PRMT5-mediated arginine methylation activates AKT kinase to govern tumorigenesis

AKT is involved in a number of key cellular processes including cell proliferation, apoptosis and metabolism. Hyperactivation of AKT is associated with many pathological conditions, particularly cancers. Emerging evidence indicates that arginine methylation is involved in modulating AKT signaling pathway. However, whether and how arginine methylation directly regulates AKT kinase activity remain unknown. Here we report that protein arginine methyltransferase 5 (PRMT5), but not other PRMTs, promotes AKT activation by catalyzing symmetric dimethylation of AKT1 at arginine 391 (R391). Mechanistically, AKT1-R391 methylation cooperates with phosphatidylinositol 3,4,5 trisphosphate (PIP3) to relieve the pleckstrin homology (PH)-in conformation, leading to AKT1 membrane translocation and subsequent activation by phosphoinositide-dependent kinase-1 (PDK1) and the mechanistic target of rapamycin complex 2 (mTORC2). As a result, deficiency in AKT1-R391 methylation significantly suppresses AKT1 kinase activity and tumorigenesis. Lastly, we show that PRMT5 inhibitor synergizes with AKT inhibitor or chemotherapeutic drugs to enhance cell death. Altogether, our study suggests that R391 methylation is an important step for AKT activation and its oncogenic function.

Recently, protein arginine methylation has been identified as one of the most common PTMs 16 , which governs enzymatic activity, protein-protein interaction, protein localization, and stability 17,18 . Protein arginine methyltransferases (PRMTs) are responsible for transferring the methyl group to the guanidino nitrogen atoms of arginine, forming three types of methylated arginine residue: monomethylarginine (MMA), asymmetric dimethylarginine (aDMA), and symmetric dimethylarginine (sDMA) 19 . In mammals, nine PRMT family members have been identified, which are classified into three distinct types according to their catalytic activity. Type I enzymes (PRMT1, 2, 3, 4, 6, and 8) catalyze the formation of aDMA, whereas type II enzymes (PRMT5 and PRMT9) promote sDMA. Type III enzyme (PRMT7) only generates MMA. Notably, PRMT1 is the main type I enzyme responsible for about 80% of total arginine methylation, whereas PRMT5 is the dominant type II enzyme 20 . These PRMTs are involved in various fundamental cellular processes including DNA replication, transcription, RNA processing, DNA repair, and protein degradation [21][22][23] . Deregulation of PRMTs is associated with human diseases including cancers 24,25 .
Emerging evidence showed that PRMT5 is involved in regulating activation of the AKT signaling pathway. Depletion of PRMT5 impaired AKT phosphorylation and suppressed lung cancer cell proliferation 26 . Moreover, AKT phosphorylation was decreased in purified hematopoietic stem and progenitor cells isolated from Prmt5-null mice 27 . However, the precise mechanism by which PRMT5 controls AKT activation is unknown. Here we reported that PRMT5 directly methylates AKT1-R391 and cooperates with PI3K to promote AKT kinase activity and its oncogenic function.

Results
PRMT5 is required for AKT activation. To determine whether PRMTs are involved in regulation of AKT activation, we performed a CRISPR-Cas9-based screen to knock out individual PRMT gene in MCF7 breast cancer cells. PRMT8 was excluded in this screen, because its expression is restricted to the brain 28 . Notably, knockout of PRMT5, but not other PRMTs, markedly blocked phosphorylation of AKT and its downstream substrates including glycogen synthase kinase-3β (GSK-3β) and forkhead box class O 3a (FOXO3A) (Fig. 1a and Supplementary Fig. 1a-g). This observation was confirmed in other breast cancer cells including MDA-MB-231, T-47D, and MDA-MB-436, colon cancer cells DLD-1, and cervical cancer cells HeLa (Supplementary Fig. 1h-l), supporting that PRMT5-mediated AKT activation is unlikely to be cell-type dependent. Consistently, knockdown of PRMT5 by short hairpin RNA (shRNA) or pharmacological inhibition of PRMT5 by its specific inhibitor GSK591 29 led to the inactivation of AKT (Fig. 1b, c). Importantly, AKT1 immunopurified from PRMT5-knockout cells exhibited much lower kinase activity than AKT1 immunopurified from control cells in the in vitro kinase assay with GSK-3β as a substrate (Fig. 1d). Moreover, AKT activation was also largely abrogated in PRMT5depleted MCF7 cells in response to insulin stimulation (Fig. 1e).
To determine whether PRMT5 also regulates the activation of AKT2 and AKT3, we generated AKT1/2-KO and AKT1/3-KO in MDA-MB-231 cells that express all the isoforms of AKT (Supplementary Fig. 1m, n). Notably, knockout of PRMT5 in these cells also reduced AKT phosphorylation, suggesting impaired activation of AKT2 and AKT3 ( Supplementary Fig. 1o, p). Altogether, these results indicate a direct link between AKT kinase activity and PRMT5.
AKT is a critical downstream effector of PRMT5 on cell proliferation and tumor growth. As the main type II enzyme, PRMT5 mostly functions as an oncoprotein by governing cellular proliferative/apoptotic signaling 30 . In support of this notion, we found that depletion of endogenous PRMT5 led to reduced colony formation and anchorage-independent cell growth (Supplementary Fig. 2a, b). Moreover, re-introduction of PRMT5-WT, but not the PRMT5-E444Q mutant (the enzymatically dead form of PRMT5), could rescue AKT activation and cell proliferation of PRMT5-depleted cells ( Supplementary Fig. 2c-e), suggesting that the methyltransferase activity is required for the proliferative function of PRMT5. Given that AKT is a master regulator of cell proliferation, we investigated whether AKT is a downstream effector of PRMT5. Notably, ectopic expression of the constitutively active form of AKT1 (myr-AKT1) largely rescued AKT signaling and the colony-forming ability of PRMT5-depleted MCF7 cells (Fig. 2a-c). Moreover, knockdown of PRMT5 significantly suppressed tumor growth, which could be reversed by myr-AKT1 (Fig. 2d, e and Supplementary Fig. 2f). In keeping with these observations, immunohistochemistry (IHC) analysis of Ki-67 showed decreased proliferating cells in PRMT5-depleted tumors, which could be rescued by expressing myr-AKT1 (Fig. 2f). Consistently, ectopic expression of myr-AKT1 also suppressed cell apoptosis in PRMT5-depleted tumors as evidenced by a significant reduction of TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labeling)-stained cells (Fig. 2g, h). These results collectively demonstrate that PRMT5 exerts its oncogenic function in part through activating the AKT pathway.
PRMT5 is the arginine methyltransferase of AKT1. Next, we sought to investigate whether AKT is a bona fide substrate of PRMT5. To this end, we found that PRMT5 specifically coimmunoprecipitated AKT1, but neither S6K1 nor PDK1, another two members of the AGC kinase family 31 ( Fig. 3a and Supplementary Fig. 3a). Reciprocally, AKT1 interacted with PRMT5 but not PRMT1 (Fig. 3b). Moreover, PRMT5 bound all isoforms of AKT including AKT1, AKT2, and, to a lesser extent, AKT3 ( Supplementary Fig. 3b). We further mapped which domain of AKT1 interacts with PRMT5. AKT1 consists of an N-terminal PH domain, a middle kinase domain (KD), and a C-terminal regulatory region (Tail) 2 . We found that PRMT5 bound to the AKT-KD and, to a lesser extent, the PH domain ( Supplementary  Fig. 3c, d). Consistent with these binding results, PRMT5, but not PRMT1 nor PRMT9, could catalyze the sDMA formation on AKT1, but not S6K1, in an enzymatic activity-dependent manner (Supplementary Fig. 3e-g). On the other hand, depletion of PRMT5 blocked AKT1 sDMA formation ( Supplementary  Fig. 3h). Importantly, in vitro methylation assays demonstrated that PRMT5 directly methylates AKT1 in a methyltransferase activity-dependent manner (Fig. 3c).
To identify the arginine residue(s) that is methylated by PRMT5, we analyzed the protein sequence of AKT1 using methylation prediction tools including PRmePred 32 and GPS-MSP 33 . Six arginine residues ranked top scores were selected for further analyses (Supplementary Fig. 3i). Interestingly, the substitution of R391 to K (R391K) completely blocked AKT1 sDMA formation in cells ( Supplementary Fig. 3j). Of note, a recent study on bioRxiv showed that AKT1-R15 is methylated in neuroblastoma cells (https://doi.org/10.1101/ 2020.08.12.246660). However, we found that R15K mutation fails to inhibit AKT1 sDMA signal in our cell lines (Supplementary Fig. 3j, k), indicating the site of methylation in AKT1 possibly occurs in a cell-type-specific manner. Moreover, highresolution liquid chromatography with tandem mass spectrometry (MS/MS) analyses confirmed dimethylation of AKT1-R391 in our cells ( Supplementary Fig. 3l). Notably, AKT1-R391 is evolutionarily conserved from Drosophila to Homo sapiens (Fig. 3d), indicating a potential role of this methylation event in various species.
To further confirm AKT1-R391 methylation, we generated a polyclonal antibody that specifically recognizes symmetric dimethyl R391 (R391-me2s), but not the unmodified, monomethyl, or asymmetric dimethyl R391 as assessed in the dot blot assays ( Supplementary Fig. 3m). R391K mutation completely blocked PRMT5-mediated methylation of AKT1 in vitro and in cells (Fig. 3e, f and Supplementary Fig. 3n). Moreover, AKT1-R391-me2s was abolished in PRMT5-knockout cells (Fig. 3g). Consistently, AKT2 and AKT3 were also methylated at corresponding residues, mutating of which blocked the R391-me2s signal ( Supplementary Fig. 3o, p). Lastly, we found that the expression levels of PRMT5 is positively correlated with AKT1-R391-me2s signal in breast cancer patient specimens (Fig. 3h, i). Altogether, these results suggest that PRMT5 is the physiological arginine methyltransferase responsible for the symmetric methylation of AKT1-R391.
AKT1-R391 methylation plays a critical role in AKT activation processes. We next sought to explore whether, on the other hand, PRMT5-mediated R391 methylation plays a role in PI3Kdependent AKT activation processes. Previous studies have illustrated that without the growth factor stimulation, intramolecular PH domain and KD domain interactions maintain AKT in a closed conformation, called PH-in 35,36 . Notably, AKT1-R391K mutation or depletion of PRMT5 strengthened the PH-KD interaction ( Fig. 5a and Supplementary Fig. 5a), and, inversely, ectopic expression of PRMT5 decreased KD-WT, but not the KD-R391K mutant, interaction with the PH domain ( Supplementary  Fig. 5b). Interestingly, PRMT5 also perturbed KD interaction with PH-R25C mutant ( Supplementary Fig. 5c), indicating that PRMT5 promotes AKT confirmation change independent of its binding to PIP3. Importantly, the symmetric dimethyl R391 peptide displayed a much lower PH-binding capacity than the corresponding non-modified peptide (Fig. 5b). Moreover, abrogation of AKT methylation through R391K mutation or PRMT5 depletion diminished the interaction between AKT1 and PIP3 ( Fig. 5c, d), and consequently reduced AKT1 membrane recruitment (Fig. 5e, f and Supplementary Fig. 5d-f). These data suggest that PRMT5-mediated R391 methylation may weaken intramolecular PH-KD binding to facilitate AKT conformational shift from the PH-in to PH-out state and membrane translocation.
To gain further insights into the molecular mechanism by which methylation of AKT1-R391 regulates phosphorylation of AKT, we found that the interaction between AKT1 and PDK1 was decreased in AKT-R391 methylation-deficient cells (Fig. 5g, h and Supplementary Fig. 6a, b), which mimics the behavior of the AKT1-R25C mutant ( Supplementary Fig. 6c Flag-PRMT5/MEP50 was transiently transfected into these cells. g Knockout of PRMT5 abrogates AKT1-R391 methylation. IB analysis of WCLs and IP products derived from PRMT5-knockout cells or control cells transfected with HA-AKT1. h Representative images of IHC staining for PRMT5 and R391-me2s in a breast cancer tissue array (n = 100 tissue specimens). Scale bar, 50 μm. i Quantification of cases with PRMT5 and R391-me2s staining (n = 100 tissue specimens). P-values were calculated using two-tailed χ 2 -test. Similar results were obtained in n ≥ 2 independent experiments in a-c and e-g.  We also observed that AKT1-R391K mutation or depletion of PRMT5 impairs the interaction between AKT1 and Sin1 ( Supplementary Fig. 6f, g), the subunit of mTORC2 complex mediating its association with substrate 37 . These results suggest that AKT-R391 methylation enhances AKT interaction with its upstream kinases, PDK1 and mTORC2.
Lastly, analysis of the AKT1 structure showed that R391 participates in a conserved network of salt bridges and hydrogen bonds with E319, A329, and K386 ( Supplementary Fig. 6h). Notably, this serves to position a loop that harbors the conserved "DYG" motif, which participates in a network of interactions with the AKT1 activation loop and the pT308 side chain (PDB: 3CQW) 38 , and serves as a crucial regulatory region of many kinases 39 . Modeling of symmetrically dimethylated R391 onto AKT1 showed that this network of interactions is disrupted both due to a loss of hydrogen bond donors on the dimethylated R391 side chain as well as the steric clashes that occur ( Supplementary Fig. 6i). Given the connection of this region to the DYG motif and activation loop, which includes T308, we speculate that dimethylation of R391 may result in a promotion of T308 phosphorylation via an allosteric effect.
Deficiency in AKT1-R391 methylation suppresses AKT activation and oncogenic function. Given that AKT1-R391 methylation is critical for AKT signaling activation, we next investigated whether it governs AKT biological functions. To this end, we reconstituted AKT1-WT and the methylation-deficient mutant of AKT1 (R391K) in DLD-1 cells, in which both AKT1 and AKT2 were knocked out (DLD-1-AKT1/2 −/− ). Compared to AKT1-WT, the AKT1-R391K mutation significantly reduced AKT activation at the basal level and in response to stimulation by growth factors (Fig. 6a, b and Supplementary Fig. 7a). Consistently, the kinase activity of R391K was dramatically decreased as shown by in vitro kinase assay with GSK-3β as a substrate (Fig. 6c). We also observed similar phenotypes in MCF7 cells ( Supplementary Fig. 7b-e). Moreover, AKT-pS473 was dramatically enriched at the membrane in cells expressing AKT1-WT, but not the R391K mutant ( Supplementary Fig. 7f). Previous studies demonstrated that TSC1-TSC2 complex and oncogenic Ras mutant directly associate with mTORC2 to promote its kinase activity 40,41 . Interestingly, PRMT5 also regulated TSC2 or oncogenic Ras-mediated phosphorylation of AKT-S473 (Supplementary Fig. 7g, h), probably through modulating AKT interaction with mTORC2 ( Supplementary Fig. 6f, g).
In keeping with the much lower kinase activity of R391K mutant, AKT1/2-KO cells expressing this mutated protein displayed a dramatic reduction in cell proliferation, colony formation, and anchorage-independent cell growth, compared to AKT1/2-KO cells expressing AKT1-WT protein (Fig. 6d-f and Supplementary Fig. 8a-h). Consistent with the minor effect on AKT1 arginine methylation, R15K mutation did not disturb AKT activation and its proliferative function ( Supplementary Fig. 8i-k).
To pinpoint the potential role of AKT-R391 methylation in tumorigenesis in vivo, we performed xenograft mouse assays.
Notably, compared to AKT1-WT, AKT1-R391K mutation significantly retarded tumor growth (Fig. 6g-i). TUNEL staining and IHC analysis of Ki-67 demonstrated an elevation of cell apoptosis and a decrease of proliferating cells in tumors expressing AKT1-R391K, compared to tumors expressing AKT1-WT (Fig. 6j-l). Importantly, depletion of PRMT5 markedly suppressed AKT activation and cell proliferation in AKT1-WT cells and, to a much lesser extent, AKT1-R391K cells ( Supplementary Fig. 8l, m). Altogether, these results demonstrate that PRMT5-mediated AKT-R391 methylation promotes AKT activation and its oncogenic function.
PRMT5 inhibitor synergizes with AKT inhibitor and chemotherapeutic drugs. PRMT5 has recently emerged as an attractive drug target for cancer therapies. Several PRMT5 inhibitors, such as GSK3326595, are on clinical trial for the treatment of a variety of cancers including breast cancer 42 . We next examine whether PRMT5 inhibitors suppresses breast cancer cell proliferation. Notably, GSK3326595 treatment significantly decreased cell survival in a time-dependent manner in a panel of breast cancer cells (Supplementary Fig. 9a-d), which is consistent with previous report 43 . Although many AKT inhibitors including AZD5363, MK2206, and GDC-0068 are in phase I/II clinical trials, none of these inhibitors has achieved satisfactory outcomes in monotherapy in part due to the inadequate AKT inhibition at tolerable doses. Therefore, a combination of AKT inhibitors with chemotherapy and other targeted therapy is a potential option for cancer treatment 44 . Interestingly, co-treatment of PRMT5 and AKT inhibitors exhibited a synergistic effect on promoting cell death in various breast cancer cells including estrogen receptor/ progesterone receptor (ER/PR) positive cell lines MCF7 and T-47D, and triple-negative breast cancer cell lines MDA-MB-231, BT-549, and MDA-MB-468 (Fig. 7a, b). Previous studies also demonstrated that activation of AKT contributes to chemotherapy resistance in breast cancer 45,46 . As PRMT5 inhibitor markedly blocked AKT activation, we next examined whether PRMT5 inhibition affects cellular sensitivity to chemotherapeutic drugs. Notably, GSK3326595 sensitized breast cancer cells to etoposide and cisplatin (Fig. 7c-f). Of note, we did not observe additional effects of PRMT5 inhibitor and cisplatin in MDA-MB-468 cells, as this cell line is susceptible to cisplatin. These results suggest that targeting PRMT5 is a potential strategy to improve the cellular response to AKT inhibitors or chemotherapeutic agents.

Discussion
The AKT/PKB protein kinases are evolutionally conserved from Drosophila to human, which are regulated by multiple upstream pathways in response to growth factor, nutrients, and hormone 2 . Hyperactivation of the AKT oncogenic pathway is a hallmark of human cancers, which can be caused by genetic alterations in AKT itself or its upstream regulators 5 . Notably, PIK3CA mutation or PTEN loss was frequently observed in a variety of human cancers 47,48 , both of which lead to PIP3 accumulation and subsequent aberrant AKT activation 49,50 . Here we revealed that AKT1-R391 methylation promotes AKT kinase activity by weakening PH-KD interaction, leading to a conformational shift from the closed state to semi-open state and subsequently promoting PI3K (PIP3)-dependent membrane translocation (open state), and interaction with PDK1 and mTORC2 for phosphorylation at T308 and S473, respectively (Fig. 7g). Interestingly, this methylation event is enhanced when AKT is in an open state, suggesting a positive feedback loop between AKT-R391 methylation and AKT activation. Importantly, deficiency in AKT-R391 methylation attenuates breast tumorigenesis. Therefore, our results demonstrate that AKT1-R391 methylation is a critical step for PI3K-mediated AKT activation and oncogenic function. As an epigenetic modifier of histone and non-histone proteins, PRMT5 exerts its biological functions at both cytoplasm and nucleus 51 . PRMT5 was first characterized as a transcriptional repressor by catalyzing symmetric dimethylation of histone H3 arginine 8 and H4 arginine 3, leading to transcriptional repression of tumor suppressor genes 52 . PRMT5 also regulates transcription programs by methylating transcription factors, mRNA splicing factors, and RNA-binding proteins 25 . In the cytoplasm, PRMT5 has been reported to regulate RAS-ERK and nuclear factor-κB (NF-κB) pathways by methylating epidermal growth factor receptor (EGFR) and NF-κB/p65, respectively 53,54 . Here we revealed that PRMT5 exerts its oncogenic functions, in part, by promoting the AKT pathway through directly methylating AKT1 at R391. Surprisingly, we found that a significant portion of PRMT5 is associated with the membrane, which is probably derived from Golgi membranes as previously reported 55 . Further study is necessary to classify the mechanism and function of the membrane-associated PRMT5.
PRMT5 is overexpressed in various cancers and correlated with poor prognosis, low survival, and drug resistance [56][57][58] . Notably, >50% of breast tumors display higher PRMT5 expression compared to normal breast tissues 56 . Moreover, overexpression of PRMT5 is associated with chemotherapeutic resistance in breast cancer 59 . Furthermore, depletion of PRMT5 significantly inhibited breast cancer cell proliferation and tumor growth in mouse model 60 . Thus, PRMT5 is considered as a potential therapeutic target for breast cancer. Recently, several PRMT5 inhibitors have been developed including EPZ015666, GSK591, and GSK3326595 61 . Notably, the efficacy of GSK3326595 is being tested in clinical trials 42 . Our data show that PRMT5 inhibitor enhances cellular sensitivity to AKT inhibitor and chemotherapy drugs. Of note, we did not observe an apparent association between PI3KCA mutations and sensitivity to PRMT5 inhibitor in these breast cancer cells we examined. Further studies using more breast cancer cell lines and mouse models are required to demonstrate whether PI3KCA mutations dictate the efficacy of PRMT5 inhibitor. As AKT inhibitor monotherapy had limited clinical activity in advanced breast cancer patients in part due to inadequate AKT inhibition at tolerable doses 62 , our study provides a rational for the combination of PRMT5 and AKT inhibitors or chemotherapeutic drugs to achieve better clinical outcomes.
Cell fractionation. This assay was performed using Subcellular Protein Fractionation Kit following the manufacturer's instructions (Thermo Fisher, 78840). The membrane fraction contains plasma, mitochondria, and ER/Golgi membranes, but not nuclear membranes.
Immunoblot and immunoprecipitation analyses. Cells were rinsed with ice-cold phosphate-buffered saline (PBS) and lysed with EBC buffer (50 mM Tris pH 7.5, 120 mM NaCl, 0.5% NP-40) or Triton buffer (40 mM HEPES pH 7.4, 150 mM NaCl, 2.5 mM MgCl 2 , and 1% Triton) supplemented with protease inhibitor (Thermo Fisher, A32953) and phosphatase inhibitors (phosphatase inhibitor cocktail sets I and II, Calbiochem). The cell lysates were centrifuged at 13,200 r.p.m. for 10 min at 4°C. The protein concentrations were measured using Nanodrop (Thermo Fisher) with Bio-Rad protein assay reagent. Equal amounts of whole cell lysates were resolved by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotted with the indicated antibodies. For immunoprecipitation, the lysates (1000-3000 μg) were incubated with 50% slurry of agarose conjugated antibody for 3-5 h at 4°C. The immunoprecipitates were washed four times with NETN buffer (20 mM Tris pH 8.0, 150 mM NaCl, 1 mM EDTA, and 0.5% NP-40) or Triton buffer before being resolved by SDS-PAGE. Anti-FLAG agarose beads (A2220) and anti-HA agarose beads (A2095) were purchased from Sigma. Anti-AKT1 antibody conjugated to agarose (sc-5298 AC) and anti-PRMT5 antibody conjugated to agarose (sc-376937 AC) were purchased from Santa Cruz Biotechnology.
Purification of GST-tagged protein from bacteria. Recombinant GST-GSK-3β protein was purified from the BL21(DE3) Escherichia coli transformed with pGEX-GSK-3β. Single colony was grown in 5 mL Luria-Bertani (LB) medium overnight at 37°C. The culture was then 1:100 inoculated into 300 mL LB medium until an optical density of 0.4-0.6. The protein expression was induced by 0.1 mM IPTG (isopropyl-β-D-thiogalactoside) at 25°C for 16 h. The bacteria cells were collected and re-suspended in GST buffer (25 mM Tris pH 8.0, 5 mM dithiothreitol (DTT), 150 mM NaCl) and sonicated (6 cycles of 5 s each at 50%). After centrifugation, the supernatant was incubated with glutathione sepharose beads for 3 h at 4°C. The protein-bound glutathione beads were washed three times with GST buffer and eluted with elution buffer (10 mM L-Glutathione, 50 mM Tris-HCl pH 8.0).
In vitro methylation assays. PRMT5 in vitro methylation assays were performed as previously described 66 . Briefly, 5 μg of recombinant GST-AKT1 proteins purified from HEK293T cells or GST-AKT1-KD-WT or R391K proteins purified from bacteria were incubated with immunoprecipitated HA-PRMT5 or HA-PRMT5-E444Q in the presence of 1 μL of adenosyl-L-methionine, S-[methyl-3 H] (1 mCi/ml stock solution, Perkin Elmer). The reactions were performed in the methylation buffer (50 mM Tris pH 8.5, 20 mM KCl, 10 mM MgCl 2 , 1 mM β-mercaptoethanol, and 100 mM sucrose) at 30°C for 1 h and stopped by adding 3× SDS loading buffer and was resolved by SDS-PAGE. The separated samples were then transferred from the gel to a polyvinylidene difluoride membrane, which was then sprayed with EN 3 HANCE (Perkin Elmer) and exposed to X-ray film.
Mass spectrometric analysis of AKT1-R391 methylation. HEK293T cells were transfected with HA-AKT1. Forty-eight hours post transfection, the cells were lysed in Triton buffer, following HA immunoprecipitation. The immunoprecipitates were resolved by SDS-PAGE and visualized using GelCode blue staining reagent. The protein band containing HA-AKT1 was excised and digested with trypsin. Peptides were analyzed on an EASY nLC 1200 in-line with the Orbitrap Fusion Lumos Tribrid mass spectrometer (ThermoScientific). Peptides were pressure loaded at 800 bar and separated on a C18 reversed phase column (Acclaim PepMap RSLC, 75 µm × 50 cm (C18, 2 µm, 100 Å)) (Thermo Fisher) using a gradient of 2-35% B in 180 min (Solvent A: 0.1% FA; Solvent B: 80% ACN/0.1% FA) at a flow rate of 300 nL/min at 45°C. Mass spectra were acquired in datadependent mode with a high resolution (60,000) Fourier Transform mass spectrometry (FTMS) survey scan followed by MS/MS of the most intense precursors with a cycle time of 3 s. The automatic gain control target value was 4.0e5 for the survey MS scan. Fragmentation was performed with a precursor isolation window of 1.6 m/z, a maximum injection time of 50 ms, and higher-energy collisional dissociation (HCD) collision energy of 35%; the fragment ions were detected in the Orbitrap at a 15,000 resolution. Spectra were searched against a custom database containing human AKT1 and a database of common contaminants using Max-Quant. The false discovery rate, determined using a reversed database strategy, was set at 1% at the peptide and modification site levels. Fully tryptic peptides with a minimum of seven residues were required including cleavage between lysine and proline. Two missed cleavages were permitted. Sites of modification were manually verified.
IHC staining. A breast cancer tissue microarray containing 100 cases of invasive carcinoma (BC081120f) was purchased from Biomax. The xenograft tumors were fixed in 10% neutral buffered formalin for 24 h and then paraffin-embedded and processed at the Histology & Immunohistochemistry Laboratory at Medical University of South Carolina (MUSC). Formalin-fixed, paraffin-embedded (FFPE) sections were deparaffinized using xylene and rehydrated in graded ethanol. Sections were heated in boiled citrate buffer (pH 6.0) for for 15 min. The remaining steps were performed using the ImmPRESS® Excel Amplified Polymer Kit (Vector Laboratories, MP-7601). Briefly, the sections were incubated with BLOXALL Blocking Solution (SP-6000) for 10 min, washed with wash buffer (10 mM sodium phosphate, pH 7.5, 0.9% saline and 0.1% Tween 20) for 5 min, and blocked with 2.5% Normal Horse Serum for 20 min. Sections were then incubated with anti-PRMT5 antibody (Proteintech, 18436-1-AP, 1:1000), anti-R391-me2s antibody (1:1000), or anti-Ki-67 antibody (Cell Signaling Technology, 9027, 1:800) diluted in normal horse serum for overnight at 4°C. Sections were then washed with PBST buffer for 5 min and incubated with Goat anti-rabbit IgG (1:500) for 15 min at room temperature. After washing with PBST, sections were incubated with ImmPRESS Polymer Reagent for 30 min followed by washing with PBST. Afterward, all sections were developed using ImmPACT DAB EqV working solution until desired stain intensity and counterstained with hematoxylin, and mounted using SHURMount Mounting Media (General Data, #682188). The levels of PRMT5 and R391-me2s were determined by H-Score as described previously 67 .
TUNEL assay. This assay was performed using the CF488A-TUNEL Apoptosis Assay Kit (Biotium, #30063). The FFPE sections were deparaffinized using xylene and rehydrated in graded ethanol. After washing with PBS, sections were permeabilized with 20 mg/mL proteinase K in PBS for 30 min at 37°C. After incubating with 100 μL TUNEL Equilibration Buffer (Component 99965) for 5 min, sections were incubated with 50 μL of TUNEL reaction mix for 2 h at 37°C and counterstained with 4′,6-diamidino-2-phenylindole (DAPI) and mounted using vibrance antifade mounting medium (Vector Laboratories, H-1700). The TUNELpositive cells were determined by ImageJ.
Immunofluorescence. Cells grown on glass coverslips were fixed with 4% paraformaldehyde for 15 min at room temperature, washed three times with PBS, and then permeabilized with 0.05% Triton X-100 for 10 min at room temperature. Following three washes of 5 min in PBS, the coverslips were blocked with 5% bovine serum albumin for 30 min and then incubated with anti-AKT1 (Cell Signaling Technology, 2938, 1:300) or anti-AKT-pS473 (Cell Signaling Technology, 4060, 1:400) antibody for overnight at 4°C. Following three washes of 5 min in PBST, the coverslips were incubated with goat anti-Rabbit IgG Alexa Fluor 488 (Invitrogen, #A-11008, 1:4 00) for 1 h at room temperature in the dark. Following three washes of 5 min with PBST, the coverslips were stained with DAPI and mounted using vibrance antifade mounting medium (Vector Laboratories, H-1700).
Colony formation assays. Cells were seeded in 6-well plates (400 cells per well) and cultured for 8-12 days until the formation of visible colonies. The colonies were fixed with 10% acetic acid and 10% methanol for 20 min, and then stained with 0.4% crystal violet in 20% ethanol for 20 min. After staining, the plates were washed with distilled water and air dried. The colonies were counted manually.
Soft agar assays. The anchorage-independent cell growth assay is a common method used to detect the ability of malignant transformation of cells and were performed as described previously 63 . Briefly, 1 × 10 4 or 2 × 10 4 cells were mixed with noble agar at a final concentration of 0.4% and layered over the bottom layer containing 0.8% noble agar. Completed DMEM medium (500 μl) was added to keep the top layer moist. The cells were then cultured for 3-4 weeks and stained with 1 mg/ml iodonitrotetrazolium chloride for colony visualization and counting manually.
Mouse xenograft assays. DLD-1-AKT1/2 −/− (2 × 10 6 ) cells stably expressing AKT1-WT or AKT1-R391K mutant or DLD-1 cells depleted PRMT5 and/or overexpressing myr-AKT1 were injected into the flank of 5-week-old female nude mice (The Jackson Laboratory). Tumor size was measured every 2 days with an electronic caliper. The tumor volume was calculated using the following formula: L × W 2 × 0.5, where L is the longest diameter and W is the shortest diameter. After 19-21 days, mice were killed. The solid tumors were dissected and weighed. All mice are housed in 22°C, 50-60%, humidity and a 12 h-light/12 h-dark cycle. All mouse experiments were conducted under Protocol #IACUC-2018-00604 approved by the MUSC Institutional Animal Care and Use Committee.
Statistical analysis. The in vitro experiments were repeated at least two times unless stated otherwise. As indicated in the figure legends, all quantitative data are presented as the mean ± SD or mean ± SEM of three biologically independent experiments or samples. Statistical analyses were performed using GraphPad Prism 9 and Excel version 16.48. Statistical significance was determined by two-tailed Student's t-test or two-way analysis of variance. P-value < 0.05 was considered significant.