Tissue specimens and cell lines

A tissue microarray containing 167 melanoma tissues and 30 normal tissues were provided by Xi’an Alenabio Co., Ltd. (Xi’an, China). All patients signed informed consent forms. Studies involving human participants were reviewed and approved by Ethics Committee of Shanghai Tenth People’s Hospital.

Five human melanoma cell lines used here, A375, SK-MEL-28, MUM-2B, MUM-2C and OM431, were purchased from Cell Resource Center, Institute of Basic Medicine, Chinese Academy of Medical Sciences (Beijing, China). MUM-2B and MUM-2C cells were grown in H-DMEM medium + 10% FBS. A375 cells were cultured in DMEM medium + 10% FBS. SK-MEL-28 and OM431cells were cultured in 1640 medium + 10% FBS. All cells were incubated in a 37 °C incubator with 5% CO2.

Immunohistochemistry (IHC) staining

First, the tissue slides were put in the oven at 65 °C for 30 min. Then, the slides were put into xylene for dewaxing, which were subsequently washed using alcohol (China National Pharmaceutical Group Co., Ltd, Beijing, China). Next, the slides were repaired with 1 × EDTA (Beyotime Biotechnology Co., Ltd, Shanghai, China) and blocked with 3% H2O2 and serum. After that, the slides were incubated with primary antibodies and secondary antibodies at 4 °C overnight, and then were stained with DAB and hematoxylin (Baso DiagnosticsInc., Zhuhai, China). Finally, the slides were sealed with neutral resin (China National Pharmaceutical Group Co., Ltd, Beijing, China). The images were scored as negative (0), positive [1,2,3,4], +  + positive [5,6,7,8], or +  +  + positive [9,10,11,12], with reference to the sum of the staining intensity (varied from weak to strong) and staining extent scores, which graded as 0 (0%), 1 (1–25%), 2 (26–50%), 3 (51–75%), or 4 (76–100%). Antibodies were listed in Additional file 1: Table S1.

Plasmid construction and lentivirus infection

The TMED3 and CDCA8 shRNA target sequences, as well as CDCA8 overexpression plasmids were designed by Shanghai Bioscienceres Co., Ltd. (Shanghai, China), and subsequently were inserted into the BR-V-108 vector through the restriction sites at both ends and transformed into TOP 10 E. coli competent cells (Tiangen, Beijing, China). The plasmids of positive recombinants were extracted with the EndoFree maxi plasmid kit (Tiangen, Beijing, China), and the concentration of which was determined by a spectrophotometer (Thermo_Nanodrop 2000). A375 and OM431 cells were infected via adding 20 μL 1 × 108 TU/mL lentivirus, and then were respectively cultured in DMEM medium and 1640 medium (both containing 10% FBS) in a 6-well dish (2 × 105 cells/well). Finally, the infection efficiencies and knockdown/overexpression efficiencies were evaluated by microscopic fluorescence, qRT-PCR and western blot.

RNA extraction, cDNA synthesis and qRT-PCR

After infection, total RNA of A375 and OM431 cells was extracted using TRIzol reagent (Sigma, St. Louis, MO, USA) for cDNA synthesis. qRT-PCR assay was performed via using the Promega M-MLV Kit (Promega Corporation, Madison, Wisconsin, USA) and the SYBR Green Mastermixs Kit (Vazyme, Nanjing, Jiangsu, China). GAPDH was as an internal normalization control. The relative expression of mRNA was evaluated based on the 2−△△Ct method. The primers sequences (5ʹ-3′) were listed in Additional file 1: Table S2.

Western blot assay and Co-immunoprecipitation (Co-IP)

A375 and OM431 cells were collected after being infected indicated lentivirus to harvest protein. The protein purity was quantified with BCA methods. After that, total proteins were segregated using 10% SDS-PAGE and then transferred into PVDF membranes. Next, the membranes were subject to be blocked in TBST solution with 5% non-fat milk and subsequently incubated with primary antibodies and secondary antibodies, and then washed with TBST solution for three times (10 min/time). Finally, the ECL + plusTM Western blotting system kit was used for color rendering to capture X-ray imaging.

In terms of exogenous Co-IP assay, the proteins of 293 T cells infected with TMED3-Flag were immunoprecipitated with Flag, CDCA8 or IgG and then subjected to western blot. For endogenous Co-IP analysis, the proteins of A375 and OM431 cells were immunoprecipitated with IgG and anti-TMED3. After that, the proteins in the immunocomplex were separated by 10% SDS-PAGE and used for TMED3 and CDCA8 incubation. Antibodies were showed in Additional file 1: Table S1.

Cell proliferation detection and colony formation assay

In this study, cell proliferation was assessed by three methods including MTT assay, Celigo cell counting assay and CCK8 experiment. For MTT assay, A375 and OM431 cells with indicated lentivirus were seeded in a 96-well plate (2000/well). 20 μL MTT (5 mg/mL) and 100 μL DMSO were added. OD value at 490 nm wavelength was detected using a microplate reader (Tecan infinite, Mannedorf Zurich, Switzerland) for 5 days. In terms of Celigo cell counting assay, the cell images were taken by Celigo image cytometer (Nexcelom Bioscience, Lawrence, MA, USA), and a continuous 5-day cell proliferation curve was drawn. Another method was CCK8 experiment, A375 and OM431 cells with indicated lentivirus were cultured in a 96-well plate at the density of 2500 cells/well. 10 μL CCK8 reagent was added, and then the 96-well plate was oscillated for 2–5 min. OD value was detected for continuous 5 days by Microplate Reader (Tecan infinite) at 450 nm.

Colony formation assay

Lentivirus infected cells were seeded in a 6-well plate (1000 cells/well). 5 days later, the cells were washed with PBS, fixed with 1 mL 4% paraformaldehyde and stained by 500 μL Giemsa (Dingguo, Shanghai, China) to calculate the number of colonies.

Cell migration assay

Cell migration was assessed by transwell assay and wound-healing assay. A375 and OM431 cells infected with indicated lentivirus were seeded in a 24-well plate (1 × 105 cells/well). Subsequently, the cells were loaded into the upper chamber containing serum-free medium. Then, the upper chamber was transferred to the lower chamber with 30% FBS and incubated for 72 h. Finally, 400 µL Giemsa was used for cell staining and the cell migration ability was quantified. For wound-healing assay, cells were cultured in a 6-well plate at a density of 5 × 104 cell/well for 12 h and then wounded to generate a straight scratch. After cells were washed with PBS gently, serum-free medium was used to culture cells at 37 °C. Pictures were taken by a light microscope (DFC500; Leica) at appropriate time after wounding.

Flow cytometry assay

The abilities of cell apoptosis were assessed by flow cytometry using the Annexin V-APC single-staining or Annexin V-APC/PI double-staining method. In detail, lentivirus-infected A375 and OM431 cells were seeded in a 6-well plate at a density of 2 mL/well and cultured for 5 days. Then, 10 μL Annexin V-APC was added for staining at room temperature in the dark. The cell apoptosis level was measured by using FACSCalibur (BD Biosciences, San Jose, CA, USA). For double-staining experiment, cells were treated as described above. Differently, 5 μL Annexin V-APC and 5 μL PI solution were successively added before cells were performed on FACSCalibur detection.

Human apoptosis antibody array

Human Apoptosis Antibody Array was employed in A375 cells to visualize the changes of apoptosis-related protein expression in response to TMED3 depletion. After the cells were lysed, the Handling Array membranes were blocked in 2 mL 1 × Wash Buffer II and incubated with cell lysates and 1 × Biotin-conjugated Anti-Cytokines overnight at 4 °C. Finally, the signals of membranes were tracked by chemiluminescence imaging system.

PrimeView human gene expression array

Total RNA was extracted as described previously. The quality and integrity of RNA were determined by Nanodrop 2000 (Thremo Fisher Scientific, Waltham, MA, USA) and Agilent 2100 and Agilent RNA 6000 Nano Kit (Agilent, Santa Clara, CA, USA). Referring to the manufacturer’s instruction, RNA sequencing was performed with Affymetrix human GeneChip PrimeView and the data were scanned by Affymetrix Scanner 3000 (Affymetrix, Santa Clara, CA, USA). The statistical significance of raw data was completed by using a Welch t-test with Benjamini–Hochberg FDR (|Fold Change|≥ 1.3 and FDR < 0.05 as significant). Significant difference analysis and functional analysis based on Ingenuity Pathway Analysis (IPA) (Qiagen, Hilden, Germany) was executed, and |Z—score|> 0 is considered valuable.

The cancer genome atlas (TCGA) database analysis

The relationship between TMED3 and CDCA8 was established based on 103 primary SKCM tumor RNAseq samples from the TCGA database. Pearson correlations of gene expression values were calculated using the R cor.test function and plotted using the R packet ggplot2.

The construction of nude mouse tumor formation model

Female BALB-c nude mice (Four-week-old) were obtained from Shanghai Lingchang Animal Research Co., Ltd., and their care was conducted according to the ARRIVE guidelines. All animal experiments conformed to the European Parliament Directive (2010/63/EU) and were approved by the Ethics Committee of Shanghai Tenth People’s Hospital. The mice were randomly divided into indicated groups (10 mice/group) and were subcutaneously injected 5 × 106 A375 cells with shTMED3 or shCtrl. The tumor volume was tested during the entire feeding period. 0.7% sodium pentobarbital was injected intraperitoneally and the fluorescence was observed by the in vivo imaging system (IVIS Spectrum, Perkin Elmer). 17 days later, the mice were sacrificed and the tumors were removed to weigh and photograph, and finally frozen in liquid nitrogen and stored at − 80 °C.

Statistical analysis

Statistical analysis was carried out using GraphPad Prism 8 (San Diego, CA, USA) and SPSS 19.0 (IBM, SPSS, Chicago, IL, USA). All data were represented as the mean ± SD from at least three repeated experiments. Student’s t-test (for comparisons of two groups) and one-way ANOVA (for multiple group comparisons) were used to analyze the statistical significance. The Spearman correlation analysis and Mann–Whitney U analysis were used to assess the association between TMED3 expression and pathological characteristics of MM patients. P < 0.05 was considered to be statistically significant. All the experiments were in triplicate.