USP39 promotes hepatocellular carcinogenesis through regulating alternative splicing in cooperation with SRSF6/HNRNPC

Abnormal alternative splicing (AS) caused by alterations in spliceosomal factors is implicated in cancers. Standard models posit that splice site selection is mainly determined by early spliceosomal U1 and U2 snRNPs. Whether and how other mid/late-acting spliceosome components such as USP39 modulate tumorigenic splice site choice remains largely elusive. We observed that hepatocyte-specific overexpression of USP39 promoted hepatocarcinogenesis and potently regulated splice site selection in transgenic mice. In human liver cancer cells, USP39 promoted tumor proliferation in a spliceosome-dependent manner. USP39 depletion deregulated hundreds of AS events, including the oncogenic splice-switching of KANK2. Mechanistically, we developed a novel RBP-motif enrichment analysis and found that USP39 modulated exon inclusion/exclusion by interacting with SRSF6/HNRNPC in both humans and mice. Our data represented a paradigm for the control of splice site selection by mid/late-acting spliceosome proteins and their interacting RBPs. USP39 and possibly other mid/late-acting spliceosome proteins may represent potential prognostic biomarkers and targets for cancer therapy.

Antibody binding was detected by incubation with biotinylated anti-rabbit IgG antibody and visualized by reaction with DAB Substrate (Boster, California, USA).

Cell lines
The HCC cell line PLC-8024 (TCHu119) was obtained from the Institute of Virology of the Chinese Academy of Sciences (Beijing, China).The HCC cell line SNU-449 (CRL-2234) was obtained from the American Type Culture Collection (ATCC).Cell line 293T (SCSP-502) was obtained from the Cell Bank affiliated to Shanghai Institute of Biochemistry and Cell Biology.The mouse HCC cell line Hepa1-6 (CTCC-ZHYC-0566) was obtained from Meisen Chinese Tissue Culture Collections (MeisenCTCC).The mouse embryo fibroblast cell line 3T3 was a kind gift from Dr. Xijun Ou.All cell lines were authenticated using short tandem repeat profiling and routinely tested for mycoplasma contamination.The cells were cultured in Dulbecco's modified Eagle's medium (DMEM) (Gibco, California, USA), supplemented with 10% fetal bovine serum (Gibco), and 1% penicillin/streptomycin mixture (Gibco).All cell lines used in this study were incubated at 37°C in a humidified incubator containing 5% CO2.

siRNA transfection
siRNA sequences were designed using the siDRECT website (http://sidirect2.rnai.jp/)and the sequences are listed in Table S5.siRNA transfection was performed using siRNA-mate reagent (GenePharma, Shanghai, China) according to the manufactures' instruction.

Tet-On system mediated overexpression and knockdown
Tet-On 3G systems are inducible gene expression systems with two elements, namely, the Tet-On 3G transactivator protein and a gene of interest under the control of a TRE3G promoter.The PiggyBac transposon system has been shown to be highly efficient in mediating gene transfer.(3)Therefore, it has been modified to deliver the multiplex Tet-On 3G system.Briefly, the KANK2 gene sequence and two shRNAs targeting the KANK2-L isoform (shKANK2-L) and KANK2-S isoform (shKANK2-S) were cloned into the pBX-093 plasmid (PB5-HS4-SV40-puro-2A-tetON3G-pA-HS4-TRE-AzaminGreen-2A-Tet3G-RNAi-GpA-HS4-PB3, a kind gift from Dr. Wei Huang, Southern University of Science and Technology, Shenzhen, China) respectively.Cells were subsequently cotransfected with the constructed pBX-093 plasmid and pBX-090 plasmid (pN1-CMV-PGK-piggybac, a kind gift from Dr. Wei Huang) and sorted based on Azamin Green.Before the functional assays, cells were treated with 1μg/ml doxycycline (DOX) to induce the overexpression of KANK2 or knockdown of KANK2-S and KANK2-L.Primers and KANK2-shRNAs sequences are listed in Table S5.

RNA extraction, PCR and qRT-PCT analysis
RNA extraction was performed using TRIzol reagent (Vazyme Biotech, Nanjing, China) according to the manufacturer's instructions.1μg of total RNA was used to synthesize the first strand of cDNA using Hifair® Ⅲ 1st Strand cDNA Synthesis SuperMix (Yeasen, Shanghai, China, #11141ES60,).
For PCR analysis, Green Taq Mix (#P131-03, Vazyme) was used according to the manufacturer's instructions.PCR products were analyzed by 6 / 30 electrophoresis on a 1% agarose gel.
For qRT-PCR analysis, the SYBR® Green Premix Pro Taq HS qPCR Kit (Accurate Biology, Changsha, China, #AG11701) was used according to the manufacturer's instructions.The relative changes in gene expression were calculated using the 2 -ΔΔCt method.The primer sequences are listed in Table S5.

Cell viability assay, foci formation assay and soft agar colony formation assay
CCK8 assay was performed to analyze cell viability.Cells were digested and re-cultured in 96-well plates at 2000 cells per well in 100μl of medium.Each well was added 10μl CCK8 solution (MCE, New Jersey, USA, HY-K0301) and incubated for 3 h at 37 °C.Optical density was measured using a microplate reader at a wavelength of 452 nm within 5 days.Triplicate repeats were performed to determine the variance and significance.The values were plotted by averaging triplicate wells.
For the foci formation assay, cells were digested and re-cultured in 6-well plates at 500 cells per well in 2 ml medium.After incubation for 9 days, natural monolayer colonies were formed.The cells were subsequently washed with PBS, fixed with 4% paraformaldehyde, stained with purple crystals for 10 min, and washed with PBS three times.The results were photographed, and the number of clones was counted.
For the soft agar colony formation assay, 2000 PLC-8024 cells were mixed 7 / 30 with 10% FBS, 1× DMEM, and 0.35% agar as the upper layer, whereas the bottom layer contained 10% FBS, 1× DMEM and 0.6% agarose.The colonies were photographed under a microscope and the clone numbers were counted 20 days later.

Cell cycle flow cytometry
Cells were grown to a density of 70% in 6-well dishes and collected for staining with propidium iodide (PI, Sangon, Shanghai, China, #A601112-0100) according to the manufacturer's instructions.A CytoFLEX cytometer (Beckman Coulter, California, USA) was used to measure changes in cell cycle, and Modfit software was used to analyze the results.
A double thymidine block was used to synchronize cells at the G1/S boundary.Briefly, the cells were treated with 2mM thymidine (Thd, Sigma, #T1895-1G) for 18h and then washed to remove thymidine.After a 9-hour incubation in fresh medium, the cells were treated with the second round of Thd (2 mM) and synchronized at the G1/S boundary.Cells were released by washing with pre-warmed 1x PBS and incubating the cells in pre-warmed fresh media.Cells were collected at 0, 3h for analysis of cell cycle by DNA staining using PI.
To synchronize the cells at the G2/M boundary, the cells were treated with 2mM Thd for 24h and then washed to remove thymidine.After 3-hour incubation in fresh medium, cells were successively treated with 100ng/ml nocodazole (MCE, #HY-13520) and then synchronized at the G2/M boundary.Cells were 8 / 30 released by washing with pre-warmed 1x PBS and incubating the cells in prewarmed fresh media.Cells were collected at 0, 4h for analysis of cell cycle by DNA staining using PI.
Biotinylated RNA pulldown and RNA pulldown coupled with mass spectrometry Biotinylated RNA-pulldown assays were performed as previously described( 4).
Briefly, exon with 60bp intron sequences at the 5'/3' splice site were cloned into pcDNA3.1,and the plasmid was linearized with the following primers: Forward: TGCTCTGATGCCGCATAGTT, Reverse: GCCCACTACGTGAACCATCA.The linearized DNA fragment was used as a template to generate biotin-labeled RNA for precipitation from PLC-8024 cell lysate.Subsequently, WB assay was performed to detect precipitation.
RNA pulldown coupled with mass spectrometry was performed as previously described.(5 These probes were incubated with PLC-8024 cell lysates, followed by an 9 / 30 additional incubation with streptavidin beads.After on-bead digestion and desalination, the extracted peptides were analyzed by LC-MSMS.
A 10%-volume of lysate was used as the input.RNA in the precipitates and input was subsequently extracted using the acid phenol-chloroform method.

RNA seq samples preparation and analysis
USP39 knockdown and control PLC-8024 cells (three biological replicates per sample) and the tumor or para-tumor tissues collected from Usp39 HOE and wild-type (WT) mice were suspended in TRIzol reagent (Invitrogen) and sent to Novogene for RNA sequencing.RNA-seq libraries prepared using oligo (dT) beads and rRNA removal methods were pooled and sequenced using an Illumina platform.Paired-end reads were mapped to the Homo sapiens GRCh38(hg38) transcriptome and the Mus musculus reference genome GRCm39(mm39) using the STAR RNA-seq aligner.Alignment files were filtered and sorted using SAMTools.The number of reads mapped to each gene was quantified based on the processed alignment files using htseq-count, and differential expression analysis was performed using DESeq2.Genes with a Padj value < 0.05 and |log2(fold change)|>1 were determined to be differentially expressed.GO and KEGG enrichment analyses were performed using the online tool MSigDB (https://www.gsea-msigdb.org/gsea/msigdb/).
Analysis of pre-RNA splicing efficiency was performed based on the processed alignment files.Briefly, paired-end reads were aligned to the human reference genome GRCh38 using the STAR aligner, and then filtered and sorted using SAMTools.For each splice site, the number of reads covering the first base at the 5' end of the intron (i.e., 5' intron coverage), number of reads covering the last base at the 3' end of the intron (i.e., 3' intron converage), number of reads covering the last base at the 3' end of the upstream exon (i.e., 3' exon coverage), and number of reads covering the first base at the 5' end of the downstream exon (i.e., 5' exon coverage) were quantified using bedtools.

Statistical analysis
Statistical analyses were performed using SPSS 24.0 or GraphPad Prism 8.0.
The mRNA levels of USP39 and KANK2-L isoforms in paired non-tumor and tumor samples were compared using a paired two-tailed Student's t-test.
USP39 expression levels in unpaired clinical samples were compared using unpaired two-tailed Student's t-test.USP39 expression levels in HCC samples with different tumor stages and neoplasm histologic grades were compared using the Kruskal-Wallis test.Differences in overall survival and disease-free survival were calculated using Kaplan-Meier plots and log-rank tests.
Correlations between two statistical variables were analyzed using Pearson's correlation analysis.An unpaired two-tailed Student's t-test was used to compare observations, such as the number of foci and the relative expression ) Briefly, biotinylated DNA probes were designed to target the exon1-2 junction or the exon1-3 junction of KANK2 and their