The E3 ubiquitin ligase CHIP mediates ubiquitination and proteasomal degradation of PRMT5

Abstract

Protein arginine methyltransferase 5 (PRMT5) is an important member of the protein arginine methyltransferase family that regulates many cellular processes through epigenetic control of target gene expression. Because of its overexpression in a number of human cancers and its essential role in cell proliferation, transformation, and cell cycle progression, PRMT5 has been recently proposed to function as an oncoprotein in cancer cells. However, how its expression is regulated in cancer cells remains largely unknown. We have previously demonstrated that the transcription of PRMT5 can be negatively regulated by the PKC/c-Fos signaling pathway through modulating the transcription factor NF-Y in prostate cancer cells. In the present study, we demonstrated that PRMT5 undergoes polyubiquitination, possibly through multiple lysine residues. We also identified carboxyl terminus of heat shock cognate 70-interacting protein (CHIP), an important chaperone-dependent E3 ubiquitin ligase that couples protein folding/refolding to protein degradation, as an interacting protein of PRMT5 via mass spectrometry. Their interaction was further verified by co-immuoprecipitation, GST pull-down, and bimolecular fluorescence complementation (BiFC) assay. In addition, we provided evidence that the CHIP/chaperone system is essential for the negative regulation of PRMT5 expression via K48-linked ubiquitin-dependent proteasomal degradation. Given that down-regulation of CHIP and overexpression of PRMT5 have been observed in several human cancers, our finding suggests that down-regulation of CHIP may be one of the mechanisms underlying PRMT5 overexpression in these cancers.

Graphical abstract

1. Introduction

Protein arginine methyltransferase 5 (PRMT5) is a type II methyltransferase that can symmetrically methylate arginine residues of histones and non-histone substrates [1]. The symmetric methylation on histone H4 at arginine 3 (H4R3) and/or histone H3 at arginine 8 (H3R8) is generally thought to result in transcriptional repression of target genes such as suppressor of tumorigenicity 7 [1, 2], nonmetastatic 23 [1], p53 [3], and RBs (RB1, RBL1, RBL2) [4]; whereas methylation of non-histone substrates including E2F1, p53, RelA/p65, epidermal growth factor receptor (EGFR), RAD9, and programmed cell death 4 generates more diverse cellular effects [5, 6]. For example, the methylation of E2F1 at R111 and R113 by PRMT5 reduces its ability to suppress cell growth and to promote apoptosis, conferring a survival advantage to tumor cells [7]. Also, methylation of p65 at R30 activates NF-κB signaling pathway and facilitates the expression of its target genes including tumor necrosis factor (TNF), TNF receptor-associated factor 1, interleukin-8, and interleukin 1A [8]. It has been proposed that PRMT5 functions as an oncoprotein by either silencing the expression of tumor suppressors or activating the signaling molecules that are crucial for cancer cells [5]. In fact, recent studies have shown that up-regulation of PRMT5 expression correlates with the development and progression of several human cancers, such as breast cancer [9], gastric cancer[10], colorectal cancer [7], ovarian cancer [11], leukemia and lymphoma [2]. However, how PRMT5 expression is regulated in cancer cells remains largely unknown.

We have previously demonstrated that in human prostate cancer cells, PRMT5 can be transcriptionally activated by nuclear factor Y (NF-Y), and that the protein kinase C (PKC)/c-Fos signaling pathway negatively regulates PRMT5 expression through transcriptional down-regulation of NF-Y [12]. Recent research has also found that MYC directly up-regulates the transcription of the core small nuclear ribonucleoprotein particle (snRNP) assembly genes, in which PRMT5 is the key enzymatic component [13]. In addition to the transcriptional regulation of PRMT5 expression, PRMT5 is also regulated by miR-92b/96 in mantle cell lymphoma [2]. Research from the same group also demonstrates that down-regulation of another three miRNAs (miR-19a, miR-25, and miR-32) in several lymphoid cancer cell lines leads to an increase of PRMT5 protein expression [4]. Recently, it has been observed that treatment of three different human cancer cell lines (ovarian, colon, and melanoma) with the heat shock protein 90 (Hsp90) inhibitor 17-AAG reproducibly down-regulates the expression of PRMT5 at the protein level [14]. Given the role of Hsp90 in the regulation of protein folding and degradation, it is reasonable to postulate that PRMT5 may be a putative client protein for Hsp90 [14].

Ubiquitination is one of the most important post-translational modifications that regulates diverse cellular signaling [15]. To execute the ubiquitination process, the consecutive action of three enzymes including the E1 ubiquitin-activating enzyme, the E2 ubiquitin-conjugating enzyme, and the E3 ubiquitin ligases are required for the attachment of ubiquitin to a substrate [16, 17]. The ubiquitin-proteasome system (UPS) is often utilized to fine-tune the expression of target proteins that are associated with cancer development and progression. As a mechanism of quality control for protein folding, ubiquitin-dependent proteasomal degradation is often coupled with the molecular chaperone system to remove misfolded proteins [16, 18, 19]. In this system, E3 ubiquitin ligases appear to be the key regulators that function together with the chaperone system to regulate protein degradation. Carboxyl terminus of heat shock cognate 70-interacting protein (CHIP), also known as STUB1/STIP1 homology and U-Box containing protein 1, is a chaperone-dependent E3 ubiquitin ligase [20, 21]. CHIP contains three tandem tetratricopeptide repeat (TPR) motifs, through which it interacts with the chaperones including heat shock protein 70 (Hsp70) and Hsp90; and a U-box domain, which is responsible for ubiquitination of the chaperone-bound substrates. Recently, CHIP has been proposed as a tumor suppressor since lower expression of CHIP promotes cell proliferation and/or inhibits apoptosis in breast cancer [22, 23], gastric cancer [24], pancreatic cancer [25], and colorectal cancer [26]. Specifically, the role of CHIP in these cancers is to control the expression of several crucial proteins, such as ErbB2 [22], hypoxia-inducible factor-1a [27], c-Myc [28], p65 [26], and EGFR [25].

In the present study, we demonstrated that PRMT5 can undergo polyubiquitination both in vivo and in vitro. We also provided evidence that the E3 ubiquitin ligase CHIP couples to the molecular chaperone system (Hsp70/Hsp90) and mediates ubiquitin-dependent proteasomal degradation of PRMT5. Our work provides a new mechanism underlying PRMT5 overexpression in cancer cells.

2. Materials and methods

2.1. Cell culture and reagents

Prostate cancer cell line LNCaP, human embryonic kidney 293T (HEK293T) and COS-1 cells were purchased from American Type Culture Collection (ATCC), and were maintained in RPMI 1640 or DMEM medium containing 10% heat-inactivated fetal bovine serum (FBS) supplemented with penicillin/streptomycin, sodium pyruvate and L-glutamine. All cells were maintained at 37°C in a humidified incubator containing 5% CO₂. Cycloheximide (CHX) and MG132 were purchased from Sigma. GA and 17-AAG were purchased from Tocris Bioscience.

2.2. Plasmid construction

The pCMV-Myc-PRMT5 expression plasmid was previously constructed [12], and was used as a template to generate methyltransferase activity-deficient mutant pCMV-Myc-PRMT5-R368A [29], and a series of truncated fragments covering the residues 229–637, 284–637, 352–637, and 451–637. For mutagenesis, nucleotide substitutions (from lysine/K to arginine/R) were introduced into PRMT5 using ligation PCR as described previously [12, 30]. pCMV-FLAG-PRMT5 was generated by subcloning PRMT5 into pCMV-FLAG expression vector (Sigma). Various truncated mutants and single-point mutations of PRMT5 were generated using PCR or ligation PCR, and then subcloned into pCMV-FLAG or pCMV-HA (Clontech). The chaperone interaction-deficient K30A mutant (Lysine/K to alanine/A at position 30) and E3 ubiquitin ligase activity-deficient H260Q mutant (histidine/H to glutamine/Q at position 260) for CHIP were generated using the same methods. Two truncated fragments of CHIP were amplified by PCR using primers specific for ΔU-box (forward primer: 5′-ccggaattcggatgaagggcaaggagg-3′ and reverse primer: 5′-cggggtaccgaggtaggagtgcagctc-3′) and ΔTPR (forward primer: 5′-ccggaattcggatcgcgaagaagaagcg-3′ and reverse primer: 5′-cggggtaccgtagtcctccacccagcc-3′), and then were subcloned into pCMV-FLAG. To express CHIP as a fusion with GST, the cDNA encoding CHIP was subcloned into pGEX-4T2 vector. For BiFC plasmid construction, pCMV-Myc and pCMV-HA were used to generate pBiFC-VN155(I152L)-N and pBiFC-VC155-N vectors, followed by the subcloning of the cDNAs encoding PRMT5 and CHIP into either of these two BiFC cloning vectors. cDNAs encoding wild-type (WT) ubiquitin, ubiquitin-K48R and ubiquitin-K63R were kind gifts from Dr. Chittaranjan Das lab (Purdue University), and were then subcloned into pCMV-HA vector. All plasmid constructs were verified by enzymatic digestion or DNA sequencing.

2.3. In vivo ubiquitination assay

Cells were co-transfected with the plasmid encoding HA-Ubiquitin and Myc-PRMT5 or its various mutants, along with plasmids encoding FLAG-CHIP or CHIP mutants for the indicated time, followed by the treatment with MG132 (10 μM) for another 6 h. Whole cell lysate (WCL) was prepared, and 500 μg of the WCL was used for immunoprecipitation (IP) using the antibodies against PRMT5, HA and Myc, followed by the detection of respective proteins by immunoblotting (IB). For the detection of protein ubiquitination, a final concentration of 10 mM NEM (Sigma, E3876-5G) was added to the IP buffer in order to inhibit protein deubiquitination.

2.4. Co-immunoprecipitation and immunoblotting

Cells were harvested and washed twice with cold phosphate buffered saline (PBS), and then lysed by sonication in lysis buffer (10 mM Tris-HCl pH 7.4, 1.5 mM MgCl₂, 10 mM KCl) containing 1 mM phenylmethlysulfonyl fluoride (PMSF), 1 mM dithiothreitol (DTT), protease cocktail, 25 mM okadaic acid and 1% Triton X-100 as described previously [31]. For the preparation of soluble and insoluble samples, supernatant was collected and saved as soluble fraction, and pellets were resuspended in the same volume of lysis buffer and sonicated on ice, and the boiled pellets were saved as insoluble fraction. For co-immunoprecipitation (Co-IP), cells were treated with or without 17-AAG for 24 h, and the cell lysate was prepared by sonication in IP buffer (50 mM Tris-HCl, pH 7.4, 50 mM NaCl, 1 mM EDTA, 0.1% Triton X-100, 1.5 mM sodium orthovanadate, 50 mM sodium fluoride, 10 mM sodium pyrophosphate, 10 mM beta-glycerolphosphate, 1mM PMSF and protease cocktail), and IP was performed following the same procedure as described previously [12, 32]. The antibodies used for IB analysis were: anti-β-actin (Cell Signaling Technology, 8H10D10), anti-PRMT5 (Millipore, 07–405), anti-CHIP (Santa Cruz, G-2 sc-133066), anti-FLAG M2 (Cell Signaling Technology, 9A3), anti-HA (Cell Signaling Technology, 6E2), anti-GST (BD Biosciences), and anti-Myc (GenScript, A00704-100). Secondary HRP-conjugated antibodies were purchased from Amersham Biosciences.

2.5. Mass spectrometry analysis of PRMT5 interacting proteins in LNCaP cells

For the identification of PRMT5 interacting proteins using mass spectrometry, LNCaP cells were transfected with the plasmids encoding FLAG-PRMT5 and HA-Ubiquitin for 42 h, followed by the treatment with MG132 for another 6 h. WCL was used for IP of FLAG-PRMT5 with anti-FLAG antibody, or the control IgG, followed by trypsin digestion and quantitative mass spectrometry analysis as described before [33]. Three independent experiments were performed, and E3 ligases that were specifically identified in the anti-FLAG immunoprecipitates but not in the IgG control were considered as putative E3 ligases for PRMT5 interaction.

2.6. GST pull-down assay

pGEX-4T2-CHIP was transformed into Escherichia coli strain BL21, and a single colony of the transformed bacteria was inoculated into 200 ml LB medium and cultured at 37 °C till the optical density value reached 0.6. CHIP expression was induced by adding 1.0 mM isopropyl-beta-D-thiogalactopyranoside into the culture for 4 h. For cell lysate preparation, pelleted bacteria were resuspended in ice cold lysis buffer (50 mM Tris-HCl pH 7.4, 50 mM NaCl) and disrupted by sonication, followed by centrifugation at 15,000×g for 30 min at 4°C. For GST pull-down assay, plasmid encoding Myc-PRMT5 was transfected into HEK293T cells using FuGENE 6 following the manufacturer’s instructions and incubated for 24 h. The transfected cells were then lysed, and WCL was prepared. Approximately 500 μg of WCL was incubated with the same molar ratio of GST and GST-CHIP at 4 °C for overnight, followed by the incubation with glutathione-Sepharose beads (GE Healthcare) for another 2 h. The beads were washed three times with lysis buffer and boiled in 2×SDS loading buffer and subjected to SDS-PAGE gel analysis [34].

2.7. BiFC assay

BiFC assay was performed essentially the same as previously described to analyze the interaction between PRMT5 and CHIP in COS-1 cells [35]. Briefly, COS-1 cells were grown on coverslips in a 12-well plate for 24 h, and the BiFC plasmids encoding Myc-VN155-PRMT5 and HA-VC155-CHIP, along with FLAG-Cerulean were co-transfected into COS-1 cells for 24 h. Cells were then fixed with 3.7% paraformaldehyde, and stained with 4′6-Diamidino-2-Phenylindole (DAPI) for 5 min at room temperature (RT) under dark condition. The fluorescent images were acquired by Nikon A1 confocal microscope.

2.8. Luciferase assay

HEK293T cells were transiently transfected with 1μg of pCMV-FLAG (Vector) or pCMV-FLAG-CHIP (CHIP), along with 500 ng of the PRMT5 proximal promoter reporter gene, plus 100 ng of pRL-TK for 24 h using Lipofectamine® 3000 Transfection Reagent (Invitrogen), and the relative luciferase activity was determined using Dual-Luciferase® Reporter Assay system (Promega) as described previously [12].

2.9. Reverse transcription and real-time PCR

For real-time PCR analysis, total RNA was purified using TRIzol® Plus RNA Purification Kit (Life Technologies), and 2 μg of RNA was then reverse transcribed to cDNA using the High Capacity cDNA Reverse Transcription Kit (Invitrogen) according to the manufacturer’s protocol. Human PRMT5 and GAPDH primers used for real-time PCR were the same as described previously [12]. For real-time PCR, StepOne Real-Time PCR (Applied Biosystems) was performed by using SYBR Select Master Mix. All real-time PCR reactions were performed in triplicate with at least three independent experiments, and the relative expression of each gene was normalized to GAPDH [36].

2.10. RNA interference

Endogenous CHIP was depleted in cells using siGENOME Human STUB1/CHIP (10273) siRNA SMARTpool (Dharmacon, Lafayette, CO), and siGENOME Non-Targeting siRNA Pool (Dharmacon, Lafayette, CO) was used as a negative control. For siRNA experiments, the indicated siRNA was transfected into HEK293T cells using DharmaFECT 1 Transfection Reagent (Dharmacon) according to the manufacturer’s protocol. After cells were transfected for 72 h, WCL was prepared, and the ubiquitination pattern or the expression level of CHIP was analyzed by immunoblotting.

2.11. Analysis of cell apoptosis by flow cytometry

Plasmid encoding FLAG-CHIP (or Vector only) was transfected into cells for 48 h, followed by the treatment of 17-AAG for another 24 h. Both floating and adherent cells were collected for flow cytometry analysis using Annexin V-APC/7-amino-actinomycin D Apoptosis Detection Kit (KeyGEN Biotechnology, Nanjing, China). Briefly, HEK293T cells were trypsinized and washed with filtered PBS twice, resuspended in 200 μl binding buffer with 2 μl Annexin V-APC, and then incubated at RT for 15 min. Supernatant was gently removed after 300×g centrifugation for 2 min, followed by adding 2 μl of 7-ADD into 200 μl binding buffer and incubated at RT for 5 min in the dark. At least 50, 000 cells were resuspended in 800 μl of PBS. Three independent experiments were performed using a BD Accuri C6 flow cytometer at a low flow rate with a minimum of 1×10⁴ cells, and the percentage of apoptotic cells was determined.

2.12. Cell growth analysis

To determine the role of CHIP in 17-AAG-induced cell growth inhibition, HEK293T cells were seeded and grown on coverslips in a 6-well plate at a cell density of 1×10⁵ cells/well, and siRNA control (siCon) or siCHIP was transfected into cells for 48 h using DharmaFECT 1 Transfection Reagent, followed by the treatment with or without 17AAG (100 nM) for another 24 h. Total cell number was counted using hemocytometer, and the percentage of cell growth over the control was determined [12].

2.13. Sequence alignment and visualization of PRMT5 structure

Sequence alignment and ubiquitination site prediction were performed using several online alignment and prediction software (http://bdmpub.biocuckoo.org/prediction.php, http://www.ubpred.org/, and http://protein.cau.edu.cn/cksaap_ubsite/), and the crystal structures of PRMT5 and MEP50 were retrieved from PDB database (accession code 4GQB), and processed by PyMOL software (http://www.pymol.org/). The illustration of protein domain of PRMT5 and CHIP was created using DOG1.0 software [37].

2.14. Statistical analysis

The GraphPad Prism 6 Software (Graphpad Software, San Diego, CA, USA) was used to perform all statistical analysis. Data were presented as mean ± SD from at least three independent experiments. Comparison between two groups was conducted by using Student’s t test. Two-way ANOVA was used to compare the means of two independent variables, followed by Tukey’s post-hoc test. p value less than 0.05 was considered to be statistically significant.

3. Results

3.1. PRMT5 undergoes polyubiquitination in LNCaP cells

We have previously shown that PRMT5 is transcriptionally activated by NF-Y in LNCaP prostate cancer cells, and that treatment of cells with the PKC activator phorbol-12-myristate-13-acetate (PMA) down-regulates PRMT5 expression [12]. During the course of these experiments, we noticed that PMA-induced PRMT5 down-regulation appeared to be partially reversed by the proteasome inhibitor MG132 (Fig. 1A), suggesting that PRMT5 might undergo proteasomal degradation. Given that polyubiquitination is a prerequisite for the proteasomal degradation of many cytosolic proteins [38], we sought to determine whether PRMT5 is subjected to polyubiquitination. To this end, LNCaP cells transfected with the plasmid encoding HA-Ubiquitin were treated with or without MG132, followed by immunoprecipitation of endogenous PRMT5 with anti-PRMT5 antibody. Indeed, polyubiquitination of endogenous PRMT5 was readily detectable in the absence of MG132, and the presence of MG132 further enhanced the polyubiquitination of PRMT5 (Fig. 1B). This result provides evidence that endogenous PRMT5 is polyubiquinated. To determine whether exogenously expressed PRMT5 also undergoes polyubiquitination, we co-expressed Myc-PRMT5 with HA-Ubiquitin (or HA-Vector) in LNCaP cells in the presence of MG132, and then immunoprecipitated Myc-PRMT5 with anti-Myc antibody followed by immunoblotting of HA-Ubiquitin with anti-HA antibody. As shown in Fig. 1C, Myc-PRMT5 was clearly polyubiquitinated. A reverse immunoprecipitation using anti-HA antibody was performed to further confirm the polyubiquitination of the exogenously expressed Myc-PRMT5 in LNCaP cells (Fig. 1D). Taken together, we conclude that PRMT5 can undergo polyubiquitination at both endogenous and exogenous level in LNCaP cells.

Fig. 1.

Ubiquitination of PRMT5 in LNCaP cells. (A) Proteasome inhibitor MG132 partially restores PMA-induced reduction of PRMT5 expression. LNCaP cells were treated with the PKC activator PMA (100 nM) in the presence or absence of MG132 (10 μM) for 24 h, and the whole cell lysate (WCL) was prepared and subjected to immunoblotting (IB). NF-YA-s (a positive control) represents the short isoform of NF-YA. (B) Ubiquitination of endogenous PRMT5 in LNCaP cells. LNCaP cells were transfected with HA-Vector or the plasmid encoding HA-Ubiquitin for 42 h, followed by the treatment with DMSO (−) or MG132 (+) for another 6 h. WCL was immunoprecipitated with anti-PRMT5 antibody and probed with anti-HA or anti-PRMT5 antibody. (C) Ubiquitination of exogenous PRMT5 in LNCaP cells. Myc-PRMT5 was co-expressed with either HA-Vector or HA-Ubiquitin in LNCaP cells for 48 h, and WCL was immunoprecipitated with IgG or anti-Myc antibody, respectively, followed by immunoblotting of β-actin, HA-Ubiquitin, and Myc-PRMT5. * indicates non-specific band. (D) HA-Ubiquitin was co-expressed with either Myc-Vector or Myc-PRMT5 in LNCaP cells for 48 h, and WCL was immunoprecipitated with anti-HA antibody, followed by immunoblotting with β-actin and Myc antibodies. PRMT5-(Ub)n in B, C and D indicates polyubiquitination of PRMT5.

3.2. PRMT5 polyubiquitination involves multiple lysine residues

Since covalent attachment of multiple ubiquitin molecules to specific lysine residues of target proteins is a prerequisite for recognition and subsequent degradation by proteasome [39], we were interested in identifying the lysine residues that are responsible for PRMT5 ubiquitination. We generated a series of deletion mutants based on PRMT5 structure (Fig. 2A, top), and co-expressed them with HA-Ubiquitin in LNCaP cells to map the ubiquitination sites. As shown in Fig. 2A, all of these mutants appeared to undergo polyubiquitination in the presence or absence of MG132 treatment. Significantly, the PRMT5 mutants 229–637, 284–637 and 352–637 were highly ubiquitinated when compared with full-length PRMT5, whereas the ubiquitination pattern of the PRMT5 mutant 451–637 remained unchanged in the presence of MG132 treatment, suggesting that the major ubiquitination sites of PRMT5 are located between residues 229 to 451. Based on the crystal structure of PRMT5, we then focused on ten most surface-exposed lysine (K) residues (highlighted in Fig. 2B) within the region of residues 229–451. We mutated these lysine residues to arginine (R) at the indicated sites, including positions at 240 and 241 (1), 248 (2), 259 (3), 275 (4), 302 (5), 329 and 333 (6), 343 (7), 354 (8), 380 (9), and 387 (10). The expression level of all mutants was comparable, however, mutated individual lysine did not significantly change the ubiquitination pattern of PRMT5 (Fig. 2C). Next, we mutated the first five K (M1), the middle four K (M2) and the last three K (M3) to R in combination. As shown in Fig. 2D, all three mutants (M1, M2, and M3) showed a dramatic decrease of polyubiquitination, suggesting that multiple lysine residues might be involved in the polyubiquitination of PRMT5.

Fig. 2.

Multiple lysine residues are involved in the polyubiquitination of PRMT5. (A) A series of PRMT5 truncated mutants were generated (top) and were co-expressed with HA-Ubiquitin in the presence or absence of MG132 in LNCaP cells, and the whole cell lysate (WCL) was then immunoprecipitated with anti-Myc antibody and subjected to immunoblotting for HA-Ubiquitin detection using anti-HA. The membrane was then stripped and re-probed with antibody against Myc. (B) Illustration of predicted surface-exposed lysine residues in PRMT5. Lysine residues exposed on the surface of PRMT5 are highlighted in white with an indicated number. Green represents the catalytic domain of PRMT5 while the light blue indicates the TIM domain of PRMT5, the yellow represents MEP50. (C) The effect of K/R mutations on PRMT5 polyubiquitination. The individual lysine (K) was mutated to arginine (R), and the mutants were co-expressed with HA-Ubiquitin in LNCaP cells in the presence of MG132 for 48 h, and the WCL was subjected to IP with anti-Myc antibody and IB for the detection of PRMT5 polyubiquitination with anti-HA antibody. (D) The polyubiquitination of the indicated PRMT5 mutants was similarly analyzed as described in C.

3.3. Co-chaperone E3 ubiquitin ligase CHIP interacts with PRMT5

E3 ubiquitin ligases are critical regulators of the ubiquitination process for specific substrates [40]. To identify E3 ubiquitin ligases specific for PRMT5, HA-Ubiquitin was co-expressed with FLAG-PRMT5 in LNCaP cells for 42 h and treated with MG132 for another 6 h, followed by immunoprecipitation using anti-FLAG antibody or IgG. The immunoprecipitates were subjected to mass spectrometry analysis. Two ubiquitin E3 ligases CHIP and TRIM21, and one sumo E3 ligase RanBP2, were specifically identified from three independent experiments (Table 1). Given the high coverage of CHIP, we selected CHIP for further validation as a potential interacting protein of PRMT5. Since HEK293T cells have a higher transfection efficiency, we co-expressed Myc-PRMT5 with FLAG-CHIP in HEK293T cells for 48 h in the presence of MG132 and then performed immunoprecipitation with anti-FLAG antibody. Compared with FLAG-Vector or IgG control, Myc-PRMT5 was co-immunoprecipitated with FLAG-CHIP (Fig. 3A), suggesting the specific interaction between Myc-PRMT5 and FLAG-CHIP in cells. Their interaction was further validated using GST pull-down assays (Fig. 3B), as evidenced by the enrichment of PRMT5 by GST-CHIP when compared with GST only. To ascertain the physiological interaction between CHIP and PRMT5, the WCL from HEK293T was prepared and subjected to reciprocal co-immunoprecipitation with either anti-CHIP antibody or anti-PRMT5 antibody. As shown in Fig. 3C, PRMT5 and CHIP were specifically co-immunoprecipitated with either antibody, demonstrating that PRMT5 and CHIP also interact with each other at the endogenous level. Since CHIP and PRMT5 can be both localized to the cytoplasm and nucleus [41], we then determined where they interact in cells using bimolecular fluorescence complementation (BiFC) technique [35, 42]. Given that COS-1 cells have a better cytoplasm/nucleus ratio for visualization of subcellular localizations, we transiently transfected plasmids encoding HA-VC155-CHIP and Myc-VN155-PRMT5 along with a plasmid encoding FLAG-Cerulean into COS-1 cells for 24 h. As shown in Fig. 3D, the Venus signal (BiFC signal) in the transfected cells was predominantly localized in the cytoplasm, suggesting that the interaction between CHIP and PRMT5 likely occurred in the cytoplasm. In line with this, we also found that both CHIP and PRMT5 were co-localized in the cytoplasm by co-expressing them fused to full-length Cerulean and Venus, respectively (Fig. 3E).

Table 1.

Identification of putative E3 ligases for PRMT5 by mass spectrometry

Accession	Gene	Description	Coverage	Unique Peptides	Repeatability
H3BUD0	CHIP	E3 ubiquitin ligase	18.92	3	3
F5H012	TRIM21	E3 ubiquitin ligase	10.53	4	3
P49792	RanBP2	E3 SUMO ligase	2.73	4	2

Fig. 3.

CHIP interacts with PRMT5 both in vitro and in vivo. (A) FLAG-CHIP interacts with Myc-PRMT5 in HEK293T cells. Myc-PRMT5 was co-expressed with FLAG-vector or FLAG-CHIP in HEK293T cells for 48 h, and MG132 was applied for another 6 h. Whole cell lysate (WCL) was prepared and immunoprecipitated with IgG or anti-FLAG antibody for overnight, and immunoprecipitates were analyzed by immunoblotting using Myc, β-actin, and FLAG antibodies. Input: 5% of WCL. (B) GST-CHIP interacts with Myc-PRMT5 in vitro. GST and GST-CHIP were expressed in E. coli. and then immobilized to glutathione agarose beads. The same amount of HEK293T WCL containing overexpressed Myc-PRMT5 (Myc-PRMT5-WCL) was then incubated with GST or GST-CHIP, and the pull down fraction was used for immunoblotting analysis. (C) HEK293T cells were treated with 17-AAG (100 nM) for 24 h, along with MG132 treatment for another 6 h. WCL was used for IP using antibodies against CHIP or PRMT5, and IgG was used as a negative control. (D) The two BiFC plasmids encoding the Myc-VN155-PRMT5 and HA-VC155-CHIP along with pFLAG-Cerulean were co-transfected into COS-1 cells for 24 h. Shown are representative fluorescent images of transfected cells (Cerulean) and the interaction between PRMT5 and CHIP (BiFC). Nuclei were stained with DAPI. Scale bar is 5 μm. (E) Co-localization of PRMT5 and CHIP in cells. Myc-Venus-PRMT5 was co-expressed with HA-Cerulean-CHIP in COS-1 cells for 24 h, and their co-localization was analyzed by confocal microscopy. Scale bar is 5 μm. (F) A schematic for CHIP and its mutants. CHIP-FL represents full-length CHIP, which includes three major domains: (TPR)₃, Charged and U-box domains. K30A indicates chaperone-interaction deficient mutant; H260Q indicates ubiquitination deficient mutant; ΔU-box represents U-box deletion mutant; ΔTPR represents TPR deletion mutant. (G) TPR domain of CHIP is required for PRMT5 interaction. Myc-PRMT5 was co-expressed with FLAG-CHIP or its mutants in HEK293T cells for 48 h, and treated with MG132 for another 6 h. WCL was used for immunoprecipitation as described above. Arrow indicates FLAG-CHIP or its mutants.

CHIP contains a TPR domain involved in the interaction with chaperones at the N-terminus, an U-box domain that possesses ubiquitin ligase activity at the C-terminus, and a linker known as charged domain in between [20]. In order to determine which region of CHIP is required for PRMT5 interaction, we then generated a series of CHIP mutants (chaperone interaction-deficient mutant K30A, ubiquitination-deficient mutant H260Q, and TPR or U-box deletion mutant), to map the PRMT5 interaction domain in CHIP (Fig. 3F). We co-expressed these CHIP mutants as FLAG fusion proteins with Myc-PRMT5 in HEK293T cells and performed immunoprecipitation with anti-FLAG antibody and immunoblotting for Myc-PRMT5 with anti-Myc antibody. Although both H260Q and U-box deletion (ΔU-box) mutants co-immunoprecipitated comparable amount of Myc-PRMT5 when compared with the CHIP-FL, the binding of Myc-PRMT5 to the K30A and TPR deletion (ΔTPR) mutants was almost abolished (Fig. 3G). This result suggests that the TPR domain of CHIP is necessary for the interaction with PRMT5, and that the binding of PRMT5 and chaperons to CHIP may share the same binding motif. However, the interaction between CHIP and PRMT5 is independent of the E3 ligase activity of CHIP. Taken together, these results demonstrate that CHIP and PRMT5 can interact both in vitro and in vivo, and the interaction likely occurs in the cytoplasm.

3.4. CHIP negatively regulates PRMT5 expression

The finding that CHIP interacts with PRMT5 prompted us to determine whether CHIP regulates PRMT5 expression. We first co-expressed Myc-PRMT5 with increasing amounts of FLAG-CHIP in HEK293T cells for 48 h, and then detected the expression of Myc-PRMT5. As shown in Fig. 4A, FLAG-CHIP dose-dependently decreased Myc-PRMT5 protein expression. We also confirmed that there was no significant effect of FLAG-CHIP on the PRMT5 promoter-driven reporter gene activity and PRMT5 mRNA expression (Fig. 4B, C). Similarly, the expression of the methyltransferase activity-deficient mutant Myc-PRMT5-R368A was also down-regulated by FLAG-CHIP, suggesting that CHIP-mediated degradation of PRMT5 is independent of the catalytic activity of PRMT5 (Fig. 4D). Next, we sought to investigate the impact of CHIP expression on the half-life of PRMT5. FLAG-CHIP or FLAG-Vector was co-expressed with Myc-PRMT5 in cells for 36 h, and treatment of CHX was applied for the indicated times. As shown in Fig. 4E and Supplementary Fig. S1, CHIP expression (though vanished at 9 h) reduced the half-life of Myc-PRMT5 from 5.5 h to 3 h, suggesting that CHIP can promote PRMT5 degradation. The identification of PRMT5 as a substrate of CHIP for proteasomal degradation is particularly intriguing, given that many proteins regulated by CHIP also require the molecular chaperone system Hsp90/Hsp70 for protein folding [20, 21]. We next determined whether the two CHIP mutants K30A and H260Q might affect the expression of PRMT5. Interestingly, we found that both CHIP and H260Q, but not K30A mutant, significantly attenuated Myc-PRMT5 expression (Fig. 4F), suggesting that the molecular chaperone system is required for PRMT5 recognition and its subsequent degradation by CHIP.

Fig. 4.

CHIP negatively regulates PRMT5 expression. (A) Overexpression of CHIP dose-dependently decreases PRMT5 expression. pMyc-PRMT5 was co-transfected with pFLAG-Vector (Vector) or an increasing amount of pFLAG-CHIP into HEK293T cells for 48 h. Antibodies against PRMT5, FLAG and β-actin were used for immunoblotting. Representative blots from three independent experiments are shown, and the images were analyzed by Image J software and relative expression of Myc-PRMT5 is presented as mean ± SD (on the right), ****p<0.0001 one-way ANOVA. (B) Overexpression of CHIP has no effect on the PRMT5 promoter activity. One microgram of pFLAG-Vector (Vector) or pFLAG-CHIP (CHIP) was co-transfected with 0.5 μg of the PRMT5 proximal promoter reporter gene, along with 100 ng of pRL-TK into HEK293T cells for 24 h, and the relative luciferase activity was determined and analyzed. (C) Overexpression of CHIP has no effect on PRMT5 mRNA expression. Three microgram of pFLAG-Vector (Vector) or pFLAG-CHIP (CHIP) was transfected into HEK293T cells for 24 h, and the PRMT5 mRNA level was determined by real-time PCR. (D) CHIP promotes degradation of PRMT5 and its methyltransferase activity-deficient mutant. pMyc-PRMT5-WT or pMyc-PRMT5-R368A was co-transfected with pFLAG-CHIP into HEK293T cells for 48 h. Antibodies against Myc, FLAG and β-actin were used for immunoblotting (IB). (E) CHIP promotes the turnover rate of PRMT5 in HEK293T cells. HEK293T cells were transfected with either pFLAG-vector or pFLAG-CHIP along with Myc-PRMT5 for 36 h, followed by the treatment with 10 μg/mL cycloheximide (CHX) for different times, and the turnover rate of Myc-PRMT5 was determined by immunoblotting. Representative results are shown (Top). Bottom: Quantitative result analyzed by Image J is presented as means ± SD from three independent experiments. Dashed line indicates the time required for exogenous PRMT5 being degraded to 50%. Statistical significance (**p<0.01; ****p<0.0001) was determined by two-way ANOVA followed by Tukey’s test. (F) The effect of CHIP and its mutants on the expression of PRMT5. Myc-PRMT5 was co-expressed with CHIP or CHIP mutants (K30A and H260Q) for 48 h, and the expression level of Myc-PRMT5 was determined by immunoblotting. n.s. in B, C, and F indicates no significance (Student’s t test).

3.5. CHIP mediates the down-regulation of PRMT5 expression and cell growth inhibition by 17-AAG

The molecular chaperone proteins (Hsp90 and Hsp70) cooperate with the ubiquitination/proteasomal system to regulate the degradation of unfolded or misfolded proteins. CHIP is one of the major E3 ubiquitin ligases involved in this ubiquitin/molecular chaperone system [20, 43, 44]. Our result that the K30A mutant failed to decrease PRMT5 expression is consistent with previous reports that PRMT5 is a client protein of Hsp90 [14, 45]. This led us to hypothesize that the degradation of PRMT5 may be regulated by the ubiquitin/molecular chaperone system involving CHIP, Hsp90, and Hsp70. In support of this hypothesis, we indeed found that Hsp90 inhibitors 17-AAG and GA, both of which target Hsp90 ATPase binding domain, dose-dependently decreased PRMT5 protein expression in HEK293T cells (Fig. 5A) and in LNCaP cells (Fig. 5B). Further, overexpression of CHIP enhanced 17-AAG-mediated down-regulation of PRMT5 (Fig. 5C). In addition, overexpressed FLAG-CHIP increased 17-AAG-induced cell death from 14.05% to 23.39% (Fig. 5D). To understand the role of endogenous CHIP in the regulation of PRMT5 expression, siRNA SMARTpool targeting CHIP was used to knock down CHIP in HEK293T cells. Significantly, knockdown of CHIP completely inhibited 17-AAG-induced down-regulation of PRMT5 (Fig. 5E), indicating that PRMT5 expression can be regulated by the ubiquitin/molecular chaperone system in cells. We next sought to determine the effect of CHIP on 17-AAG-induced cell growth inhibition/cell death. 17-AAG indeed significantly inhibited cell growth, which is consistent with previous reports [23, 46], and knockdown of CHIP partially rescued cell growth inhibition by 17-AAG (Fig. 5F). Taken together, these results suggest that 17-AAG-induced cell growth inhibition/cell death is likely mediated by CHIP-dependent down-regulation of PRMT5 expression.

Fig. 5.

CHIP mediates the down-regulation of PRMT5 expression and cell growth inhibition by 17-AAG. (A and B) Hsp90 inhibitors dose-dependently inhibit PRMT5 expression in HEK293T and LNCaP cells. HEK293T (A) and LNCaP cells (B) were treated with increasing amounts of 17-AAG (1 nM-100 nM) or GA (10 nM-1μM) for 24 h, and the whole cell lysate (WCL) was subjected to immunoblotting (IB). (C) Overexpression of CHIP enhances the down-regulation of PRMT5 induced by 17-AAG. HEK293T cells were either transfected with pFLAG-Vector (Vector) or pFLAG-CHIP (CHIP) for 48 h, and then treated with 17-AAG for another 24 h before preparing WCL for IB. (D) Overexpression of CHIP increased 17-AAG-induced apoptosis in HEK293T cells. HEK293T cells were either transfected with pFLAG-Vector (Vector) or pFLAG-CHIP (CHIP) for 48 h, followed by the treatment with 17-AAG for another 24 h. Both floating and adherent cells were collected and labeled with Annexin V-APC and 7-amino-actinomycin D (7-ADD) for flow cytometry analysis. The percentage of apoptotic cells (Q2+Q3) was calculated and normalized to the Vector control, and the percentage of apoptotic cells is represented as means ± SD from three independent experiments (**p<0.01; ***p<0.001). (E) Knockdown of CHIP blocks the reduction of PRMT5 induced by 17-AAG treatment. HEK293T cells were transfected with siRNA Control (siCon) or siRNAs targeting CHIP (siCHIP) for 60 h, and 17-AAG was applied for another 24 h before preparing WCL for IB. (F) Knockdown of CHIP partially reverses 17-AAG-induced cell growth inhibition. HEK293T cells were transfected with siRNA Control (siCon) or siRNAs targeting CHIP (siCHIP) for 60 h, followed by the treatment with 17-AAG for another 24 h. The total cell number was counted using hemocytometer, and is presented as the percentage of the control. Statistical significance (*p<0.05; **p<0.01) Representative blots from three independent experiments are shown in A, B, C, and E.

3.6. CHIP promotes PRMT5 degradation through K48-linked ubiquitination

CHIP is an E3 ubiquitin ligase that mediates protein degradation by ubiquitinating its substrates [47]. Since PRMT5 undergoes ubiquitination and CHIP negatively regulates PRMT5 expression, we were interested in determining whether PRMT5 is subjected to CHIP-mediated proteasomal degradation. To this end, FLAG-CHIP and Myc-PRMT5 were co-expressed in the absence or presence of the proteasome inhibitor MG132. As shown in Fig. 6A, treatment with MG132 attenuated the inhibitory effect of FLAG-CHIP on Myc-PRMT5 expression (Fig. 6A), suggesting that the down-regulation of PRMT5 expression by CHIP is mainly through the proteasomal degradation pathway. To demonstrate that CHIP is capable of ubiquitinating PRMT5, we performed in vivo ubiquitination assays in HEK293T cells by transiently co-expressing Myc-PRMT5 with FLAG-CHIP in the presence of HA-Ubiquitin. Immunoprecipitaion results showed that overexpression of CHIP increased the ubiquitination of PRMT5 when compared with FLAG-Vector only (Fig. 6B). However, both K30A and H260Q mutants had a reduced activity when compared with FLAG-CHIP (Fig. 6B), suggesting that the chaperone binding activity of CHIP and the U-Box region are required for CHIP-induced ubiquitination of PRMT5. However, H260Q mutant not only decreased PRMT5 expression (Fig. 4F), but also abolished the ubiquitination of PRMT5 (Fig. 6B), and this led us to investigate whether PRMT5 moves to the insoluble fraction as suggested previously [48, 49]. As shown in Fig. 6C, H260Q mutant did not increase the level of insoluble PRMT5.

Fig. 6.

CHIP promotes PRMT5 degradation through K48-linked ubiquitination. (A) MG132 inhibits the down-regulation of PRMT5 expression by CHIP. pFLAG-Vector (Vector) or pFLAG-CHIP (CHIP) was co-transfected without (−) or with pMyc-PRMT5 (+) for 48 h, and DMSO or MG132 was applied for another 6 h. Immunoblotting was performed to detect the indicated proteins. (B) CHIP, but not its mutants, promotes polyubiquitination of PRMT5. Myc-PRMT5 was co-expressed with HA-Ubiquitin, along with FLAG-Vector or FLAG-CHIP or its mutants into HEK293T cells for 48 h. Myc-PRMT5 was immunoprecipitated using anti-Myc antibody and was immunoblotted with anti-HA antibody. (C) Effect of CHIP and its mutants on the expression of PRMT5. Myc-PRMT5 was co-expressed with FLAG-CHIP and CHIP mutants (K30A and H260Q) for 48 h, and the expression level of Myc-PRMT5 (soluble and insoluble) was determined by immunoblotting. (D) Knockdown of CHIP blocks PRMT5 ubiquitination induced by 17-AAG treatment. pMyc-PRMT5 was co-transfected with pHA-Ubiquitin, along with siControl (siCon) or siCHIP for 60 h, and cells were treated with DMSO or 17-AAG for another 12 h. Whole cell lysate (WCL) was used for immunoprecipitation, and was immunoblotted with anti-HA antibody. (E) CHIP mediates PRMT5 polyubiquitination through K48-linkage chain. FLAG-CHIP was co-expressed with Myc-PRMT5, along with HA-Ubiquitin-WT (HA-Ub-WT) or the indicated mutants in HEK293T cells for 48 h, and WCL was used for immunoprecipitation with anti-Myc antibody and immunoblotted with anti-HA antibody. PRMT5-(Ub)n in B, C and D denotes the polyubiquitination of PRMT5.

In contrast, knockdown of CHIP decreased 17-AAG-induced polyubiquitination of PRMT5 (Fig. 6D), indicating the necessity of CHIP in ubiquitinating PRMT5. Since CHIP can function either as a partner of Ubc13-Uev1a to induce the formation of K63-linked polyubiquitin chains [50], or a mediator for K48-linked proteasomal degradation [41], we next sought to determine which types of ubiquitination may occur in PRMT5. To this end, two ubiquitin mutants, HA-Ubiquitin-K48R (HA-Ub-K48R) and HA-Ubiquitin-K63R (HA-Ub-K63R), were co-expressed with Myc-PRMT5 in the presence of FLAG-CHIP for 48 h, and anti-Myc antibody was used for immunoprecipitaion. As shown in Fig. 6E, a substantially reduced ubiquitination of PRMT5 was observed when Myc-PRMT5 was co-expressed with HA-Ub-K48R, but not HA-Ub-K63R, when compared with HA-Ub-WT, demonstrating that CHIP mediates K48-linked polyubiquitination of PRMT5. These results further support our finding that the CHIP/chaperone system (Hsp90/Hsp70) is involved in proteasomal degradation of PRMT5.

4. Discussion

PRMT5 is an emerging arginine methyltransferase that can epigenetically suppress the transcription of tumor suppressor genes and regulate the function of several signaling molecules through symmetrically dimethylating arginine residues of histones and non-histone substrates [51, 52]. Recently, overexpression of PRMT5 has been demonstrated to promote cell growth or inhibit cell death in multiple cancer cell lines, and is correlated with cancer development and progression in cancer patients [4, 7, 9, 11, 53]. The de-regulation of PRMT5 expression may occur at four different levels including transcription, post-transcription, translation and post-translation. We and others have previously demonstrated that PRMT5 can be transcriptionally activated by NF-Y [12] or post-transcriptionally regulated by miR-92b/96 [2]. However, whether the expression of PRMT5 can be regulated at post-translationally level remains elusive. In the present study, we first showed that PRMT5 undergoes polyubiquitination, and further demonstrated that CHIP as an E3 ubiquitin ligase interacts with PRMT5 and targets PRMT5 for ubiquitin-dependent proteasomal degradation. Results also revealed that 17-AAG-induced cell death and PRMT5 down-regulation is mediated through a CHIP-dependent mechanism.

Ubiquitination is a common type of post-translational modifications (PTMs) that regulates various cellular processes. The functional consequences of protein ubiquitination are highly dependent on the ubiquitination pattern (monoubiquitination vs polyubiquitination) and the ubiquitination linkage types [16]. At present, eight inter-ubiquitin linkage types such as K6, K11, K27, K29, K33, K48, K63, and linear ubiquitination have been reported [17, 39, 54]. Among them, K63 and K48 are the two most well-known ubiquitin-linked types. K63 ubiquitin linkage is involved in protein trafficking, and K48 ubiquitin linkage leads to proteasomal degradation [54]. Accumulated evidence suggests that CHIP can function as an E3 ubiquitin ligase, and thereby is responsible for fine-tuning protein homeostasis through K48-linked proteasomal degradation [54, 55]. Consistent with this, our results suggest that CHIP is required for ubiquitinating and targeting PRMT5 for proteasomal degradation. Several lines of evidence from our study support this conclusion. First, PRMT5 could undergo polyubiquitination, which is a prerequisite for proteasomal degradation (Fig. 1B–D). Second, co-immunoprecipitation, GST pull-down, and BiFC assays demonstrated the interaction between PRMT5 and CHIP both in vitro and in vivo (Fig. 3A–D). Third, the TPR domain of CHIP was sufficient for the interaction with PRMT5 (Fig. 3E and F), which is consistent with previous reports that the TPR domain is necessary for the interaction between CHIP and its substrates [56, 57]. Fourth, overexpression of CHIP dose-dependently decreased PRMT5 expression and shortened the half-life of PRMT5 (Fig. 4A, D and E). Fifth, overexpression of CHIP, but not its mutant (K30A, H260Q), mediated PRMT5 K48-linked ubiquitination (Fig. 4F, Fig. 6B, D); whereas knockdown of CHIP blocked 17-AAG-induced ubiquitination (Fig. 5D, Fig. 6C). However, H260Q also decreased the expression level of PRMT5 when overexpressed. Contrary to previous report that H206Q brings substrates into the insoluble fraction [28, 57], we did not see any significant increase of PRMT5 in the insoluble fraction (Fig. 6C). Therefore, it remains to be investigated whether CHIP may cooperate with other E3 ligases to ubiquitinate PRMT5 [58].

CHIP-mediated client protein degradation is often coupled with the molecular chaperone system including Hsp90 and Hsp70 [21]. Hsp90 inhibitors such as GA and 17-AAG have been on clinical trials in several human cancers [23, 46]. Their effects are mainly attributed by the disruption of chaperone function of Hsp90, and subsequent targeting of its client proteins for proteasomal degradation through associating with Hsp70 and E3 ubiquitin ligases [20, 43, 44]. Recent evidence has also shown that 17-AAG decreases PRMT5 protein expression (but not mRNA level) in ovarian cancer cell lines, suggesting that PRMT5 may be a potential client protein of Hsp90 [14, 45]. We showed here that Hsp90 inhibitors GA and 17-AAG dose-dependently inhibited PRMT5 protein expression in HEK293T cells and LNCaP cells (Fig. 5A and B), and overexpression of CHIP enhanced 17-AAG-induced PRMT5 reduction and cell death (Fig. 5C and D). Given that overexpressed PRMT5 is correlated with the development and progression of several human cancers [5], our results suggest that CHIP likely mediates the inhibitory effect of 17-AAG on cancer cell growth by promoting PRMT5 polyubiquitination and degradation via the chaperone/proteasomal degradation system.

Recent reports have shown that overexpression of CHIP blocks oncogenic signaling pathways, inhibits cell migration and anchorage independent growth, and induces cell death; whereas depletion of CHIP expression increases tumor formation and metastasis in mouse models [55, 59]. Interestingly, several studies have also demonstrated that the expression of CHIP in a number of cancers, such as breast cancer, gastric cancer, pancreatic cancer, and colorectal cancer [24, 26], is lower than the corresponding normal tissues, and that lower expression of CHIP appears to contribute to a lower survival rate (Supplementary Fig. S2). In these cancers, CHIP actually functions as a tumor suppressor by degrading a number of important oncogenic proteins, such as hypoxia inducible factor 1α [27], p65 [60], androgen receptor [59], c-Myc [28], EGFR [25], and histone deacetylase 6 [61]. Interestingly, PRMT5 is also overexpressed in these cancers. It is therefore tempting to hypothesize that the major tumor suppressor role of CHIP is through promoting the degradation of multiple oncogenic proteins such as PRMT5. In conclusion, the present study demonstrates that PRMT5 undergoes polyubiquitination, and that CHIP mediates ubiquitin-dependent proteasomal degradation of PRMT5 (Fig. 7). Given that lower expression of CHIP, and overexpression of PRMT5 have been observed in a number of cancers, it will be necessary to further evaluate the negative regulatory role of CHIP on PRMT5 expression in human cancer tissues.

Fig. 7.

The CHIP/chaperone system is involved in the regulation of PRMT5 expression. Molecular chaperone system (Hsp90 and Hsp70) plays an important role in maintaining the stability of a number of client proteins. PRMT5 is an aggregation-prone protein, and is also a client protein of Hsp90. Hsp90 assists the folding of PRMT5 into a fully functional molecule (denoted as correct folding). If misfolded PRMT5 is not corrected, it will trigger the CHIP/chaperone system, thereby ubiquitinating misfolded PRMT5 for subsequent proteasomal degradation. This model is also supported by the finding that treatment of Hsp90 inhibitors, such as 17-AAG and GA, enhances CHIP-mediated ubiquitination and degradation of PRMT5.

Highlights.

• PRMT5 undergoes polyubiquitination in LNCaP cells.

• The E3 ubiquitin ligase CHIP interacts with PRMT5 both in vivo and in vitro.

• CHIP/chaperone system mediates the ubiquitination and proteasomal degradation of PRMT5.

• CHIP mediates the inhibitory effect of 17-AAG on cell growth and down-regulation of PRMT5.

Abbreviations

PRMT5

protein arginine methyltransferase 5

EGFR

epidermal growth factor receptor

TNF

tumor necrosis factor

NF-Y

nuclear factor Y

PKC

protein kinase C

Hsp90

heat shock protein 90

UPS

ubiquitin-proteasome system

CHIP

carboxyl terminus of heat shock cognate 70-interacting protein

TPR

tetratricopeptide repeat

Hsp70

heat shock protein 70

HEK293T

human embryonic kidney 293T

CHX

cycloheximide

PMA

phorbol-12-myristate-13-acetate

GA

geldanamycin

17-AAG

17-(Allylamino)-17-demethoxygeldanamycin

WT

wild-type

MS

mass spectrometry

GST

Glutathione S-transferase

MEP50

methylosome protein 50

WCL

whole cell lysate