TRMU Confers Resistance of Melanoma Cells to Vemurafenib through Modulating Mitochondrial Activities

Abstract

Approximately 50% of melanoma patients carry a mutation in the BRAF gene, and over 90% of these mutations lead to the substitution of valine 600 with glutamic acid (V600E). Vemurafenib is an FDA-approved kinase inhibitor for BRAF^V600E; while the drug elicits effective remission of metastatic melanoma, relapse typically occurs within several months after therapy. Recent studies documented critical roles of reversible modifications in RNA in modulating resistance to cancer therapy. Herein we explored the contributions of epitranscriptomic alterations to vemurafenib resistance by assessing the differential expression of epitranscriptomic reader, writer and eraser (RWE) proteins in IGR37 metastatic melanoma cells and the isogenic vemurafenib-resistant cells (IGR37xp). Our results revealed altered expressions of multiple epitranscriptomic RWE proteins, including markedly elevated expressions of MTO1 and TRMU—which act sequentially to produce 5-taurinomethyl-2-thiouridine (τm⁵s²U) at the 34th position of human mitochondrial (mt) tRNA^Glu, tRNA^Gln and tRNA^Lys—in the resistant line. We also observed elevated oxidative phosphorylation in IGR37xp relative to IGR37 cells. Moreover, we found that genetic depletion of TRMU in IGR37xp cells results in diminished oxidative phosphorylation and resensitizes IGR37xp cells to vemurafenib. Together, we uncovered a role of TRMU in conferring vemurafenib resistance in melanoma through modulating oxidative phosphorylation.

Graphical Abstract

INTRODUCTION

Melanoma is the fifth most common cancer incidence in the U.S.¹ Approximately 50% of melanoma patients carry mutations in the Ser/Thr-kinase BRAF, most commonly with valine 600 being mutated to a glutamic acid (V600E), which leads to constitutive activation of the mitogen-activated protein kinase (MAPK) pathway, thereby promoting cancer cell survival, proliferation and metastasis.² Vemurafenib, a selective chemical inhibitor of the oncogenic BRAF kinase, is an FDA-approved drug that is used in combination with MEK inhibitor cobimetinib for treating metastatic melanoma with BRAF V600 mutations.³ However, patients relapse with lethal drug-resistant diseases within months after therapy, which is largely attributed to mutation accrual in a small population of cells after drug administration.^4,5

Studies have been conducted to compare the genomic and transcriptomic landscapes before and after the development of the resistance phenotype so as to identify pathways driving vemurafenib resistance. By harnessing whole-exome sequencing of patients’ biopsies before and after vemurafenib treatment, van Allen et al.⁶ revealed a plethora of mutations in genes associated with cancer signaling pathways and underscored a role of mutations of genes encoding components of the MAPK pathway in conferring vemurafenib resistance in melanoma cells. In addition, CRISPR screening experiments showed that depletion of an E3 ubiquitin-protein ligase (CUL3), transcriptional adapters (TADA1/2B), cytokine-like nuclear factor N-PAC (NF1/2), and mediator of RNA polymerase II transcription subunit 12 (MED12) conferred resistance to vemurafenib.⁷ Microarray-based transcriptomic analysis also unveiled anaplastic lymphoma kinase (ALK) as a driver for vemurafenib resistance.⁸

Epitranscriptomics involves the studies about chemical modifications on RNA and their regulatory roles in biological processes.⁹ These modifications are deposited and removed by ‘writer’ and ‘eraser’ enzymes, respectively, and modulate the binding affinities of RNA to proteins. N⁶-methyladenosine (m⁶A) is a well-studied epitranscriptomic modification, and it is installed, removed, and recognized by METTL3/METTL16,¹⁰ ALKBH5/FTO,¹¹ and YTH domain family (YTHDF) proteins,¹² respectively. m⁶A in mRNAs is known to regulate the stabilities and translation efficiencies of the mRNAs.¹³

A previous study showed that the sensitivity to BRAF inhibitors was dictated by a subset of highly expressed mRNAs with METTL3-installed m⁶A in their 5’-untranslated regions.¹⁴ Apart from modifications in mRNAs, tRNA modifications are also linked with therapeutic resistance. For instance, enzymes modifying U34 in cytosolic tRNAs (e.g., ELP1, ELP3, CTU1, CTU2) are upregulated in BRAF V600E melanoma and their depletion sensitized melanoma cells to vemurafenib treatment.¹⁵ In esophageal cancer, NAT10-installed N⁴-acetylcytidine (ac⁴C) on tRNA confers resistance to an EGFR inhibitor, which can be overcome by coadministration with an NAT10 inhibitor.¹⁶ Moreover, METTL1/WDR4-installed m⁷G in tRNAs promotes resistance to tyrosine kinase inhibitor lenvatinib in hepatocellular carcinoma¹⁷ and to topoisomerase 2 inhibitor doxorubicin in osteosarcoma patients.¹⁸ It remains unexplored whether other epitranscriptomic modulators also contribute to the development of acquired resistance toward vemurafenib in melanoma cells.

We reasoned that a comprehensive profiling of the changes in expression levels of epitranscriptomic reader, writer, and eraser (RWE) proteins constitutes a crucial step toward understanding the role of epitranscriptomic regulation in acquired resistance. In this study, we employed a previously developed parallel-reaction monitoring (PRM)-based targeted proteomic method¹⁹ to examine the differential expression of epitranscriptomic RWE proteins in IGR37, a metastatic melanoma cell line with the BRAF^V600E mutation, and its isogenic vemurafenib-resistant counterpart, i.e., IGR37xp cells.^8,20 We found that increased expression of TRMU, a mitochondrial tRNA-modifying enzyme, contributes to the acquired resistance of melanoma cells to vemurafenib through modulating oxidative phosphorylation.

MATERIALS AND METHODS

Cell Culture

IGR37 and the isogenic vemurafenib-resistant IGR37xp melanoma cells were kindly provided by Dr. Stephanie Kreis.²⁰ The IGR37 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Gibco) supplemented with 10% fetal bovine serum (FBS, Corning) and 1% penicillin–streptomycin solution (GE Healthcare). IGR37xp cells were maintained in DMEM complete medium with 5 μM vemurafenib (Selleckchem) to maintain its drug-resistant phenotype. The cells were cultured in a humidified incubator with 5% CO₂ at 37 °C.

Stable isotope labeling by amino acids in cell culture (SILAC) experiments involved culturing cells in heavy or light DMEM media for at least 6 passages. The SILAC heavy medium was prepared by adding [¹³C₆,¹⁵N₂]-l-lysine (Lys-8) and [¹³C₆]-l-arginine (Arg-6, Cambridge Isotope Laboratories) to lysine, arginine-depleted DMEM base medium (Thermo Fisher) until their final concentrations reached 0.798 and 0.398 mM, respectively, whereas the light medium was prepared in a similar way with the use of unlabeled lysine and arginine. The media also contained 10% dialyzed FBS (HyClone, Cytiva) and 1% penicillin-streptomycin (Thermo Fisher).

Cell Lysis and Proteomic Sample Preparation

IGR37 and IGR37xp cells were harvested, and subsequently lysed using CelLytic M cell lysis reagent (Sigma) supplemented with a protease inhibitor cocktail (Sigma). Protein concentrations were quantified by using Quick Start Bradford Protein Assay (Bio-Rad). Following a filter-aided sample preparation (FASP) method²¹ with a minor modification, 25 μg each of heavy- and light-labeled lysates were combined (Figure 1), and denatured twice with 8 M urea in 50 mM NH₄HCO₃ in a poly(ether sulfone) (PES) membrane centrifugal filter unit (VWR) with 30 kDa molecular weight cutoff. The denatured samples were reduced with 20 mM dithiothreitol at 37 °C for 1 h, alkylated with 55 mM iodoacetamide at room temperature in the dark for 20 min, followed by washing four times with 50 mM NH₄HCO₃. The samples were digested with MS-grade trypsin (Pierce) at a 1:50 trypsin-to-protein ratio (by mass) in 50 mM NH₄HCO₃ at 37 °C overnight. The tryptic peptides were collected by centrifugation, dried in a Speed-vac, desalted using Pierce C18 tips (Thermo Fisher), and redissolved in 0.1% formic acid for LC-MS analysis.

Figure 1.

SILAC in conjunction with scheduled PRM for assessing the differential expression of epitranscriptomic RWE proteins in IGR37 and IGR37xp cells. Epitranscriptomic RWE proteins were quantified as the average ratios of their constituting peptides. In forward SILAC experiments, light-isotope-labeled IGR37 lysates were mixed at a 1:1 ratio (by mass) with heavy-isotope-labeled IGR37xp lysates, whereas reverse SILAC experiments were conducted in the opposite way. The mixed cell lysate was digested with trypsin and subjected to scheduled LC-PRM analysis. The LC-MS/MS data were processed using Skyline.

LC-PRM Analysis and Data Processing

The tryptic peptide mixtures were subjected to LC-PRM analysis on an Orbitrap Fusion Lumos Tribrid mass spectrometer (Thermo) coupled with an Easy nLC 1000 (Thermo) and a Flex nanoelectrospray ion source (Thermo). The mass spectrometer was equipped with a high-field asymmetric-waveform ion mobility spectrometry (FAIMS). The compensation voltages (CV) were set at −40, −60, and −80 V, each for 1/3 of a 3-s cycle, where the carrier gas flow rate was 4.2 L/min. The analytical column was packed in-house using 3 μm Reprosil-Pur C18-AQ resin (Dr. Maisch GmbH HPLC) in a ~ 25 cm long, 75-μm i.d. fused silica column. The trapping column was also prepared in-house using 5 μm Reprosil-Pur C18-AQ resin (Dr. Maisch GmbH HPLC) in a 4 cm long, 150 μm i.d. fused silica column. The LC gradient included 6–43% of buffer B (0.1% formic acid in 80% acetonitrile) at a flow rate of 0.3 μL/min. The spray voltage was 2 kV, and the ion transfer tube temperature was 320 °C. Precursor ions were isolated in the quadrupole at an isolation window of 2.0 Th, fragmented in the higher-energy collisional dissociation (HCD) cell at a normalized collision energy (NCE) of 30. The MS/MS were acquired in the Orbitrap at a resolution of 50 000.

Tryptic digestion mixture of bovine serum albumin (BSA, Bio-Rad) was used to define the RT-iRT relationship, following previously published procedures (Figure 1).¹⁹ Precursor ions for light- and heavy-labeled forms of tryptic peptides of epitranscriptomic RWE proteins were monitored using a previously reported scheduled PRM method¹⁹ with a 10 min retention time window, where the maximum number of concurrent precursor ions was set at 45. The extracted-ion chromatograms (XICs) were obtained from manual analysis using Skyline (Version 21.2)²² or Xcalibur with ≥ 4 fragment ions from coeluting light- and heavy-labeled peptide pairs. The chromatographic peak areas were exported from Skyline and processed using a customized R (4.3.1) script in Rstudio (2023.06.1 + 524) with the following packages: tidyverse (2.0.0), ggplot2 (3.4.2), ggrepel (0.9.3), ggh4x (0.2.4), openxlsx (4.2.5.2), scales (1.2.1), and drc (3.0–1). The script was uploaded to a GitHub repository (https://github.com/odingsy/Epitranscriptomicsreader-writer-and-eraser-proteins-in-melanoma-drug-resistance/blob/main/proteomics.R). The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE²³ partner repository with the data set identifier PXD063781.

shRNA Stable Knockdown (KD)

Plasmids for shRNA-mediated stable knockdown of TRMU were designed by inserting stem-loop sequences into the AgeI and EcoRI sites of the PLKO.1-puro plasmid.²⁴ The sequences of the plasmids were confirmed by Sanger sequencing.

Plasmids for lentivirus-based stable KD experiments were constructed by following previously published protocols.²⁴ Briefly, 3 μg of pCMV-dR8.2 dvpr (Addgene plasmid #8455), 1 μg of pLTR-G (Addgene plasmid #17532) and 4 μg of stable KD plasmids were cotransfected into a 10 cm dish of HEK293T cells using 50 μL of PolyFect Transfection Reagent (Invitrogen); lentiviral experiments with 4 μg control shRNA (shCtrl) were performed in parallel. The transfection medium was replaced with fresh DMEM 24 h later. After another 48 h, the virus-containing media were collected and passed through a 0.45 μm PES membrane and introduced separately to IGR37 and IGR37xp cells. After a 48-h incubation, the transfected cells were selected with 500 ng/mL puromycin for 1 week. The cells were then harvested for assessing knockdown efficiency as well as for survival and Seahorse assays.

CRISPR-Mediated Knockout of MTO1

The CRISPR plasmid was designed by CRISPick.^25,26 The single guided oligonucleotides were inserted into the PX330 plasmid (Addgene plasmid 42230) according to the original protocol.²⁷ Sequences at the inserted region were validated by Sanger sequencing.

IGR37xp cells were seeded in 24-well plates at 60% confluency. One μg of PX330 plasmid was added together with 3 μL TransIT-X2 Transfection Reagent (MIR6000, Mirus) for 24 h. On the next day, DMEM containing 500 ng/mL puromycin was incubated with the transfected cells for 3 days. The cells were subsequently seeded at a density of approximately 3 cells per well in 100-μL medium. After single cells were repopulated into colonies (a week), they were transferred to 24 wells and MTO depletion was verified by Western blot analysis.

Cell Survival Assay

IGR37 and IGR37xp cells with different genetic backgrounds and the respective controls were seeded in a 96-well plate at a density of 10⁵ cells per well. After a 24-h incubation, the culture media were changed to the media containing the indicated concentrations of vemurafenib. After a 48-h incubation in vemurafenib-containing media, the media were changed to 10% CCK-8-containing DMEM (ck04-13, Dojindo). Absorbance at 450 nm was recorded using a Synergy H1 (BioTek) plate reader. The data were plotted, and LC₅₀ was derived by a customized R script (https://github.com/odingsy/Epitranscriptomicsreader-writer-and-eraser-proteins-in-melanoma-drug-resistance/blob/main/phenotypicAssays.R). Statistical significance of relative survival differences between shTRMU and shCtrl groups at each dose was determined using two-way ANOVA, with Bonferroni correction for multiple comparisons.

Western Blot

Cells were washed and pelleted in 1× PBS and lysed with CelLytic M cell lysis reagent (Sigma) supplemented with 1% protease inhibitor cocktail. Cell lysate was boiled at 95 °C for 5 min in Laemmli loading buffer (Bio-Rad). Fifteen μg of denatured lysate was resolved on a 10% SDS-PAGE gel at 90 V for 30 min, followed by 130 V for 90 min. The separated proteins were transferred onto a nitrocellulose membrane at 100 V for 60 min at 4 °C. The membrane was blocked with 5% milk in PBS-T (PBS with 0.1% Tween 20) for 60 min, and then incubated with primary antibodies recognizing METTL3 (67733-1-AP, Proteintech, 1:2000), METTL14 (26158-1-AP, Proteintech, 1:2000), PUS7 (PA5-54983, Invitrogen, 1:2000), MTO1 (15650-1-AP, Proteintech, 1:1000) and GAPDH (Santa Cruz, sc-32233, 1:5000) at 4 °C overnight with shaking. After washing with PBS-T for three times, the membrane was incubated with antirabbit secondary antibody (Sigma, A0545, 1:5,000), or antimouse secondary antibody (Santa Cruz, m-IgGκ BP-HRP, 1:5,000), washed for three times with PBS-T, and the protein bands were visualized using Amersham ECL Western Blot Detecting Reagent (GE Healthcare) and imaged with a LI-COR imaging system.

Seahorse Assay for Monitoring Mitochondrial Activity

Mitochondrial ATP production rate was measured using Standard XF Real-Time ATP Rate Assay (Agilent) on a Seahorse XF Pro Analyzer (Agilent) by following the manufacturer’s recommended procedures. Briefly, IGR37, IGR37xp cells, and IGR37xp cells stably transfected with shTRMU or shCtrl were seeded for 48 h prior to measurement in a Seahorse plate at a density of 2 × 10⁴ cells per well. The chamber was hydrated overnight in a non-CO₂ incubator at 37 °C. On the same day of measurement, the cells were washed with XF Real-Time ATP Rate Assay Medium twice and placed in the non-CO₂ incubator at 37 °C for 1 h prior to measurement. Oligomycin (10×, 15 μM) and rotenone + antimycin A (10×, 5 μM) were transferred into the respective injection ports according to the manufacturer’s recommended procedures. Cartridge calibration was performed before the actual runs initiated. After the run, the whole proteome was harvested and quantified using Quick Start Bradford Protein Assay (Bio-Rad) to ensure uniform seeding density across different biological replicates. The data were processed using web-based Seahorse Analytics tools (Agilent) and plotted using GraphPad Prism (Version 10.2.2).

RESULTS AND DISCUSSION

We set out to ask whether changes in expression of epitranscriptomic RWE proteins contribute to vemurafenib resistance in metastatic melanoma cells. To this end, we employed scheduled LC-PRM, in conjunction with SILAC-based metabolic labeling, to assess the differential expression of these proteins using an isogenic pair of metastatic melanoma cells that carry BRAF^V600E mutation and are sensitive (IGR37) or resistant (IGR37xp) to vemurafenib.^8,20 The LC-PRM results allowed for the quantifications of 110 out of 154 epitranscriptomic RWE proteins (71.4%) in the library (Table S1). Among all the quantified proteins, 4 and 11 were consistently up- and down-regulated, respectively, by at least 2-fold in IGR37xp relative to IGR37 cells (Figure 2a,c). We also validated the PRM quantification results of selected proteins by Western blot (Figure 3), suggesting the accuracy and robustness of the LC-PRM method in quantifying the epitranscriptomic RWE proteins.

Figure 2.

LC-PRM-based targeted proteomic analysis revealed the reprograming of the epitranscriptomic RWE proteome associated with vemurafenib resistance in melanoma cells. (a) A scatter plot displaying log₂-transformed RWE protein abundance ratios in IGR37xp/IGR37 cells from the mean ratios of two forward and two reverse SILAC labeling experiments. 110 out of 154 proteins (71.4%) in the library were quantified. Red dots indicate those proteins with SILAC protein ratios being >2 or <0.5. All the upregulated proteins are annotated with the protein names whereas the only the top 5 downregulated proteins are labeled to avoid overcrowding. The results for all quantified proteins are shown in Table S1. (b) Extracted-ion chromatograms (XICs) for monitoring the formation of the y₃, y₄, y₇, and y₈ ions from the [M+2H]²⁺ ion of a TRMU peptide, TPNPDIVCNK. Peptide quantification was achieved by calculating the ratio of heavy/light in the forward labeling experiments and the ratio of light/heavy in the reverse labeling experiments, respectively. These peptide ratios were used to represent ratios in protein expression between vemurafenib-resistant and sensitive cells. (c) Mean and standard deviation of top 10 ratios (log₂-transformed) of up- and down-regulated epitranscriptomic RWE proteins in IGR37xp and IGR37 cells (n = 4).

Figure 3.

Western blot validation of quantification results of representative proteins. (a) Western blot images and (b) quantification results for confirming the differential expression of PUS7, METTL3, METTL14 and MTO1 proteins in IGR37xp cells maintained in 5 μM vemurafenib over IGR37 cells. (c) XICs and (d) quantification results based on the formation of the y₆, y₇, y₈, y₉, and y₁₀ ions from the [M+2H]²⁺ ions of a PUS7 peptide, FGTTAVP-TYQVGR, of the y₄, y₅, y₆, y₇, y₈, and y₉ ions from an METTL3 peptide, NPEAALSPTFR, of the y₃, y₄, y₅, and y₇ ions from an ME an METTL14 peptide, ETGITANEK and of y₂, y₃, y₄, and y₅ ions from an MTO1 peptide, LGFVVGR.

Among all the differentially expressed proteins associated with vemurafenib resistance, tRNA mitochondrial 2-thiouridylase (TRMU) and mitochondrial tRNA translation optimization 1 (MTO1) are of particular interest (Figure 2). In this vein, we found that TRMU and MTO1 proteins were 3.0- and 2.6-fold more abundant in IGR37xp than IGR37 cells, respectively (Figures 2 and 3). We also confirmed the differential expression of MTO1 in these two lines of cells using Western blot (Figure 3), though we were unable to validate the quantification result for TRMU by Western analysis owing to the lack of a good quality antibody.

MTO1 and TRMU act sequentially to produce 5-taurinomethyl-2-thiouridine (τm⁵s²U) at the 34th position of tRNA^Glu, tRNA^Gln and tRNA^Lys in human mitochondria. In particular, MTO1, together with mitochondrial GTP-binding protein 3 (GTPBP3), deposits taurine at the C5 position of uridine 34 to form 5-taurinomethyluridine (τm⁵U),²⁸ whereas TRMU works as a 2-thiouridylase to replace the oxygen atom at the C2 position of τm⁵U with sulfur.²⁹ τm⁵s²U at the 34th position of tRNA^Glu, tRNA^Gln and tRNA^Lys pairs with wobble base A/G more specifically than an unmodified uridine.³⁰ Mutations in TRMU could lead to Leigh syndrome,³¹ acute infantile liver injury even liver failure,^32,33 and hearing loss.³⁴ So far there has been no study linking TRMU with therapeutic resistance.

Considering the important role of τm⁵s²U modification in mt tRNAs in stimulating mitochondrial translation,³⁵ we next assessed whether elevated expression of TRMU in IGR37xp cells confers increased mitochondrial activity. Our Seahorse assay results showed that IGR37xp cells indeed exhibit elevated oxidative phosphorylation, as reflected by augmented oxygen consumption rate (OCR), compared to parental IGR37 cells (Figure 4a). In this vein, it is worth noting that we attempted, but failed to detect τm⁵s²U in the digestion mixture of total tRNA samples isolated from IGR37 or IGR37xp cells by LC-MS/MS, which is likely attributed to relatively low level of this modified nucleoside in total tRNA.

Figure 4.

(a) Standard XF Real-Time ATP Rate Assays measuring real-time changes in oxygen consumption rates (OCR) of IGR37xp maintained in 5 μM vemurafenib and of IGR37 cells and (b) in IGR37xp cells treated with shTRMU and shCtrl. (c) A CCK-8 assay measuring survival of IGR37xp cells treated with two different sequences of shRNA targeting TRMU (shTRMU1 and shTRMU2, n = 3) or control shRNA (shCtrl), and exposed with different concentration of vemurafenib. Statistical significance of relative survival differences between the two shTRMU and shCtrl groups at each dose was determined using two-way ANOVA, with Dunnett correction for multiple comparisons. Error bars represent 95% confidence interval of the fitted model. (d) LC₅₀ was estimated from the fitted dose response curves from shTRMU1 (n = 3) and shTRMU2 (n = 3). One-way ANOVA with Dunnett correction for multiple comparison was applied. Error bars represent standard derivation of the mean.

We next explored whether elevated expression of TRMU contributes to augmented resistance of IGR37xp cells to vemurafenib. Our results showed that shRNA-mediated depletion of TRMU in IGR37xp cells led to significantly augmented sensitivity to vemurafenib, as reflected by a change in LC₅₀ from 32.1 in the control shRNA-treated cells to 15.7 and 14.2 μM for shTRMUl and shTRMU2, respectively (Table S2, Figure S1, and Figure 4c,d). We also found that genetic depletion of TRMU leads to diminished OCR in IGR37xp cells (Figure 4b). Moreover, we observed that genetic ablation of MTO1 in IGR37xp cells led to increased sensitivity to vemurafenib (Figure S2).

To study neomycin-related ototoxicity, He et al.³⁶ observed that TRMU knockdown could sensitize hearing cells to neomycin-elicited death, which could be rescued by N-acetylcysteine (NAC). We, however, found that cotreatment with 2 μM NAC did not alter appreciably the vemurafenib sensitivity of IGR37xp cells treated with control shRNA or shTRMU (Figure S3), indicating that vemurafenib resistance does not occur through augmented production of reaction oxygen species.

Drug resistance is the main obstacle in established therapeutic regimes for cancer.⁴ To identify molecular targets for restoring sensitivity to therapy, researchers employed various high-throughput approaches to discover candidate genes associated with the resistance phenotype.^6,7 Here we explored how alterations in the epitranscriptome contribute to therapeutic resistance to vemurafenib by employing a targeted quantitative proteomic approach to identify systemically those epitranscriptomic RWE proteins that are differentially expressed between IGR37 metastatic melanoma cells and the isogenic vemurafenib-resistant cells (IGR37xp).

We discovered that MTO1 and TRMU, which act sequentially to catalyze the formation of τm⁵s²U,^28,29 are expressed at pronouncedly higher levels in the vemurafenib-resistant over parental IGR37 metastatic melanoma cells (Figures 2 and 3). In addition, genetic depletion of MTO1 and TRMU resensitized melanoma cells to vemurafenib toxicity (Figure 4 and Figure S2). We also found that IGR37xp cells exhibit augmented oxidative phosphorylation than IGR37 cells, and genetic depletion of TRMU in IGR37xp cells led to markedly attenuated oxidative phosphorylation. These findings highlight the importance of mitochondrial tRNA τm⁵s²U modification and the ensuing regulation of oxidative phosphorylation in vemurafenib resistance.

Selective BRAF^V600E inhibition, which is the mechanism of action of vemurafenib, elicited increased dependence of melanoma cells on oxidative phosphorylation,³⁷ and augmented mitochondrial fusion and metabolism.^38,39 Combination therapy, relying on vemurafenib together with depolarizing mitochondrion inner membrane³⁷ or inhibiting oxidative phosphorylation proteins,⁴⁰ was proposed for restoring sensitivity of melanoma cells to vemurafenib. In addition, embryonic stem cells (ESCs) from Mto1^−/− mice exhibit impaired mitochondrial translation and oxidative phosphorylation, as well as defective mitochondrial protein import, leading to cytosolic protein aggregation and unfolded protein response (UPR).⁴¹ As a tRNA-modifying enzyme acting on U34 to form τm⁵s²U, which is critical for wobble base pairing with mRNA, TRMU’s function in modulating vemurafenib sensitivity in melanoma cells likely occurs through modulating mitochondrial function, which is indeed supported by our Seahorse assay results (Figure 4a,b).

CONCLUSIONS

In conclusion, we characterized comprehensively the reprogramming of epitranscriptomic RWE proteome that are associated with vemurafenib resistance in melanoma. Our results revealed that several RNA-modifying enzymes exhibit altered expression in vemurafenib-resistant IGR37xp melanoma cells over parental IGR37 cells. We also demonstrated that the elevated expression of a mitochondrial tRNA-modifying enzyme, TRMU, confers vemurafenib resistance through stimulating oxidative phosphorylation in IGR37xp cells. Our findings suggest a role of TRMU in melanoma resistance to BRAF inhibition. In this vein, it is worth noting that mechanisms of drug resistance can vary among different cell lines;⁴² therefore, it will be important to examine whether the findings made in the current study can be extended to other metastatic melanoma cell lines in the future.

Supplementary Material

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jproteome.5c00805.

Experimental procedures, sequences of oligonucleotides to induce knockdown and the list of proteins quantified by the SILAC-based PRM on RWE proteins. Table S1. The quantification results for the epitranscriptomic RWE proteins in IGR37xp and IGR37 cells. Table S2. Oligonucleotides used in this study. Figure S1. LC-MS/MS quantification results for a tryptic peptide of TRMU protein in IGR37xp cells after treatment with different shRNAs. Figure S2. Genetic depletion of MTO1 led to increased sensitivity of IGR37xp cells to vemurafenib. Figure S3. Survival plots for IGR37xp cells treated with increasing concentrations of vemurafenib and in the presence or absence of N-acetylcysteine. Figure S4. Uncropped Western blot images. (PDF)

Data Availability Statement

The RWE PRM data has been deposited into ProteomeXchange Consortium via the PRIDE partner repository with the data set identifier PXD063781.