Misfolded proteins in the endoplasmic reticulum (ER) are removed through multistep processes termed ER-associated degradation (ERAD). Valosin-containing protein (VCP) plays a crucial role in ERAD as the interaction of ubiquitin fusion degradation protein 1 (Ufd1) with VCP via its SHP box motif (228F-S-G-S-G-N-R-L235) is required for ERAD. However, the mechanisms by which the VCP–Ufd1 interaction is regulated are not well understood. Here, we found that the serine 229 residue located in the Ufd1 SHP box is phosphorylated in vitro and in vivo by cyclic adenosine monophosphate-dependent protein kinase A (PKA), with this process being enhanced by either forskolin (an adenylyl cyclase activator) or calyculin A (a protein phosphatase inhibitor). Moreover, a phosphomimetic mutant (S229D) of Ufd1 as well as treatment by forskolin, calyculin A, or activated PKA strongly reduced Ufd1 binding affinity for VCP. Consistent with this, the Ufd1 S229D mutant significantly inhibited ERAD leading to the accumulation of ERAD substrates such as a tyrosinase mutant (C89R) and 3-hydroxy-3-methylglutaryl coenzyme A reductase. However, a non-phosphorylatable Ufd1 mutant (S229A) retained VCP-binding ability and was less effective in blocking ERAD. Collectively, our results support that Ufd1 S229 phosphorylation status mediated by PKA serves as a key regulatory point for the VCP–Ufd1 interaction and functional ERAD.
Most newly synthesized membrane and secretory proteins undergo a transition from a nascent polypeptide chain into the folded form in the endoplasmic reticulum (ER) . However, although cells have evolved to adopt highly complex processes that ensure relevant protein folding and secretion [1,2], intrinsic errors in protein synthesis inevitably lead to protein misfolding, which occurs more frequently under various ER stress conditions [2,3]. In such instances, several signaling pathways involving protein kinase RNA-like ER kinase, inositol-requiring enzyme 1, and activating transcription factor 6 are activated as sensors, serving to initiate the unfolded protein response to facilitate protein folding or, ultimately, trigger apoptosis [1,4]. Accordingly, the ER stress/unfolded protein response pathway is closely associated with the pathological features of diverse diseases . In addition, unfolded and misfolded proteins are eliminated through the ubiquitination system-linked cellular response termed ER-associated degradation (ERAD) [2,3,6]. The ERAD pathway functions as an essential route for relieving the burden imposed by structurally aberrant proteins, thereby maintaining protein quality control in the ER [2,6,7].
For ERAD, misfolded proteins are processed through multiple steps that include recognition, retrotranslocation, polyubiquitination, and proteasomal degradation [3,7,8]. Various proteins are spatiotemporally engaged in each step of the ERAD pathway to promote its progression [7,8]. In particular, valosin-containing protein (VCP, also termed p97), the mammalian homolog of Cdc48 in yeast, plays a key role in ERAD by extracting protein substrates out of the ER and transporting them to the cytoplasmic proteasome for ubiquitin-dependent degradation [9–11]. VCP belongs to the type II AAA+ (ATPases associated with diverse cellular activities) family and has a homohexameric conformation [11,12]. VCP-driven ATP hydrolysis is believed to induce a mechanical force for dislocating ERAD substrates [13,14]. Furthermore, VCP is also recognized as a critical regulator of Golgi reassembly, membrane fusion, autophagy, endocytosis, cell cycle, and DNA damage response [15,16], which is consistent with the observation that signaling proteins such as hypoxia-inducible factor 1-α  and inhibitor of κB-α  are subject to ERAD. In turn, VCP exerts its pleiotropic functions through the formation of protein complexes with specific binding cofactors. To date, a wide range of proteins have been identified as VCP cofactors. These contain various but distinct VCP-binding domains, such as ubiquitin regulatory X (UBX), UBX-like, and peptide N-glycosidase/ubiquitin-associated), or VCP-binding motifs, such as the SHP box (named after the first identified Shp1, the yeast ortholog of p47), VCP-interacting motif (VIM), and VCP-binding motif, which consist of a relatively short length of amino acids [17,19–23].
Among the numerous VCP-binding partners, ubiquitin fusion degradation protein 1 (Ufd1) and nuclear protein localization protein 4 (Npl4) are the most well-established VCP cofactors for ERAD [10,24–27]. Ufd1 and Npl4 bind to VCP via the SHP box motif (previously termed binding site 1) and the UBX-like domain, respectively. The Ufd1–Npl4 heterodimer recognizes VCP via a bipartite mechanism, thereby forming a ternary complex of VCP–Ufd1–Npl4 . In accordance with this, protein structural determinations of the VCP–Ufd1 complex performed by ourselves and others revealed that the Ufd1 SHP box recognizes a unique region of the VCP N-terminal domain that is spatially distinct from the regions to which the Npl4 UBX-like domain and the VIM motif bind [28,29].
Nevertheless, relatively little is known regarding the regulation of the VCP–Ufd1 interaction. Considering that a conserved serine or threonine is present within the SHP box motif of VCP cofactors, we hypothesized that phosphorylation at this site might modulate Ufd1 binding to VCP. In this study, we assessed the ability of the cyclic adenosine monophosphate (cAMP)-dependent protein kinase A (PKA) pathway based on the presence of a PKA consensus motif in the SHP box motif to effect Ufd1 phosphorylation and the potential role of this phosphorylation in regulating the VCP–Ufd1 interaction and functional ERAD.
Reagents and antibodies
Most chemicals including Dulbecco's modified Eagle's medium, bovine serum albumin, paraformaldehyde, poly-l-lysine, cycloheximide (CHX), MG132, anti-FLAG M2 affinity agarose gels, and antibodies against α-tubulin (T5168), FLAG (F1804), and β-actin (A5316) were obtained from Sigma–Aldrich (St. Louis, MO, U.S.A.). Lipofectamine 2000, Lipofectamine RNAiMAX, Opti-MEM I, anti-V5 antibody (R960-25), goat serum, and Alexa Fluor-conjugated secondary antibodies were purchased from Thermo Fisher Scientific (Waltham, MA, U.S.A.). H-89, forskolin, calyculin A, FK506, okadaic acid, and anti-calreticulin antibody (ab4109) were obtained from Abcam (Cambridge, MA, U.S.A.). Antibodies against HA (#3724), Myc (#2278), and phospho-(Ser/Thr) PKA substrate (#9621) were purchased from Cell Signaling Technology (Danvers, MA, U.S.A.). Antibodies against VCP (sc-20799), Ufd1 (611642), phospho-serine (05-1000), and glutathione S-transferase (GST) (27-4577-01) were purchased from Santa Cruz Biotechnology (Dallas, TX, U.S.A.), BD Biosciences (San Diego, CA, U.S.A.), Merck Millipore (Billerica, MA, U.S.A.), and GE Healthcare (Princeton, NJ, U.S.A.), respectively.
The plasmids of human UFD1 and VCP cloned into p3XFLAG-CMV-10 and pcDNA3-HA expression vectors, respectively, were previously described . The S229D, S229A, S231A, and S231D mutations were introduced into the FLAG-Ufd1 template using the QuikChange II Site-Directed Mutagenesis Kit (Agilent Technologies, Santa Clara, CA, U.S.A.) according to the manufacturer's protocol. The wild-type (WT) and mutated UFD1 inserts were subcloned into the BamHI-XhoI sites of a pGEX-6P-1 vector to generate GST-tagged plasmids. Plasmids pcDNA5FRT/TO-DVC1-Strep-HA (#113481) and pcDNA5FRT/TO-p47-Strep-HA (#113475) were gifts from Hemmo Meyer (Addgene, Watertown, MA, U.S.A.) . The S254D mutation of DVC1 and T254D mutation of p47 were generated using the QuikChange II Site-Directed Mutagenesis Kit. The specific polymerase chain reaction primers for the point mutations were obtained from Bionics (Seoul, Republic of Korea) and the mutated and recombinant plasmids were confirmed by DNA sequencing (Bionics). Expression plasmids of V5-VCP and FLAG- and V5-3-hydroxy-3-methylglutaryl coenzyme A reductase (HMGCR) were provided by Ho Chul Kang (Ajou University, Suwon, Republic of Korea). The tyrosinase (TYR) C89R mutant cloned into the pcDNA3.1-Myc vector was a gift from Petek Ballar (Ege University, Turkey) . The plasmid containing the HA-tagged PKA catalytic subunit was provided by Pietro De Camilli (Yale University, New Haven, CT, U.S.A.). All plasmids were purified using an EndoFree Plasmid Maxi Kit (Qiagen, Hilden, Germany).
Human cell culture, transfection, and chemical treatment
Human embryonic kidney 293T (HEK293T) cells and HeLa cells (human cervical carcinoma) were purchased from the American Type Culture Collection (Manassas, VA, U.S.A.) and grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and penicillin/streptomycin (HyClone, Logan, UT, U.S.A.) at 37°C in a humidified atmosphere of 5% CO2 and 95% air and were routinely subcultured at 2- or 3-day intervals. For transient transfection, the indicated Ufd1 and/or VCP expression plasmids or corresponding empty vectors were mixed with Lipofectamine 2000 in Opti-MEM I, then the mixtures were added to 70–80% confluent cell cultures for 1 day. Cells were treated with dimethyl sulfoxide as a vehicle or various chemicals including forskolin, H-89, FK506, okadaic acid, or calyculin A as indicated at 24 h post-transfection.
Immunoblotting (IB) and immunoprecipitation (IP)
Cells were harvested in cold lysis buffer (50 mM Tris pH 7.4, 150 mM NaCl, 1 mM ethylenediaminetetraacetic acid (EDTA), 1 mM egtazic acid (EGTA), 1 mM dithiothreitol (DTT), 1 mM Na3VO4, 5 mM NaF, and 1% Triton X-100) containing protease and phosphatase inhibitor cocktails (GenDEPOT, Barker, TX, U.S.A.) and placed on ice for 20 min with occasional vortexing. After clearance by centrifugation (15 000×g, 30 min, 4°C), the protein concentration in the cell lysates was determined using bicinchoninic acid protein assay reagents and bovine serum albumin as a protein standard (Pierce, Rockford, IL, U.S.A.). For IP of FLAG-Ufd1 proteins, cell lysates (1.0–1.5 mg) were mixed with 20 µl anti-FLAG M2 affinity gels for 4 h at 4°C. Alternatively, for IP of HA- and V5-tagged VCP proteins, cell lysates (2.0 mg) were incubated with 5.0 µg of anti-HA and -V5 antibody, respectively, for 4 h at 4°C, and then with 40 µl Protein A/G PLUS-Agarose IP reagent (Santa Cruz Biotechnology) for an additional 2 h . Similarly, cell lysates were processed for endogenous Ufd1 IP with an Ufd1-specific antibody or normal IgG (Santa Cruz Biotechnology) as a negative control. The immune complexes were washed with the cell lysis buffer five times and the starting cell lysates and IP products in Laemmli sample buffer were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS–PAGE) on 8–10% resolving gels and transferred to Immobilon-P polyvinylidene difluoride membranes (Merck Millipore). Following blocking with 5% nonfat milk in Tris-buffered saline containing 0.1% Tween-20 (TBST), membrane blots were incubated with primary antibodies against FLAG (1 : 2000), HA (1 : 1000), Myc (1 : 1000), V5 (1 : 1000), VCP (1 : 1000), GST (1 : 2000), phospho-(Ser/Thr) PKA substrate (1 : 500), β-actin (1 : 1000), Ufd1 (1 : 1000), phospho-serine (1 : 500), or α-tubulin (1 : 2000) for 2 h at room temperature, washed three times with TBST, and then incubated with horseradish peroxidase-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories, West Grove, PA, U.S.A.) for 1 h. The resulting immune complexes were detected using the SuperSignal West Pico chemiluminescent substrate (Pierce).
GST fusion protein pull-down assay
The N-terminal GST-tagged plasmids of Ufd1 were expressed in the Escherichia coli strain BL21 (Real Biotech Corporation, Banqiao City, Taiwan) using the supplier's non-heat shock transformation protocol. Single colonies were grown in Luria–Bertani broth containing ampicillin (100 µg/ml) and chloramphenicol (50 µg/ml) and GST-Ufd1 fusion proteins were induced with 0.1 mM isopropyl β-d-thiogalactopyranoside (Merck Millipore) for 6 h at 30°C. The E. coli cells (∼0.5 g pellet wet weight) were harvested by centrifugation (5000×g, 20 min, 4°C) and resuspended in 5 ml lysis buffer (phosphate-buffered saline (PBS) containing protease inhibitors and 1 mM DTT). After incubation with 1.0% Triton X-100 for 30 min at 4°C, the cell suspensions were lysed by sonication (5 min, on ice) and then clarified by centrifugation (18 000×g, 30 min, 4°C). To affinity purify the GST fusion proteins, 0.1 ml of glutathione Sepharose 4B beads (GE Healthcare) were washed twice with PBS supplemented with 0.5% Triton X-100 and mixed with 1 ml of the bacterial extracts for 2 h at 4°C . After washing five times with the lysis buffer, the glutathione beads were further incubated with HEK293T cell lysates (1.0–2.0 mg) for 4 h at 4°C and then washed five times with the cell lysis buffer. Approximately 30 µl of the resulting beads were analyzed by SDS–PAGE and IB.
In vitro phosphorylation
The fully active PKA catalytic subunit (∼0.25 µg) purchased from New England Biolabs (Ipswich, MA, U.S.A.) was mixed with purified GST-Ufd1 fusion proteins (2.0 µg). Ponceau S staining was used for detection of purity and amounts of GST fusion proteins. The phosphorylation reactions were performed in a buffer (50 mM Tris–HCl, 10 mM MgCl2, 0.2 mM ATP, 0.1 mM EDTA, 2 mM DTT, and 0.01% Brij 35, pH 7.5) according to the manufacturer's protocol. After 1-h incubation at 30°C, reaction mixtures (50 µl) were stopped by the addition of Laemmli sample buffer and then analyzed by SDS–PAGE and IB using the phospho-(Ser/Thr) PKA substrate antibody (1 : 500).
Stoichiometry measurement of Ufd1 phosphorylation
Ufd1 phosphorylation efficiency by PKA was determined by indirect measurements of free orthophosphate. Firstly, GST-Ufd1 fusion protein conjugated to glutathione beads was phosphorylated by PKA in in vitro condition as described above. For accurate measurements, increased GST-Ufd1 substrate amount (24.0 µg) that corresponds to 0.34 nmol based on the 70 kDa molecular mass were used for phosphorylation reaction. An equivalent amount of GST alone was used as a control. After 1 h reaction, the glutathione beads were washed three times with 1 ml TBST, and then incubated with 15 units of calf-intestinal alkaline phosphatase (New England Biolabs) in a volume of 50 µl following the manufacturer's protocol. After the dephosphorylation reaction for 1 h at 37°C, the mixtures were centrifuged and supernatants (20 µl) were carefully transferred into 96-well and 80 µl of malachite green assay reagent (Sigma–Aldrich, #MAK307) was added to the supernatant samples. After further incubation for 1 h at room temperature, the absorbance of color changes (from yellow to green) due to generated free orthophosphate was measured at 620 nm using a microplate reader (model VMAX, Molecular Devices, San Jose, CA, U.S.A.) according to the manufacturer's protocol. The orthophosphate concentration in the samples was determined from a standard curve obtained with serially diluted (0–30 µM) orthophosphate solution. The orthophosphate values of GST alone samples were subtracted from those of GST-Ufd1 samples.
Cells were plated on 60-mm dishes and transfected with FLAG-Ufd1 together with the Myc-TYR-C89R or FLAG-HMGCR as an ERAD substrate (total 6 µg at a 1 : 1 ratio) or their relevant empty vectors. Cell lysates were prepared as described above after 1 day and the protein levels of transfected proteins were examined by IB analysis . The cotransfected cells were left untreated or treated with CHX (50 µg/ml) for up to 2 h for CHX chase experiments.
Ufd1 gene knockdown
For Ufd1 knockdown, double-stranded small interfering RNA (siRNA) targeting human Ufd1 sequence (5′-GUG GCC ACC UAC UCC AAA UUU-3′) or control siRNA (5′-UUC UCC GAA CGU GUC ACG UUU-3′) from Bioneer (Daejeon, Republic of Korea) were mixed with Lipofectamine RNAiMAX in Opti-MEM I, and the mixtures were added to cells for 2 days.
Immunostaining and cell imaging
HeLa cells were plated onto 18-mm circular coverslips coated with poly-l-lysine and transfected for 1 day with the indicated plasmids. As a marker for ER, anti-calreticulin immunostaining was used. Cells were then fixed with 4% paraformaldehyde for 20 min, permeabilized with PBS containing 0.1% Triton X-100 for 15 min, and blocked for 30 min with PBS containing 10% bovine serum albumin and 5% goat serum. The coverslips were stained for 2 h with anti-FLAG (1 : 200), -Myc (1 : 100), and/or -calreticulin (1 : 100) antibodies diluted in the blocking buffer in a humidified air-tight container, followed by staining with Alexa Fluor-conjugated secondary antibodies (1 : 200) for 1 h at room temperature. Cells were mounted using Prolong Gold Anti-fade Reagent (Thermo Fisher Scientific) and fluorescence images were captured with a Zeiss LSM 710 confocal microscope (Carl Zeiss GmbH, Jena, Germany) as previously described .
All experiments were performed independently at least three times with similar results. The intensities of immunoblots were measured using ImageJ software (National Institutes of Health, Bethesda, MD, U.S.A.). Data shown in the graphs are presented as the means ± SEM. The statistical significance of the data was determined using a one-way analysis of variance with multiple comparison tests using GraphPad Prism software (La Jolla, CA, U.S.A.).
Phosphorylation of the Ufd1 SHP box at serine 229 (S229) by PKA
The VCP-binding cofactor Ufd1 comprises a ubiquitin-binding UT3 domain and a VCP-binding SHP box motif at the N- and C-terminus, respectively (Figure 1A) [24,28,29]. VCP is composed of an N-terminal (N) domain, to which most VCP cofactors bind, and two catalytic ATPase domains (referred to as D1 and D2). The SHP motif specifically recognizes a C-terminally located lobe (Nc) of the N domain to form the VCP–Ufd1 complex (Figure 1A) [28,29]. The SHP box motif, whose consensus sequence is F-X-G-X-G-X-X-L (where X indicates any amino acid), is highly conserved among human, mouse, and rat Ufd1 protein sequences and is also present in DVC1 (also termed Spartan or C1orf124) and several UBX domain-containing proteins including p47, ASPL (also termed UBXD9), and UBXD4 (Figure 1B) [21,28]. Notably, the SHP box motifs of these proteins contain a well-conserved serine or threonine residue, which corresponds to the S229 in Ufd1 although other SHP box motif-containing proteins such as yeast Ufd1 and human UBXD5 do not harbor the conserved serine or threonine residue (Figure 1B). This suggested that S229 phosphorylation of Ufd1 may have a role in regulating its interaction with VCP. Moreover, several web-based searches for Ufd1 phosphorylation sites (i.e. Scansite (http://scansite4.mit.edu), NetPhos (www.cbs.dtu.dk/services/NetPhos), and PhosphoSitePlus (www.phosphosite.org)) predicted a considerable probability of phosphorylation at S229.
Serine 229 in the Ufd1 SHP box motif is phosphorylated by PKA.
In Ufd1, an arginine (R226) is located at the −3 position; the resulting R-X-X-S/T motif matches a known target for PKA [34–38]. PKA also exhibits high phosphorylation efficiency toward the R-R-X-S/T motif [39,40], although this is lacking in Ufd1. Hence, we first tested whether PKA could phosphorylate Ufd1 at S229 in vitro using a recombinant active PKA (catalytic subunit only) protein and purified GST fusion protein of Ufd1 (Figure 1C). The S229 of GST-Ufd1 was also mutated to a non-phosphorylatable alanine and the point mutant (S229A) was compared with the GST-Ufd1 WT with regard to PKA phosphorylation efficiency. We measured changes in Ufd1 phosphorylation levels by IB using a phospho-(Ser/Thr) PKA substrate antibody. This antibody specifically recognizes phosphorylated serine or threonine in the R-X-X-S/T motif and thus has been widely used for the detection of PKA substrates [34–36,38]. As shown in the western blot, PKA strongly phosphorylated GST-Ufd1 WT but was unable to phosphorylate its S229A mutant (Figure 1C). GST alone showed no immunoreactivity, confirming the specificity of the Ufd1 phosphorylation. In addition, a band was observed in the incubation samples containing the PKA catalytic subunit when detected using the anti-phospho-PKA substrate (P-PKA sub.) antibody, which was considered to reflect autophosphorylation of PKA.
We attempted to determine PKA-mediated Ufd1 phosphorylation efficiency by employing indirect measurements of free orthophosphate (Pi) generated from phosphorylated Ufd1. For this purpose, we firstly prepared phosphorylated GST-Ufd1 protein with PKA under the in vitro condition and then induced its dephosphorylation by alkaline phosphatase treatment as described in the Experimental section. The released Pi following the dephosphorylation reaction was then measured using malachite green phosphate assay that is an acceptable and reliable method for Pi detection . Our results indicated that ∼65% of GST-Ufd1 protein underwent phosphorylation, suggesting that Ufd1 is a sensitive PKA substrate. The S229 in Ufd1 turned out to be a prominent site to be phosphorylated by PKA in vitro (Figure 1C), thus most of Pi is likely to be liberated from the phosphorylated S229 residue. To test whether Ufd1 S229 phosphorylation might be supported in cultured cells, HEK293T cells were cotransfected with FLAG-tagged Ufd1 WT or S229A mutant in the absence and presence of the HA-tagged PKA catalytic subunit (a constitutively active form). FLAG-Ufd1 proteins were purified by IP using anti-FLAG affinity gels and Ufd1 phosphorylation status was analyzed by IB with the anti-phospho-PKA substrate antibody. The transfected FLAG-Ufd1 and HA-PKA proteins in the IP and/or cell lysate samples showed ∼45 kDa and 40 kDa bands, respectively, as expected from their molecular sizes. FLAG-Ufd1 WT, but not its S229A mutant, was clearly phosphorylated by the HA-PKA catalytic subunit (Figure 1D), which was consistent with the in vitro phosphorylation results (Figure 1C). IB analysis of cell lysates and FLAG IP samples with anti-FLAG and/or anti-HA antibodies confirmed comparable expression levels of the transfected proteins and IP efficiency. These results indicated that Ufd1 represents a novel substrate of PKA and the S229 residue in the SHP box serves as a major phosphorylation site for PKA.
Regulation of cellular Ufd1 S229 phosphorylation by the PKA pathway and a calyculin A-sensitive protein phosphatase
We next examined the feasibility of Ufd1 S229 phosphorylation by the endogenous PKA signaling pathway. As cellular PKA activity is dependent on the levels of cAMP, which is produced by adenylyl cyclase , forskolin, a known agonist of adenylyl cyclase, was used to induce PKA activation . Changes in Ufd1 phosphorylation levels were monitored by FLAG IP and IB analysis using the anti-phospho-PKA substrate antibody following FLAG-Ufd1 transfection. PKA-dependent phosphorylation of Ufd1 was augmented by stimulation with 20 µM forskolin for 1 h (Figure 2A). Conversely, 1 h treatment with 20 µM H-89, an inhibitor of PKA , reduced the Ufd1 phosphorylation (Figure 2B). These results supported the physiological relevance of Ufd1 phosphorylation by endogenous PKA. In addition, cells transfected with FLAG-Ufd1 WT or S229A mutant were stimulated with forskolin (20 µM) for 0, 15, 30, and 45 min to ascertain that the forskolin-induced Ufd1 phosphorylation by PKA occurs on the S229 residue. Consistent with the in vitro findings, Ufd1 phosphorylation was detected in FLAG IP products from WT-expressing cells 45 min after forskolin stimulation, whereas no signal was detectable in products from S229A-expressing cells (Figure 2C).
Cellular Ufd1 S229 phosphorylation is reversibly modulated by PKA and a calyculin A-sensitive protein phosphatase.
As protein phosphorylation can alternatively be induced by suppression of protein phosphatases (PPs), we also evaluated the effects of several inhibitory chemicals of PPs on Ufd1 S229 phosphorylation. For example, FK506 is an inhibitor of the calcium/calmodulin-dependent PP (PP) 2B (PP2B, also termed calcineurin)  whereas okadaic acid and calyculin A both inhibit the serine/threonine PPs PP1 and PP2A [46,47]. When FLAG-Ufd1 WT-expressing cells were treated with FK506 (10 µM), okadaic acid (100 nM), or calyculin A (20 nM) for 30 min, only calyculin A increased PKA-dependent Ufd1 phosphorylation (Figure 2D). In contrast, this calyculin A effect was not observed in FLAG-Ufd1 S229A-expressing cells (Figure 2D). As calyculin A is a more potent PP1 inhibitor than okadaic acid , this result suggested that certain types of calyculin A-sensitive PPs such as PP1 may reversibly regulate the phosphorylated status of Ufd1 S229.
Inhibitory effect of Ufd1 S229 phosphorylation on its interaction with VCP
To determine whether the PKA-dependent Ufd1 phosphorylation might affect the interaction with VCP, the major binding partner of Ufd1, HEK293T cells were transfected with FLAG-Ufd1 and/or the HA-PKA catalytic subunit and then endogenous VCP levels in the FLAG IP products were analyzed by IB. VCP in the IP and input samples showed a molecular mass of ∼100 kDa, matching with the expected 97 kDa protein size. The anti-VCP immunoblot revealed that VCP coprecipitation with FLAG-Ufd1 was relatively weakened by the presence of the HA-PKA catalytic subunit, which was associated with PKA-dependent phosphorylation of FLAG-Ufd1 as shown by IB with the anti-phospho-PKA substrate antibody (Figure 3A). Similarly, we examined the effect of the PKA activator forskolin on VCP–Ufd1 interaction by treating cells cotransfected with FLAG-Ufd1 and/or HA-VCP in the absence and presence of forskolin. The anti-HA immunoblot of the FLAG IP products revealed robust coprecipitation of HA-VCP with FLAG-Ufd1; in contrast, treatment with forskolin, which induced PKA-dependent FLAG-Ufd1 phosphorylation, strongly blocked this coprecipitation (Figure 3B).
Ufd1 S229 phosphorylation inhibits its interaction with VCP.
As PKA directly phosphorylated the S229 located in the Ufd1 SHP box, a well-defined binding motif for VCP (Figure 1), we next examined whether this phosphorylation plays a negative role in the Ufd1 interaction with VCP. Toward this end, we utilized a phosphomimetic Ufd1 mutant with a substitution of S229 with aspartic acid (S229D). FLAG-Ufd1 WT, S229D, or S229A was cotransfected with HA-VCP, then FLAG IP and HA IP products were reciprocally analyzed by anti-HA and anti-FLAG IB, respectively. The IB results revealed that the S229D mutant had a significantly lower binding affinity for HA-VCP compared with that of the WT, whereas the binding of the S229A mutant to HA-VCP was only mildly reduced (Figure 3C). As expected, FLAG-Ufd1 S229D blocked the interaction with endogenous VCP (Figure 3D). We then introduced the S229 mutations into the GST fusion protein of Ufd1 to further evaluate the differences in VCP-binding affinity. A control GST protein and GST fusion proteins of Ufd1 WT, S229D, or S229A conjugated to glutathione beads were used in pull-down assay of endogenous VCP with HEK293T cell lysates. Similar to the results from the FLAG IP products, GST-Ufd1 S229D showed a markedly weak VCP immunoreactivity in comparison with that of its WT construct, whereas GST-Ufd1 S229A still retained the binding affinity for VCP (Figure 3E). In addition, treatment with calyculin A, leading to marked Ufd1 S229 phosphorylation (Figure 2D), also abrogated the binding of FLAG-Ufd1 to endogenous VCP (Figure 3F). These results indicated that Ufd1 phosphorylation at S229 inhibits the interaction with VCP.
No effect of Ufd1 S231 phosphorylation on the interaction with VCP
Considering that the Ufd1 SHP box harbors another serine residue at position 231 (Figure 1B), we further tested the possibility that Ufd1 S231 phosphorylation may also have the potential to regulate VCP binding. We generated FLAG-Ufd1 S231D and S231A mutants and cotransfected them with HA-VCP. Then, we prepared anti-FLAG and anti-HA immunoprecipitates and repeated the IB experiments, as shown in Figure 3C. In contrast with the results from S229 mutations, however, the S231D and S231A mutants exhibited binding affinities for HA-VCP that were almost equivalent to those of the WT (Figure 4A). Moreover, we tested pull-down assays with GST fusion proteins of the Ufd1 mutants to ascertain their interaction with endogenous VCP. Consistent with the above results, GST-Ufd1 S231D and S231A mutants maintained a binding affinity for VCP as strong as that of the WT (Figure 4B). These results indicated that S231 phosphorylation of Ufd1 does not represent a critical event for modulating the VCP binding.
Phosphomimetic mutation of Ufd1 at S231 does not affect the interaction with VCP.
Similar effects of equivalent mutants of DVC1 (S254D) and p47 (T254D) on VCP binding
To determine whether the phosphorylation of equivalent residues in other SHP box motif-containing proteins such as DVC1 and p47 might similarly regulate their binding to VCP (Figure 1B) , we generated corresponding phosphomimetic mutants of DVC1 (S254D) and p47 (T254D) and evaluated possible changes in their binding affinities for VCP. HA- and V5-IP products from cotransfected cells with DVC1-Strep-HA WT or S254D with V5-VCP were examined by IB analysis. The purified DVC1 and p47 fusion proteins exhibited molecular masses of 70 and 50 kDa, respectively, as expected. Notably, whereas WT DVC1 strongly bound to VCP, the S254D mutant noticeably lost the VCP-binding property (Figure 5A). In comparison, when p47-Strep-HA WT and T254D were tested, the binding of the T254D mutant to V5-VCP was also decreased compared with that of p47-Strep-HA WT (Figure 5B). These findings were consistent with the results obtained from Ufd1 S229D mutant (Figure 3C–E), implying that phosphorylation of the conserved serine or threonine in the SHP box motifs of VCP cofactors may function as a common mechanism to impede VCP binding.
Phosphomimetic mutation of the SHP box motif in DVC1 (S254D) or p47 (T254D) blocks their VCP binding.
Impairment of ERAD function by the Ufd1 S229 phosphorylation
As Ufd1 interaction with VCP is required for the removal of misfolded proteins via the ERAD pathway, we addressed whether Ufd1 S229 phosphorylation had a functional effect on ERAD. Toward this end, we ectopically expressed the albinism-associated TYR mutant (C89R), which has been used as a substrate for ERAD assays [29,31,48]. Protein levels of Myc-tagged TYR-C89R (∼75 kDa) in the resulting cell lysates were assessed by anti-Myc IB analysis following cotransfection with FLAG-Ufd1 WT, S229D, or S229A. The protein levels of Myc-TYR-C89R were much higher in the S229D-coexpressing cells than those in the WT- or S229A-coexpressing cells and relatively higher in cells expressing Myc-TYR-C89R alone compared with those in the WT-coexpressing cells (Figure 6A). Similarly, the Ufd1 mutants were also cotransfected with FLAG-tagged HMGCR, a key enzyme in cholesterol synthesis, as another ERAD substrate [49,50]. Anti-FLAG IB of the transfected cell lysates to detect FLAG-HMGCR (∼140 kDa) and FLAG-Ufd1 (∼45 kDa) showed a higher level of FLAG-HMGCR protein in the S229D-coexpressing cells in comparison with that in the WT- or S229A-coexpressing cells (Figure 6B), consistent with the result of Myc-TYR-C89R. These results indicated that the VCP binding-deficient S229D mutant of Ufd1 was less effective in mediating ERAD function.
Ufd1 S229 phosphorylation negatively regulates the ERAD process.
To examine whether the increased ERAD substrates caused by Ufd1 S229D mutant was due to defective protein degradation rather than simply to differences in transfection efficiency, we treated Ufd1 mutants-expressing cells with CHX, an inhibitor of protein synthesis. CHX chase experiments are typically used for measuring protein degradation related to changes in protein stability. The IB analysis demonstrated that Myc-TYR-C89R protein levels declined up to 2 h after CHX treatment in the FLAG-Ufd1 WT- or S229A-expressing cells (Figure 6C). In sharp contrast, however, Myc-TYR-C89R protein levels in the S229D-expressing cells remained relatively high during the tested time period (Figure 6C). Comparison of the effects of CHX on the transfected V5-HMGCR between Ufd1 WT and the mutants yielded similar results: V5-HMGCR proteins were degraded in the WT- and S229A-expressing cells but stably accumulated in the S229D-expressing cells (Figure 6D).
We examined whether endogenous Ufd1 could be phosphorylated by PKA and the Ufd1 phosphorylation could influence VCP binding and ERAD. To address these points, cells transfected with Myc-TYR-C89R as a reporter of ERAD were treated in the absence and presence of forskolin, endogenous Ufd1 protein was then purified by IP with an Ufd1-specific antibody. As shown in Figure 6E, phospho-serine and VCP immunoreactivities in Ufd1 IP products were increased and decreased, respectively, upon forskolin treatment, as expected. The ERAD substrate levels in forskolin-treated lysates were relatively high compared with those in the untreated lysates (Figure 6E). No immunoreactivities of phospho-serine, VCP, and Ufd1 in the IP samples with control IgG confirmed the specificity of these IP experiments. These findings were consistent with the results obtained from the Ufd1- and VCP-transfected cells and indicated that PKA-mediated phosphorylation of intact Ufd1 acts to suppress VCP interaction and ERAD. In a similar context, we also tested the PKA-mediated Ufd1 phosphorylation for its implication in ERAD regulation using reexpression of Ufd1 S229A mutant. For this, we firstly knocked down Ufd1 with its specific siRNA, added back FLAG-Ufd1 WT or S229A, and then treated with or without forskolin. Myc-TYR-C89R was cotransfected for monitoring ERAD. Ufd1 siRNA-mediated knockdown and FLAG-Ufd1 reexpression were confirmed by Ufd1 and FLAG IB (Figure 6F). We observed that following forskolin treatment, Myc-TYR-C89R protein levels were much higher in the WT-reexpressing cells than in the S229A-reexpressing cells (Figure 6F), further supporting down-regulation of ERAD by PKA-dependent Ufd1 S229 phosphorylation.
In addition, we examined the potential effect of Ufd1 S229 phosphorylation on the subcellular localization of Myc-TYR-C89R by confocal imaging. MG132 (a proteasomal inhibitor) was used to clearly detect the ERAD substrate and calreticulin, an ER-resident protein, was immunostained for ER visualization . Confocal images showed that the Myc-TYR-C89R protein localized not only to the ER (the zoomed area 1) but also to the sites where calreticulin immunofluorescence was relatively sparse (the zoomed area 2) in FLAG-Ufd1 WT-expressing cells (Figure 7A). In comparison, Myc-TYR-C89R protein was retained primarily at the calreticulin-positive sites (the zoomed area 1 and 2) in Ufd1 S229D-expressing cells (Figure 7B). These results suggested that Ufd1 S229 phosphorylation may impede, at least in part, the extraction of ERAD substrate from the ER.
Effect of Ufd1 WT and S229D on the ER localization of the Myc-TYR-C89R protein.
In the ERAD pathway, biochemical but not regulatory features of Ufd1 binding and recruitment of VCP via the Ufd1 SHP box have been reported [21,25,28,29]. In the present study, we addressed a potential regulatory role of phosphorylation of the conserved S229 in the Ufd1 SHP box motif in VCP–Ufd1 binding. Here, we demonstrated that the S229 in Ufd1 serves as the main site for PKA-mediated phosphorylation and that this modification significantly interferes with the interaction with VCP. Consistent with this, we showed that ERAD substrates such as TYR-C89R and HMGCR were accumulated in the ER without being degraded upon the expression of a VCP binding-deficient mutant (S229D), supporting an inhibitory role of Ufd1 S229 phosphorylation in the ERAD pathway (Figure 8). As formation of the VCP–Ufd1 complex is required for segregating ERAD substrates and extracting them out of the ER, it is likely that the liberation of ERAD substrates becomes weakened through this Ufd1 modification.
Schematic diagram of the inhibitory effect of Ufd1 S229 phosphorylation on VCP binding and ERAD.
We previously resolved the protein structure of the binding interface between VCP and Ufd1 and identified the key residues in the Ufd1 SHP box motif, such as F225, F228, and L235, which bind to the VCP N domain ; however, these residues are not readily modified inside cells. In general, cellular protein interactions are reversibly controlled with protein phosphorylation constituting one of the most efficient mechanisms that can modulate the assembly and disassembly of protein complexes. Our results showed that Ufd1 S229 phosphorylation could be reversibly regulated by PKA and a calyculin A-sensitive phosphatase, and that the S229 phosphorylation status serves as a critical factor for VCP–Ufd1 complex formation. Conversely, the VCP binding appeared to be unaffected by phosphorylation of S231, another serine in the Ufd1 SHP box motif, suggesting a specific role of S229 phosphorylation in various functions associated with VCP–Ufd1 interaction.
The S229 in Ufd1 is located in the R-X-X-S/T amino acid sequence motif, which has been shown to undergo PKA-mediated phosphorylation. For example, the S21 residue (R-D-G-S) in tyrosine kinase Fyn , S183 (R-T-G-S) in caspase-9 , T90 residue (R-T-L-T90) in heat shock protein 90-α , and S458 (R-K-R-S) in calcium/calmodulin-dependent protein kinase kinase 1  are phosphorylated by PKA. In addition, PKA-mediated phosphorylation of the NAD-dependent protein deacetylase sirtuin-1, small GTPase RhoA, and transcription factor SOX11 occur on S434 (I-G-S-S434) , S188 (K-K-K-S) , and S133 (A-K-P-S) , respectively. Nevertheless, considering that other protein kinases such as PKC and cGMP-dependent protein kinase often show overlapping activities with PKA, and Rho-associated protein kinase can recognize the R-X-X-S/T motif [37,39,40], we cannot exclude the possibility that other protein kinases aside from PKA also play a role in the VCP binding and ERAD by modulating Ufd1 S229 phosphorylation. Alternatively, Ufd1 contains two additional R-X-X-S/T motifs (R-F-S-T21 and R-C-F-S27) in the N-terminal UT3 domain; however, even if PKA could phosphorylate these motifs, these modifications are not expected to substantially affect VCP–Ufd1 interaction because the UT3 domain functions as a ubiquitin-binding domain [16,24,28].
Among VCP cofactors, DVC1 is recognized as a recruiting factor for VCP during the DNA damage response [58,59] and p47 plays an important role in membrane events of Golgi and the ER as a VCP adaptor [60–62]. We demonstrated that phosphomimetic mutants of DVC1 (S254D) and p47 (T254D) in the respective SHP box motifs lost the binding affinity for VCP, similar to the effect of Ufd1 S229D. Although it remains to be determined whether DVC1 and p47 undergo phosphorylation at the target site, our results implied that phosphorylation of S254 in DVC1 and T254 in p47 may control the function of the nucleus and Golgi, respectively, by restricting the binding to VCP. However, as the arginine residue at the −3 upstream position of Ufd1 S229 (R-A-F-S) is absent from the sequences of DVC1 S254 (I-P-F-S) and p47 T254 (K-A-F-T), the possibility exists that phosphorylation of the SHP box motifs may be differentially co-ordinated by more than one protein kinase including PKA depending on the distinct target organelles.
Notably, hyperactive ERAD resulting from overexpression of synoviolin (also called Hrd1), a major ERAD-associated E3 ubiquitin ligase, was suggested to be a leading cause of rheumatoid arthritis [63,64]. In addition, the VIM motif-containing small VCP-interacting protein  and UBX domain-containing protein 1 (also termed SAKS1)  act as endogenous inhibitors of the ERAD pathway. Therefore, it is conceivable that uncontrolled, excessive ERAD activity may perturb the quality control of protein folding, which can negatively impact ER homeostasis. In such cases, the extent and duration of ERAD must be regulated through certain mechanisms. Based on our findings, we speculated that the S229 phosphorylation of Ufd1 might serve as a signal that could moderate Ufd1-dependent ERAD activity, thereby contributing to the maintenance of ER integrity. Further investigation to define the molecular details underlying the spatiotemporal regulation of PKA-mediated Ufd1 S229 phosphorylation during the progression of ERAD is therefore warranted.
endoplasmic reticulum-associated degradation
Human embryonic kidney 293T
3-hydroxy-3-methylglutaryl coenzyme A reductase
nuclear protein localization protein 4
protein kinase A
sodium dodecyl sulfate-polyacrylamide gel electrophoresis
Tris-buffered saline containing 0.1% Tween-20
ubiquitin regulatory X
ubiquitin fusion degradation protein 1
Q.T.N. performed gene cloning, biochemical experiments, and cell imaging, and validated the data. J.C. supported biochemical studies. J.K.Y. and S.Y.L. developed the study concept. Q.T.N. and S.Y.L. analyzed and interpreted the data. S.Y.L. supervised this study and wrote the manuscript.
This study was supported by research grants from the National Research Foundation of Korea funded by the Ministry of Science and ICT (NRF-2018R1A2B6004598).
We thank Drs. Pietro De Camilli, Ho Chul Kang, Petek Ballar, and Hemmo Meyer for providing the expression constructs.
The Authors declare that there are no competing interests associated with the manuscript.