Bioinformatics analysis

The TCGA database (https://www.cancer.gov/ccg/research/genome-sequencing/tcga) was used to analyze the differences in NSUN2 and HGH1 expression between normal and tumor tissues of the breast. Survival plots related to NSUN2 and HGH1 expression in BC patients were generated with Gepia2 (http://gepia2.cancer-pku.cn/). The STRING database (https://cn.string-db.org/) was used to analyze proteins that interact with HGH1.

Clinical samples

Thirty-nine BC tissues and nine adjacent tissues were collected to detect the NSUN2 protein level by IHC. These samples were preserved by pathologists via paraffin embedding. Five pairs of BC and adjacent tissue samples were collected for enrichment of mRNA and library construction, which were sufficient for RNA-Seq and RNA-BisSeq. These samples were stored in liquid nitrogen until use. All the clinical tissue samples were collected from 2019 to 2020 at the First Affiliated Hospital of Zhengzhou University. This study was approved by the Institutional Review Board of the First Affiliated Hospital of Zhengzhou University (the approval number 2019-KY-31) and complied with the Declaration of Helsinki.

Immunohistochemistry

First, the paraffin samples were heated at 60 °C for 1 h. Then, an automatic immunohistochemistry instrument (ROCHE) and an automatic staining instrument (LEICA) were used to incubate the sections with the antibodies. Images were scanned using the 3DHISTECH PANNORAMIC VIEWER. Concentrations of anti-NSUN2 (Sigma, HPA037896), anti-HGH1 (Proteintech, ab181031), and anti-Ki67 (Servicebio, GB111499) were used according to the supplier’s instructions.

LC-MS/TOF

RIPA lysis buffer was used to extract total protein from the breast tissues, and a 10 K ultrafiltration tube was used for FASP digestion. Mass spectrometry analysis was carried out using Eksigent NanoLC 400 liquid chromatography and AB Sciex6600*TripleTOF mass spectrometry. The human protein database was downloaded from the UniProt website, SWATH mass spectrometry data were imported into DIA-NN 1.8.1 software for database comparison analysis, and the corresponding protein data of the peptides were obtained for further analysis. Based on three biological replicates of each sample, the differentially expressed proteins with a fold change > 1.5 or < 0.67 and a p-value < 0.05 were screened for further research.

Cell lines

The human normal breast epithelial cell line MCF10A and the human breast cancer cell lines MDA-MB-468, MDA-MB-231, and MCF7 were purchased from the National Collection of Authenticated Cell Cultures of the Chinese Academy of Sciences. The human breast cancer cell line T47D was kindly provided by Dr. Yue Du’s laboratory in the Department of Pharmacy, First Affiliated Hospital of Zhengzhou University.

Animals

According to the guidelines of the Hospital Ethics Committee, this project purchased 6-week-old female BALB/c nude mice from Jiangsu JICUI Yiming Biotechnology Co., Ltd. The mice were raised in an SPF-grade environment at the Experimental Animal Center in Henan Province, with an environmental temperature of 21–23 °C and a humidity of 40-60%.

CDX mode

Six-week-old female BALB/c nude mice were used for xenotransplantation experiments. The experimental nude mice were anesthetized via isoflurane inhalation (1%) according to the animal ethics requirements. After complete anesthesia, the constructed cell lines were injected into the fat pad of the second mammary gland on the right side of the nude mice. After the injection, the wound was sutured, and the nude mice were observed until they woke up and regained their ability to move. The mice were observed every 3 days, and the nude mice were sacrificed based on the size of the tumor. The tumor size and weight were recorded, and the tumors were fixed in a 4% paraformaldehyde tissue cell fixation solution. After paraffin embedding and sectioning, immunohistochemical detection was performed.

Plasmids and transfection

The PLKO.1-shNSUN2 plasmid was kindly provided by Prof. Yun-Gui Yang (Beijing Institute of Genomics, Beijing, China). The 3FLAG-EGFP-NSUN2-WT, 3FLAG-EGFP-NSUN2-DM (C271/321A), 3FLAG-EGFP-YBX1-WT, 3FLAG-EGFP-YBX1-Mut (W65A), LV-HGH1-RNAi, and LV-HGH1 lentivirus were purchased from GENECHEM (Shanghai, China). Transient transduction was carried out using Lipofectamine 3000 (Invitrogen) according to the manufacturer’s instructions. The cells were then cultured for 48 h and harvested for western blot analysis. For lentiviral transduction, a second-generation lentivirus packaging system consisting of psPAX2 (Addgene) and PMD2. G (Addgene) was used to generate virus particles. In brief, plasmids were transfected into HEK293T packaging cells at 60% confluence using Lipofectamine 3000 (Invitrogen) according to the manufacturer’s instructions. After an additional 48 h of incubation, the supernatant was collected, filtered through a 0.45-µm filter (Millipore), and used to infect host cells in the presence of 6 µg/mL HitransG A/P (GENECHEM). The resulting stable polyclonal populations of transduced cells were then selected with puromycin or hygromycin (Solarbio, China) for two weeks, followed by validation by western blotting.

The shRNA sequences used are as follows:

shNSUN2#1: 5’- GCTGGCACAGGAGGGAATATA-3’;

RNA interference

Cells were seeded at a confluence of 40% in 6-well or 96-well plates. After 24 h, according to the manufacturer’s instructions, the cells were transfected with the indicated siRNA oligonucleotides using Lipofectamine RNA iMAX (Invitrogen). Then, the cells were cultured for 72 h and harvested for cell viability assays or Western blot analysis. The target sequences of the siRNA oligonucleotides used are as follows:

siNSUN2#1: 5’- GAGAUCCUCUUCUAUGAUCTT-3’;

siNSUN2#2: 5’- CACGUGUUCACUAAACCCUAUTT-3’;

siYBX1#1: 5’- GGAUAUGGUUUCAUCAACATT-3’;

siHGH1: 5’- AUGCUUGUUGAAGCCAUCATT-3’;

siEEF2: 5’-GCCCUCUUAUGAUGUAUAUTTAUAUACAUCAUAAGAGGGCTT-3’.

RNA extraction and real-time RT-PCR (RT-qPCR)

Total RNA was extracted from cells and tissues using TRIzol reagent (Invitrogen). Then, the RNA was reverse-transcribed into complementary DNA (cDNA) using a PrimeScript RT reagent kit (Takara, RR037A). RT-qPCR analysis was conducted using iTaq Univer SYBR Green Supermix (Bio-Rad) and QuantStudio5 Applied Biosystems (Thermo Fisher Scientific) according to the manufacturer’s instructions. The results were analyzed as the fold change in expression. The primer pairs used are listed below:

NSUN2 forward: 5’- CAAGCTGTTCGAGCACTACTAC-3’,

NSUN2 reverse: 5’- CTCCCTGAGAGCGTCCATGA-3’;

YBX1 forward: 5’- GCGGGGACAAGAAGGTCATC-3’;

YBX1 reverse: 5’- CGAAGGTACTTCCTGGGGTTA-3’;

HGH1 forward: 5’- ACATCCGCAAGATGCTTGTTG-3’;

HGH1 reverse: 5’- GAAGGATCAGGTAGGCTCCCT-3’;

GAPDH forward: 5’- GGAGCGAGATCCCTCCAAAAT-3’;

GAPDH reverse: 5’- GGCTGTTGTCATACTTCTCATGG-3’;

18S rRNA forward: 5’- GTAACCCGTTGAACCCCATT-3’;

18S rRNA reverse: 5’- CCATCCAATCGGTAGTAGCG-3’.

RNA m5C dotblot

Total RNA was extracted from cells using the TRIzol method, and mRNA was enriched using the magnetic bead method. After quantification with Qubits, 200 ng of mRNA from each group of cells was divided 3–5 times and spotted onto a Hybond-N + nylon membrane (Solarbio, Cat# YA1760). Each spot was crosslinked with UV for 5 minutes in a dark crosslinking instrument. After all the samples were crosslinked, the membrane was blocked with 5% nonfat milk and then incubated with an anti-m5C antibody (1:1000, Abcam, ab214727) overnight at 4 °C. The next day, the membrane was incubated with a secondary antibody at room temperature, and the membrane was exposed to the imaging system using a sensitive ECL exposure solution (Fisher Biotech). After exposure, the membrane was stained with 0.02% methylene blue (Sangon Biotech) to determine whether the amount of sample loaded was consistent for each group.

RNA-seq and RNA-Bisseq

Total RNA was extracted using the TRIzol method. The mRNA was enriched using the VAHTS mRNA Capture Beads Kit (Vazyme, 401-01/02) according to the manufacturer’s instructions. DNase I (NEB, M0303) and RNaseOut (Invitrogen, 10,777,019) were used to remove DNA and RNase from the enriched mRNA. RNA fragmentation reagents (Ambion, AM8740) were used to fragment the mRNA. Magnetic RNA Clean Beads (Vazyme, N412) were used to purify the mRNA samples, and Qubits was used to determine the mRNA concentration.

A KAPA stranded mRNA-seq kit (KK8429) was used to construct mRNA sequencing libraries according to the manufacturer’s instructions. The main steps included cDNA synthesis of the first and second strands, purification of the second strand synthesis products, A-tailing addition, adapter addition, purification of the ligation products, cDNA library amplification, and purification of the PCR products of the library.

mRNA C-T base conversion was performed according to the instructions of the conversion reagent (Zymo, R5001), unmethylated C was converted to U, and C was converted to T after PCR, while methylated C remained unchanged; 1 µL of diluted Luciferase Control RNA (Promega, L456A) was added as an internal reference for C-T conversion to evaluate conversion efficiency. The obtained mRNA was mixed with random primers (ACT hexamers), and cDNA I strand synthesis was performed using 100 mM DTT, 2.5 mM dNTPs, and SuperScript II. For chain II synthesis and subsequent steps, refer to the transcriptome process.

The libraries were transported on dry ice and sent to the Nanjing Jiangbei New Area Biomedical Public Service Platform for Illumina NovaSeq 6000 platform-based transcriptome sequencing.

Sequencing data analysis

The adaptors and low-quality bases were trimmed using Cutadapt and Trimmomatic software, respectively. Clean reads were mapped to the human genome (hg19) using meRanTK40’s meRanGh. The Dhfr spike-in was used to calculate the C to T conversion rate of > 99%. For each sample’s BAM file, only high-quality bases (Q ≥ 30) were used for variant calling. The mismatched types at each site were inferred based on the gene annotations on the chain. Additionally, each variant had to meet the following criteria: base quality score ≥ 30, mismatch frequency ≥ 0.1, and C + T coverage ≥ 20. Furthermore, we required that (i) the variant still met the above criteria, (ii) the variant’s signal ratio was ≥ 0.9, and (iii) the p-value was calculated using a one-sided binomial test. To identify high-confidence sites, we required that each tissue type site pass the p-value < 0.001 according to Stouffer’s Z score method. The annotation of m5C sites was performed using BEDTools intersectBed. The m5C sites that satisfied the conditions in each sample were merged, and consecutive C sites were removed. The mean methylation level of each tumor methylation site was calculated.

Combining the transcriptome sequencing data, the genes were divided into four groups: high methylation and high expression, high methylation and low expression, low methylation and high expression, and low methylation and low expression, based on the criteria of |diffmean (Tumor-Normal)| ≥ 0.05, |log2foldchange| ≥ 1, MeanT_P > 0.1, and Bs_seq_p-value < 0.05. Differentially significant genes with high methylation and high expression were selected as potential target genes for BC m5C dependency.

MeRIP-qPCR

Then, 5 µg of m5C antibody was mixed with 50 µl of Protein A Dynebeads in 200 µl of IPP Buffer and incubated at room temperature for 1 h. Total RNA was extracted from the samples using the TRIzol method, and mRNA was enriched using magnetic beads. DNA and RNA enzymes were removed from the enriched mRNAs using DNase I (NEB, M0303) and RNaseOut (Invitrogen, 10,777,019). The mRNA was fragmented using RNA fragmentation reagents (Ambion, AM8740). The mRNA samples were purified using RNA Clean Beads magnetic beads (Vazyme, N412). The concentration of mRNA was measured using Qubits, and 1/10 of the mRNA was used as the input control. The mRNA was mixed with magnetic bead antibodies and incubated at 4 °C for 4 h on a rotator. The supernatant was discarded, and the magnetic beads were washed three times with IPP buffer. Then, 200 µl of PK buffer and 40 µl of protease K were added, mixed well, and incubated at 55 °C with shaking at 1100 rpm for 60 min. After centrifugation, the supernatant was discarded, and the supernatant was transferred to a new 1.5 ml centrifuge tube. Then, 240 µl of RNA extraction reagent (ACMEC, AC13320) was added, mixed well, and incubated at 30 °C with shaking at 1100 rpm for 5 minutes. The upper layer was aspirated to a new 1.5 ml centrifuge tube. One microliter of glycogen, 20 µl of sodium acetate, and 660 µl of absolute ethanol were added. The samples were stored overnight at -80 °C. After centrifugation at 4 °C and 13,300 rpm for 30 min, the supernatant was carefully discarded, and the blue precipitate was RNA. The RNA was dissolved in an appropriate amount of nuclease-free water. For each input sample and MeRIP sample, the same amount of mRNA was reverse transcribed. The reverse transcription reaction mixture was prepared according to the instructions of the Novogene low-starting amount reverse transcription kit (HiScrip III All-in-one RT SuperMix Perfect for qPCR; R333), and RT-qPCR was subsequently performed.

RNA stability experiment (actinomycin D treatment method)

Cells were plated in 6-well plates and treated with actinomycin D (5 µg/mL, MedChemExpress) for various durations: 0, 2, 4, 6, and 8 h. Equal numbers of BC cells were collected at each time point, and total RNA was extracted using TRIzol reagent (Invitrogen). RNA from each sample was reverse-transcribed into cDNA, and RT-qPCR was used to measure the mRNA level of HGH1. The relative abundance of mRNA at each time point was calculated in reference to the t = 0 time point.

RIP-qPCR

The cell pellet was immersed in RIP lysis buffer (composed of 150 mM KCl, 10 mM HEPES (pH 7.6), 2 mM EDTA, 0.5% NP-40, 0.5 mM DTT, protease inhibitor, and RNase inhibitor) on ice for 30 min. One-tenth of the cell lysate was designated as the input control. The remaining cell lysate was then incubated either with rabbit IgG-coated beads or with anti-YBX1 (Abcam, ab76149)-coated beads for 4 h at room temperature. After this incubation, the bead-antibody-protein-RNA complex was washed five times with ice-cold washing buffer (consisting of 200 mM NaCl, 50 mM HEPES (pH 7.6), 2 mM EDTA, 0.05% NP-40, 0.5 mM DTT, and RNase inhibitor). Subsequently, the immunoprecipitated sample was digested with proteinase K, and the RNA was precipitated using glycogen (Thermo Scientific, AM9516). Total RNA was then extracted using TRIzol reagent, followed by RT-qPCR analysis.

Protein synthesis assay (Click-iT HPG Alexa Fluor method and anti-puromycin method)

Protein synthesis was detected using the Click-iT HPG Alexa Fluor 594 Protein Synthesis Kit (Thermo Fisher, C10429) according to manufacturer’s instructions. Briefly, the cells were treated with Click-iT HPG Alexa Fluor working solution and subjected to a single wash following formalin fixation. Then, cells underwent a 20-minute permeabilization period with 0.5% Triton X-100, and the cells were washed twice with 3% BSA in PBS. Next, a reaction cocktail was added to each well, followed by a 30-minute incubation at room temperature, shielded from light, and the cells were then washed with Click-iT reaction rinse buffer. After a 30-minute blocking period with 5% BSA, the cells were incubated with an anti-NSUN2 primary antibody followed by a FITC fluorescent secondary antibody. Finally, after washing with PBS, the cells were stained with Hoechst for 30 min. The samples were then examined using a fluorescence microscope equipped with appropriate filters for DAPI/Hoechst, FITC, and Alexa Fluor 594. Another method was applied in vitro using puromycin (P8230, Solarbio, China). Cells were treated with 1 µM puromycin for 1 h. Then, the cells were collected, and the proteins were extracted. Protein expression was determined through Western blotting.

Co-IP (co-Immunoprecipitation)

The protein of BC cells was extracted. NP-40 and Triton X-100 were used to prepare the lysis solution, and protein quantification was performed after 10 min of ice lysis. One hundred micrograms of protein from each group of samples was added to 500 µl of lysis solution. Part of it is used as input, while the other part is used as IP and IgG samples. Then, 20 µl of magnetic beads were added to each sample, 500 µl of TBS and 5 µg of the corresponding antibodies (HGH1 as the IP group and IgG as the negative control group) were added, and the mixture was incubated at room temperature for 1 h. After incubation, 100 µg of protein sample was added to the antibody magnetic bead suspension and incubated overnight with 4° rotation. The magnetic scaffold was used to discard the supernatant, and the protein lysis solution was used to lyse the magnetic bead-antibody protein mixture. We determined whether there was an interaction between the two proteins through Western blot.

Ribo-Seq

Two hundred microliters of ribosomal buffer (20 mM Tris-Cl, 150 mM NaCl, 5 mM MgCl2, 1 mM DTT) and 10% Triton X-100, 100 mM DTT, 50 mg/ml cyclohexide, 1 U) were added to 2 × 107 cells, and the resulting solution was combined with the µ-cracking solution prepared with L DNase. After gently blowing and beating, the ice was shaken for 10 min, and the mixture was regularly reversed. The samples were centrifuged at 4 ℃ and 20,000 × g for 10 min, after which the supernatant was transferred to a new centrifuge tube precooled on ice. Then, 120 µl of lysate was digested at room temperature with 3 µl of RNase I (Life Technologies, AM2294) for 45 min to remove RNA that was not protected by ribosomes. Finally, digestion was terminated with 4 µl of Superase In (Life Technologies, AM2694). RNA was extracted from the sample using TRIzol LS (Life Technologies, 10296-010), and residual ribosomal RNA was removed from the sample using a Ribo-Off rRNA Removal Kit (Vazyme, N406-01). Then, 16% urea PAGE was used to separate the sample RNA, the target RNA band position was 27–30 nt, and a Small RNATM PAGE Recovery Kit (Zymo R1070) was used for gel cutting and recovery. Finally, the Small RNA Amplification Library Prep Kit for Illumina (Vazyme, NR801) was used for library construction.