphorylated at Serine 1387 and Serine 939, two important sites, phosphorylated by AKT (26, 42, 43). Therefore, we cannot rule out if the immunopurified TSC2 was fully phosphorylated at S1387 and at S939. However, it is important to note that we did not observe any differential loss of phosphorylation in TSC2 at S1387 in our mass spectrometry phosphorylation analysis. Interestingly, we found T1462 was strongly phosphorylated by wild-type DYRK1A, and not by kinase-dead DYRK1A (Figure 4A). Taken together, these experiments demonstrate that DYRK1A phosphorylates TSC2 on T1462. This phosphorylation is known to be an inhibitory modification and it is a key site that gets phosphorylated by various other key regulators of the mTORC1 pathway (43). We were not able to perform in-vitro tests to check if DYRK1A phosphorylates TSC1 or other sites on TSC2 due to the lack of appropriate antibodies. The overall data suggest that DYRK1A mediates TSC2 T1462 phosphorylation to inhibit TSC activity and, therefore, promotes TORC1 activity.

TSC phosphorylation has been shown to inhibit TSC, and activate mTORC1 via RHEB, where, the activation of TSC leads to the hydrolysis of GTP-bound RHEB, which in turn leads to mTORC1 inhibition (20). We hypothesized that DYRK1A-mediated phosphorylation of TSC2 may promote mTORC1 activation in a RHEB-mediated manner. Therefore, we reasoned that overexpression of RHEB, will result in the rescue of cell size phenotype caused by DYRK1A knockdown. We overexpressed RHEB in DYRK1A knockdown cells and found a partial rescue of cell size, suggesting that DYRK1A indeed plays a role in cell size regulation at least partly upstream to RHEB (Figure 4B and C). Taken together these data suggest that DYRK1A phosphorylates TSC2, which inhibits its action, leading to the activation of mTORC1 via RHEB.

Since the TSC-TOR pathway is conserved in Drosophila (44), we sought to test if the DYRK1A-mediated regulation of the TSC-TOR pathway is also conserved. The TOR pathway is shown to promote neuromuscular junction (NMJ) development in flies (45, 46) (Figure 5B,G). Similarly, mnb, the fly homolog of DYRK1A, is also required for NMJ growth (47); however, the mechanism of mnb-mediated NMJ growth is not known. Based on our finding that DYRK1A promotes TOR activity, we hypothesized that mnb promotes larval NMJ growth through TOR. To test this hypothesis, we systematically compared the NMJ growth phenotype of mnb mutant with that of loss or gain of TOR activity at larval NMJ. Since TOR mutants are early larval lethal, we expressed dominant-negative TOR (TOR.Ted) in motor neurons using the UAS-Gal4 system (D42Gal4>UAS-TOR.Ted) (48). To activate the TOR pathway, we overexpressed RHEB in motor neurons (D42Gal4>UAS-RHEB) (49). We stained NMJ using anti-HRP (to mark neurons) and anti-DLG (to mark boutons where synaptic connections are formed). As shown in Figure 5 (A, B, G), the loss of mnb causes a reduction in the NMJ bouton numbers compared to the Wild-type control (WT, Canton S). Similarly, the expression of TOR.TED in motor neurons (D42Gal4>UAS-Tor.Ted) also causes a reduction in NMJ bouton numbers (Supplementary Figure S3 D-F). In contrast, overexpression of mnb in motor neurons increases the bouton numbers (Figure 4 C, D, H). Similarly, the activation of the TOR pathway by overexpression of RHEB in motor neurons or mutations in gigas (gig, the fly homolog of TSC2) increases the number of boutons (Fig 5 E, F, I, Supplementary Figure S3 C). The similarities in the TOR and mnb phenotypes suggests that Mnb may positively regulate TOR activity and promote NMJ growth. To test this hypothesis, we asked whether or not the NMJ phenotype of the mnb mutant is rescued by the activation of TOR. As shown in figure 5L, overexpression of RHEB rescues the NMJ bouton numbers in mnb mutants. These results indicate that RHEB functions downstream to mnb to regulate the NMJ development. Taken together, the results in mammalian cells and in flies suggest that the role of DYRK1A/Mnb in the regulation of the TOR pathway is conserved.

pcDNA3-HA3-TSC1 (Yue Xiong; Addgene #19911), pcDNA3 Flag TSC2 (Brendan Manning: Addgene #14129), LentiCRISPR v2 (Feng Zhang; Addgene #52961) plasmids were procured from Addgene. For expressing cDNA using lentivirus, LentiCRISPR v2 was modified by removing the U6 promoter and tracer RNA, and cDNA were cloned in place of SpCas9, and named LentiExp vector. RHEB sequence from pRK7-RHEB (John Blenis; Addgene #15888) was cloned into the LentiExp vector by replacing SpCas9. Expression was confirmed by western blot analysis. DYRK1A expression vector, DYRK1A shRNA, and control shRNA have been previously described (18). All constructs were confirmed by DNA sequencing. Expression of Flag-tagged proteins was induced with 250ng/ml Doxycycline for 36 to 48 h. HEK293, HEK293T, NIH3T3 cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% Fetal Bovine Serum (FBS). All cell lines were obtained from the American Type Culture Collection (ATCC). All cell lines were maintained at 37°C under 5% CO₂.

Antibodies against Actin (AC026) and HA (AE008) were from Abclonal, anti-DDDDK from MBL (M185-3L), TSC1 (4906), TSC2 (4308), S6K (2708), pS6K (9205), S6 (2317), pS6 (4856) were from CST. DYRK1A antibody was generated in house and has been described before (18).

For immunoblotting, total cell extracts were prepared. Briefly, collected cells were washed once with cold PBS and lysed in cold NP-40 lysis buffer (40mM HEPES, pH 7.4, 120mM NaCl, 1mM EDTA, 1% NP-40 [Igepal CA-630], 5% glycerol, 10 mM sodium pyrophosphate, 10 mM glycerol 2-phosphate, 50 mM NaF, 0.5 mM sodium orthovanadate and 1:100 protease inhibitors [Solarbio, P6730]) for 15 min on ice. Lysates were cleared by centrifugation, and protein concentration was measured using BCA kit. Western blot signals were quantified using ImageJ.

Short hairpin RNA (shRNA) targeting DYRK1A have been previously described (18). Lentiviral particle preparation and infection were performed as previously described (18). Lentivirus-infected HEK293 and SH-SY5Y cells were selected with 1 μg/mL(HEK293) or 2 μg/mL (SH-SY5Y) Puromycin respectively for four days.

For RHEB overexpression rescue of DYKR1A shRNA-mediated knockdown HEK293 cells, cells were first infected with lentivirus expressing shRNAs targeting control and DYRK1A, and then selected with 1μg/mL Puromycin for 48 hours. After selection cells were transduced with lentivirus expressing Flag-RHEB. These cells were further selected with 2μg/mL Puromycin for 48 hours before harvesting.

shRNA treated HEK293 cells selected for 4 days with 1ug/ml puromycin were washed once with PBS and then DMEM without serum added. Serum starvation was performed for 12 hours. For stimulation with FBS (10%), pre-warmed FBS was directly added to the serum-free media for the time periods indicated.

NIH3T3 cells were transduced with lentiCRISPR v2 (Addgene plasmid #52961) containing either no sgRNA (control) or sgRNA targeting Dyrk1a (57). Transduced cells were selected with 3μg/mL Puromycin for 4 days before harvesting. sgRNA targeting mouse Dyrk1a is provided in supplemental table S1.

Cell size was measured and analyzed with a JIMBIO FIL electronic cell counter following the manufacturer’s protocol. In short, 2.5×10⁵ cells were seeded in 6-well plates. After 24 hours of culture, cells were trypsinized and taken up in full DMEM. Each cell line was measured three times; each measurement comprised three cycles of cell counts with intervals of 1 μm ranging from 4 to 30 μm. The sum of counts of viable cells in the range of 12 to 30 μm was plotted and quantified. Three biological replicates were performed per cell type (i.e., HEK293, SH-SY5Y and NIH3T3 cells [KD/KO and corresponding Control]). Control and corresponding KD/KO were compared with multiple unpaired t-tests. p values are presented as stars above the corresponding bar graphs; *, p < 0.05. HEK293 Flag-DYRK1A cell size was measured as above, except cells were induced with 40ng/mL Doxycycline for 48 hours.

2.5×10⁵ cells were seeded in 6-well plates. After 48 hours of 40ng/mL Doxycycline induction, total RNA from fresh cells were extracted using RNA-easy Isolation Reagent (Vazyme) as per the manufacturer’s instructions. After determining the total RNA concentration and quality using a Nanodrop (Thermofisher), 1 μg RNA was used in 20 μL reaction mixture to reverse transcribe RNA using HiScript II 1st Strand cDNA Synthesis Kit (Vazyme R212) and stored at −20°C. 2μL was used to perform qPCR using ChamQ Universal SYBR qPCR Master Mix (Vazyme Q711) with 10 μM forward and reverse gene specific primers. The cycling consisted of 2 min at 95°C, followed by 40 cycles of 5 s at 95°C and 10 s at 60°C. Following completion of the final cycle, a melting curve analysis was performed to monitor the purity of the PCR products. Each sample was analyzed in triplicate. Quantification was performed by means of the comparative Ct method (2^-ΔΔCT). The mRNA expression of target genes was normalized to that of GAPDH, and the data are represented as the mean ± SEM of biological replicates.

Protein precipitation and digestion: Proteins were precipitated with TCA (trichloroacetic acid). The protein pellet was dried in Speedvac. The pellet was subsequently dissolved with 8 M urea in 100 mM Tris-Cl (pH 8.5). TCEP (final concentration is 5 mM; Thermo Scientific) and Iodoacetamide (final concentration is 10mM; Sigma) were added to the solution and incubated at room temperature for 20 and 15 minutes for reduction and alkylation respectively. The solution was diluted four times and digested with Trypsin at 1:50 (w/w; Promega).

LC/tandem MS (MS/MS) analysis of peptide: The peptide mixture was loaded on a home-made 15 cm-long pulled-tip analytical column (75 μm i.d.) packed with 3μm reverse-phase beads (Aqua C18, Phenomenex, Torrance, CA) connected to an Easy-nLC 1000 nano HPLC (Thermo Scientific, San Jose, CA) for mass spectrometry analysis. Data-dependent tandem mass spectrometry (MS/MS) analysis was performed with a Q Exactive Orbitrap mass spectrometer (Thermo Scientific, San Jose, CA). Peptides eluted from the LC system were directly electrosprayed into the mass spectrometer with a distal 1.8-kV spray voltage. One acquisition cycle includes one full-scan MS spectrum (m/z 300-1800) followed by top 20 MS/MS events, sequentially generated on the first to the twentieth most intense ions selected from the full MS spectrum at a 27% normalized collision energy. MS scan functions and LC solvent gradients were controlled by the Xcalibur data system (Thermo Scientific).

With detergent: 1×10⁶ cells were seeded in 10 cm plates. After 24 hours of culture, cells were co-transfected with 5μg pcDNA-Flag-DYRK1A/deletion and 5μg HA-TSC1 using PEI. After 48h, collected cells were washed once with cold PBS, and lysed in cold NP-40 lysis buffer (40mM HEPES, pH 7.4, 120mM NaCl, 1mM EDTA, 1% NP-40 [Igepal CA-630], 5% glycerol, 10 mM sodium pyrophosphate, 10 mM glycerol 2-phosphate, 50 mM NaF, 0.5 mM sodium orthovanadate and 1:100 protease inhibitors [Solarbio, P6730]) for 15 min on ice. Lysates were cleared by centrifugation and protein concentration was measured using BCA kit. Supernatant was then incubated with Flag beads (Smart-Lifesciences, SA042005) for ~4 h at 4°C. Finally, beads were washed with a lysis buffer without NP-40 three times and resuspended in a 1× SDS loading buffer. Samples were heated for 10 min at 95°C and separated by SDS-PAGE.

Approximately 10⁸ HEK293-Flag-DYRK1A (with 250ng/ml Doxycycline treatment) and parental cells were induced for 48h were collected and washed with PBS. Cells were swollen for 15 min in hypotonic buffer (Buffer A: 10 mM HEPES pH 7.9, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM DTT supplemented with freshly prepared protease inhibitors, Solarbio, P6730). Swollen cells were dounced 20 times in a Wheaton Dounce homogenizer till about 90% cells were lysed (as observed in a microscope). Lysates were then centrifuged at 5000-6000 g for 10 min at 4°C. Supernatant was transferred to a new tube and 0.11 volume of Buffer B (0.3M HEPES pH 7.9, 1.4M KCl, 0.03M MgCl2) was added; the supernatant was then centrifuged at 12000g for 30 min. Resulting supernatant was transferred to a new tube and incubated with Flag-beads (Smart-Lifesciences; SA042005) for ~4 h at 4°C. Finally, beads were washed three times with wash buffer containing 150 mM NaCl and 0.2% Triton X-100 and resuspended in a 1× SDS loading buffer and analyzed by SDS-PAGE.

GST-DYRK1A and GST-DYRK1A K188R were cloned in pGEX-6p1vector, proteins were expressed in BL21 strain and purified using GST beads. FLAG-TSC2 and HA3-TSC1 were co-expressed in HEK293 cells, and Flag-Affinity purified. After wash, beads were rinsed with 1× kinase assay buffer (25 mM HEPES, pH 7.0, 5 mM MgCl2, 0.5 mM DTT), and kinase assay was performed on beads bound with FLAG-TSC2 and HA-TSC1. Equal amounts of GST-DYRK1A or GST-K188R were added to the Flag beads in presence/absence of ATP (100μM), and incubated at 30°C for 30 min. Reactions were stopped by addition of 5× SDS loading dye and analyzed by SDS PAGE.

Flies were grown in standard fly food at 25°C. Following genotypes were used in this study: Canton S (Bloomington Stock Center), mnb[1] (FBal0012364, (11), gift from Francisco J. Tejedor), UAS-mnb (FBtp0114512, (58), gift from Francisco J. Tejedor), UAS-RHEB (FBst0009688, Bloomington Stock Center), D42-Gal4 (FBti0002759, Bloomington Stock Center (49), P[(45)w[+mC]=UAS-Tor.TED]II (FBti0026636, (59), Bloomington Stock Center), UAS-mCherry (FBti0147460, Bloomington Stock Center).

For Immunofluorescence staining of NMJ the third instar larvae were fixed in 4% paraformaldehyde (Himedia Cat#TCL119) for 20 min at room temperature and washed in 1X PBST (0.2% Triton X-100). Antibodies were used at the following dilutions: Mouse anti-DLG 1:500 (DSHB 4F3, (59), Rabbit anti-HRP conjugated with alexa488 (dilution 1:500, Jackson 123-545-021), Goat anti-mouse conjugated with Alexa 555 (Dilution 1:500, Invitrogen A28180). Samples were imaged with confocal microscope (Leica Stellaris 5 (PL APO 40X/1.30 oil objective) or Olympus FV3000 (UPLFLN 40X/1.30 oil objective). Images were documented and quantified using ImageJ. For the quantification of NMJ phenotypes, we marked the HRP labeled boutons and manually counted the number. The number of boutons were then normalized to the muscle area (muscle 6/7) and the normalized data were compared between genotypes.

GraphPad Prism version 7.00 was used for statistical and statistical presentation of quantitation. p values are presented in the figures above or below the compared conditions. Two-way ANOVA followed by a Sidak’s multiple comparisons test was applied to cell size data, p-S6 and p-S6k. Unpaired two-tailed Student’s t test was applied to DYRK1A protein expression. The data are represented as the mean ± SEM of three biological replicates.

Mass spectrometry data files for FLAG-DYRK1A affinity purifications and negative controls are available from Massive at /ftp://massive.ucsd.edu/MSV000085815/and ProteomeXchange (PXD020533). Original mass spectrometry data underlying this manuscript can be accessed after publication from the Stowers Original Data Repository at http://www.stowers.org/research/publications/libpb-1722.