Landscape of expression pattern of TRIM E3 ligases in human CRC sample

To identify the differentially expressed genes in CRC, we analyzed the RNA-sequencing expression profiles for CRC downloaded from the TCGA dataset (https://portal.gdc.com). A volcano plot showed that 10 TRIMs in CRC samples were differentially expressed, while other TRIMs were unchanged (Fig. 1A). Five TRIM expression was significantly upregulated (TRIM14, TRIM15, TRIM24, TRIM29, and TRIM31), and the other five TRIM genes showed decreased expression (TRIM1, TRIM3, TRIM9, TRIM22, and TRIM73) in both colon cancer (COAD) and rectal cancer (READ) (Fig. 1A, B). Interestingly, the downregulated TRIMs and the upregulated TRIMs formed two phylogenetically distinct clusters, indicating their synergistic and divergent roles in CRC cancer development (Fig. 1C).

Fig. 1: Identification of differentially expressed TRIMs in colorectal cancer.

A Volcano map showing the overall transcriptional expression in CRC of tumor tissues (n = 620) matching the TCGA data and normal tissues (n = 830) matching the TCGA normal and GTEx data. Red dots refer to significantly up-regulated genes, blue dots correspond to the down-regulated genes, and gray dots indicate the non-significant change in gene expression. B Box plot showing the mRNA expression of differentially expressed TRIMs in COAD (n = 275 for tumor tissue and n = 349 for normal tissue) and READ (n = 92 for tumor tissue and n = 318 for normal tissue) from the TCGA normal and GTEx data. C Phylogenetic analyses of differentially expressed TRIMs. The amino acid sequence of TRIMs was aligned, and a phylogenetic tree was constructed in MEGA 5.0 using the neighbor-joining method.

Higher expression of TRIM1 predicts poor prognosis in CRC

To determine the significance of these differentially expressed TRIMs in CRC, we analyzed their associations with the prognostic value of CRC patients. We used the RNA-sequencing expression profiles and corresponding clinical information for CRC from the TCGA dataset. Kaplan–Meier survival curve showed that a high mRNA level of TRIM1 was significantly associated with poor overall survival (OS) and disease-free survival (DFS) in CRC (Fig. 2A, B). Univariate and multivariate Cox regression analyses exhibited that TRIM1 was an independent prognostic factor (Fig. 2C, D). However, the associations of other TRIMs with survival rates and prognostic values in CRC were not significant. These results showed that only TRIM1 expression could predict prognosis in CRC, emphasizing its role in CRC tumorigenesis. Hence, we choose TRIM1 for the subsequent investigations.

Fig. 2: Correlation between TRIMs expression and survival rate of CRC patients.

A, B Kaplan–Meier plots for the survival of CRC patients stratified by the mRNA expression level of each differentially expressed TRIMs. The overall survival curves are shown in (A). The disease-free survival curves are shown in (B). C, D Cox regression analysis of mRNA expression of each differentially expressed TRIMs in CRC patients from TCGA data (n = 620). The p value, hazard ratio (HR), and confidence interval of each TRIM in CRC are analyzed by univariate (C) and multivariate (D) Cox regression analysis. *p < 0.05, **p < 0.01.

TRIM1 expression is downregulated in CRC

The expression of TRIM1 was observed to be downregulated in the TCGA dataset (Fig. 1). To further verify this, we have provided another three pieces of evidence. An independent CRC cohort (GSE244551) containing normal and cancer tissues was examined. The results showed that TRIM1 expression was significantly upregulated in the normal tissues (Fig. 3A). Next, we collected four pairs of clinical samples containing the cancer and their adjacent tissues and found that TRIM1 was also upregulated expressed in the adjacent normal tissues (Fig. 3B). In addition, we evaluated the protein expression of TRIM1 in cancer tissues and the adjacent normal tissues of the colon using immunohistochemistry. The results showed that TRIM1 protein was mainly located around the glandular structure of the lumen and was relatively lowly expressed in CRC tissue (Fig. 3C, D). Besides, we examined the TRIM1 expression profile of TRIM1 among different cancers in the TGCA cohort using the GEPIA web tool. Compared to the normal tissues, TRIM1 is downregulated in the six types of cancer, including BLCA, COAD, READ, SKCM, UCEC, and UCS, while upregulated in THYM cancer, suggesting the expression varies among different cancers (Fig. 3E).

Fig. 3: TRIM1 expression is significantly downregulated in colorectal cancer.

A The mRNA expression of TRIM1 of CRC tumor tissues and their corresponding adjacent normal tissues matching GSE24551 data (n = 160 for tumor tissue and n = 13 for normal tissue). B The mRNA expression of TRIM1 in the tumor tissues compared to the adjacent normal tissues from four CRC patients paired samples. C, D Immunohistochemical staining of TRIM1 protein in paired samples from six CRC patients. Representative IHC images of TRIM1 were shown (C), and the IHC scores were calculated (D). E The transcriptional expression profile of TRIM1 in 33 types of tumor tissues (T) in TCGA and normal tissues (N) matching the TCGA normal and GTEx data. Red and green labels correspond to the cancer types in which TRIM1 expression is up- and down-regulated in tumor tissue. Scale bar, 50 μm. **p < 0.01.

TRIM1 is positively correlated with clinicopathological parameters and immunotherapy biomarkers of CRC

To explore the potential roles of TRIM1 in CRC development, we next analyzed the relationship between TRIM1 mRNA expression level and its clinical outcomes. We observed positive correlations between the expression level of TRIM1and the CRC tumor stage (Fig. 4A), the EMT signaling (Fig. 4B), and two malignant tumor marker genes Ki67 and KRAS (Fig. 4C), implying that TRIM1 may play a promotive role in CRC tumorigenesis.

Fig. 4: Higher expression of TRIM1 was significantly associated with poor prognostic and immunotherapy biomarkers in colorectal cancer.

A TRIM1 expression in different CRC pathological stages using TCGA data. B Correlation between TRIM1 expression and EMT marker by GESA enrichment analysis of CRC patients’ data. C Correlation between TRIM1 expression and MMR genes in CRC patients. D, E Correlation analysis between TRIM1 expression and MSI/TMB score of CRC patients. The abscissa represents gene expression distribution, and the ordinate represents MSI (D) and TMB (E) score distribution. The value in the panel represents the paired-sample number, correlation coefficient, and correlation p value. F Correlation between TRIM1 expression and immune cell infiltration levels in COAD and READ. G Correlation between the TRIM1 expression and immune checkpoint genes. Spearman’s correlation coefficients and p values were shown. *p < 0.05, **p < 0.01.

Growing studies have proved that microsatellite instability high MSI status (MSI-H) of mismatch repair deficient (dMMR) gene may predict immunotherapeutic response in CRC. dMMR-MSI-H signatures are typically closely related to the high tumor mutation burden (TMB-H) or immune cell infiltration [11, 12]. To determine the potential role of TRIM in immunotherapeutic response, we performed correlation analyses using the TCGA RNA-seq data of CRC samples. The TRIM1 mRNA level highly correlated with the dMMR-MSI-H signature in CRC samples, including three MMR genes (MSH2, MSH6, and PMS2) and MSI score (Fig. 4C, D). TRIM1 expression had non-significant correlations with TMB but showed positive correlation with infiltrating levels of immune cells (CD8+ T cells, CD4+ T cells, macrophage, neutrophils, and dendritic cells) in CRC (Fig. 4E, F). Consistently, TRIM1 mRNA level had dramatically positive coefficients with the canonical immune checkpoint genes (Fig. 4G). Together, these results elucidated the possible role of TRIM1 in regulating immunotherapeutic response in CRC.

TRIM1 promotes cell proliferation of CRC

Clinical analyses implied that TRIM1 played a tumor-promoting role in CRC, so we next examined the biological functions of TRIM1 in CRC cells. We synthesized four pairs of siRNAs for the loss-of-function study and found that the first and the second pairs showed an excellent silencing effect (Fig. 5A). Also, for the gain-of-function study, we constructed the functional pCS2-GFP-plasmid expressing the wild-type (WT) TRIM1 and the catalytically inactive mutant ΔRING TRIM1 for over-expression in CRC cells. Overexpression of TRIM1 in SW480 and LoVo cells dramatically increased the migration rate and the colony formation of CRC cells compared with the corresponding controls (Supplementary Fig. S1). Silencing of TRIM1 efficiently decreased the colony number of SW480 cells (Fig. 5B, C), attenuated the cell proliferation both in SW480 and LoVo cells (Fig. 5D), and slowed down the migration rate (Fig. 5E, F). Notably, this inhibition effect was not due to cell death because TRIM1 siRNA treatment did not induce apparent cell death based on the detection of the lactate dehydrogenase (LDH) release and the caspase-3 activity with or without the treatment of the apoptosis stimuli cisplatin (Supplementary Fig. S2). Next, we examined whether TRIM1 affects tumorigenesis in vivo. No significant body weight was lost during the TRIM1 siRNA, suggesting that TRIM1 was not overtly toxic in vivo (Fig. 5G). We observed that TRIM1 silencing can efficiently inhibit the growth and proliferation of SW480 cells in the nude mice xenograft model (Fig. 5H). Under treatment with TRIM1 siRNA, the tumor volumes and weights were significantly lower (Fig. 5I, J). Collectively, the above data demonstrated an oncogenic role of TRIM1 in CRC.

Fig. 5: TRIM1 silencing attenuates the colony formation, proliferation, and migration of colorectal cancer cells.

A–F Colorectal cancer cells were transfected with TRIM1 siRNA for 48 h, and then subjected to the colony formation assay, cell proliferation assay, or wound scratch assay. A Knockdown efficiency of TRIM1 siRNA was detected by immunoblotting. B, C Effects of TRIM1 knockdown on the colony formation of SW480 cells. Representative images were shown (B), and the colony number was calculated (C). D Effects of TRIM1 knockdown on the cell proliferation of SW480 and LoVo cells. E, F Effects of TRIM1 knockdown on the migration of SW480 cells. Representative images were shown (E), and the wound width was calculated (F). G–J Effects of TRIM1 knockdown on the growth and proliferation of SW480 cells in the nude mice xenograft model. Effects of TRIM1 knockdown on the tumor growth. Measurement was conducted every 3 days (G, H). The tumor volume and weight were measured on the 20th day after tumor inoculation (I, J). Results are as means ± SD from three independent experiments. Scale bar, 100 μm. *p < 0.05, **p < 0.01.

TRIM1 facilitates metabolism and restrains immune response

Our results indicate TRIM1 as an essential factor in promoting the proliferation of CRC cells. To investigate the crucial roles of TRIM1 in genome-wide gene expression changes and intracellular signaling pathways, we conducted a systematically transcriptional analysis of TRIM1-transfected SW480 cells was performed. Based on the RNA-seq analyses, TRIM1 transfection in SW480 cells led to the upregulation of 736 genes and the downregulation of 961 genes (Supplementary Fig. S3A). These DEGs were assigned to GO/KEGG analyses, and the top 20 enriched pathway lists were shown. The functions were primarily divided into positive regulation of metabolism (in red) and negative regulation of innate immune (in blue) (Fig. 6A and Supplementary Fig. S3B, C). The heat map showed the upregulation of critical metabolic genes and the downregulation of immune-related genes (Fig. 6B). Consistently, TRIM1 silencing by siRNA oligonucleotides results in the increased mRNA level of immune-related genes (TNFAIP3, CCL5, and RELB) and the decreased mRNA level of metabolic genes (ARNT2 and PGK1) (Fig. 6C).

Fig. 6: TRIM1 is critical for metabolism promotion and immune suppression.

A, B Systematic RNA-seq analysis from TRIM1-overexpressed SW480 cells. A Pathway enrichment of the DEGs by GO and KEGG analyses using the DAVID online tool. The top 20 pathways were listed. The circle size represents the number of DEGs enriched in this pathway. Red labels correspond to the upregulated pathways, and blue labels refer to the downregulated pathways. B Heatmap shows the synergistic expression patterns of the DEGs involved in regulating metabolism and immune response post-TRIM1 transfection. Color change from blue to red represents the expression levels of DEGs from low to high. C Effects of TRIM1 silencing on the expression of genes involved in regulating metabolism and immune response. After siRNA treatment for 48 h, the expression of the selective genes was examined by qRT-PCR. Primers were displayed in Supplementary Table S2. D, E Effects of TRIM1 overexpression and silencing on the NF-κB activities. D The plasmid for GFP-WT TRIM1, GFP- ΔRING TRIM1 or GFP was co-transfected with the plasmid constructs for NF-κB-Luc, adapter molecule TRAF2 or TRAF6. E After transfection of TRIM1 siRNA for 48 h, SW480 cells were co-transfected with plasmid constructs for NF-κB-Luc, adapter molecule TRAF2 or TRAF6 into SW480 cells. NF-κB activity in these samples was determined using the luciferase reporter assay. F, G Effects of TRIM1 knockdown on the NF-κB pathway. After siRNA treatment for 48 h, SW480 cells were added with TNF for another 12 h. F The expression of NF-κB pathway-related protein was examined by immunoblotting. G The level of NF-κB pathway activation was quantitated by measuring the ratio of band signal intensity for phosphorylated NF-κB/ total NF-κB, and IκB/GAPDH with Image J. Results are as means ± SD from three independent experiments. *p < 0.05, **p < 0.01.

To verify the roles of TRIM1 in the negative regulation of inflammation in vitro, we examined the canonical NF-κB pathway by NF-κB-luciferase assay and immunoblotting. Over-expression of FL TRIM1, but not ΔRING TRIM1, in SW480 significantly decreased the TRAF2/TRAF6-mediated NF-κB activity (Fig. 6D). Conversely, TRIM1 silencing by siRNA oligonucleotides results in an elevated NF-κB activity (Fig. 6E). Besides, TRIM1 knockdown increased the endogenous level of NF-κB phosphorylation and IκBα degradation induced by TNF, confirming the TRIM1-mediated NF-κB pathway blockade (Fig. 6F, G).

TRIM1 interacts with and catalyzes K63-linked ubiquitination on HIF1α

To further understand the molecular mechanism underlying the signaling pathways related to TRIM1 in CRC, we next analyzed the direct Protein interaction network (PPI) to determine potential interaction baits of TRIM1 (also called MID2). Besides the well-studied microtubule-binding protein MID1 and the ubiquitin-conjugating enzyme E2 D4 UBE2D4, we were surprised to find that TRIM1 was closely associated with the transcription factor hypoxia-inducible factor-1α (HIF1α) (Fig. 7A). Coimmunoprecipitation (co-IP) assay showed that over-expressed and endogenous TRIM1 and HIF1α could interact, which was detected by immunoblotting (Fig. 7B–D). Deletion of the N-terminal RING domain did not abolish this interaction (Fig. 7E). Also, TRIM1 was observed to co-localize with HIF1α at microtubules by confocal microscopy (Fig. 7F). TRIM1 is an E3 Ub ligase, and we next evaluated whether TRIM1 ubiquitinated HIF1α in vivo. Compared with the control plasmid, co‐transfection of FL TRIM1 with HIF1α results in robust ubiquitination of HIF1α, while expression of the enzymatically inactive TRIM1 did not induce additional ubiquitination (Fig. 7G). Besides, we used a series of lysine mutants of Ub to determine the poly-Ub chain type on HIF1α. Strong ubiquitination of HIF1α appeared in reactions containing wild-type (WT), K11R, K27R, K29R, K33R, or K63-only Ub (a mutant in which all Lys residues have been mutated to Arg residues except for Lys63). However, in the sample with the K63R or K48-only ubiquitin mutant, ubiquitination was largely inhibited (Fig. 7H). Thus, our data suggested that TRIM1 interacted with HIF1α on microtubules and accelerated its K63-conjugated ubiquitination.

Fig. 7: TRIM1 interacts with HIF1α and catalyzes its K63-linked ubiquitination.

A The protein–protein interaction network (PPI) of MID2/TRIM1 by GeneMANIA. Shown are the top 10 most related proteins. B–E The interaction between TRIM1 and HIF1α by coimmunoprecipitation (co-IP) assay. SW480 cells were co-transfected with the indicated plasmids. Samples lysed were immunoprecipitated with anti-Flag, anti-GFP or anti-TRIM1 antibody, and the input and immunoprecipitated samples were detected by immunoblotting with the indicated antibodies. B, C Overexpressed TRIM1 and HIF1α were co-immunoprecipitated with each other. D Endogenous TRIM1 and HIF1α can be co-immunoprecipitated. E ΔRING TRIM1 was coimmunoprecipitated with HIF1α. F HIF1α co-localized with TRIM1 at microtubules. SW480 cells were co-transfected with GFP-HIF1α and Flag-TRIM1 plasmids for 18 h. Shown are photos of the cellular localization of HIF1α (green) and TRIM1 (red). Scale bar, 10 μm. G Overexpression of TRIM1 promotes ubiquitination of HIF1α. GFP-HIF1α expressed-SW480 cells were transfected with the empty control vector or a plasmid expressing WT TRIM1 or ΔRING TRIM1 in the presence of the WT HA-ubiquitin. At 18 h post-transfection, GFP-HIF1α was immunoprecipitated with an anti-GFP antibody, followed by immunoblotting analysis with the corresponding antibodies. H TRIM1 catalyzes K63-linked polyubiquitination of HIF1α. GFP-HIF1α expressed-SW480 cells were transfected with Flag-TRIM1 plasmid or the empty control vector in the presence of the WT and mutated HA-ubiquitin. 18 h post-transfection, GFP-HIF1α was immunoprecipitated with an anti-GFP antibody, followed by immunoblotting analysis with the corresponding antibodies.

Lys214 of HIF1α is ubiquitinated by TRIM1 and is essential for HIF1α’s activity

To precisely map the modification site(s), we affinity-purified GFP-HIF1α from SW480 cells co-transfected with either wild-type TRIM1 or empty vector. From quantitative mass spectrometry, we detected six ubiquitination peptides of HIF1α. Our data reveal that peptide -214KPPMTcLVLIcEPIPHPSNIEIPLDSK240- was highly (~77.8%) ubiquitinated in the presence of TRIM1 (Fig. 8A). In contrast, the change of modification rate for the other modified peptides was below 10% (Supplementary Table S1). MS/MS analyses assigned the major modification site to Lys214 (K214) (Fig. 8A). K214 is predicated to be located within the loop between the two helix Per-ARNT-Sim (PAS) domains of HIF1α (Fig. 8B). Mutation of Lys214 to Arg of HIF1α did not abolish the binding with TRIM1 (Fig. 8C). However, the ubiquitination signals were significantly attenuated in samples expressing the HIF1α K214R mutant (Fig. 8D). Upon activation, the transcription factor HIF1α is translocated to the nucleus and binds the consensus HREs (hypoxia-responsive element) in the target gene promoter regions to initiate expression [13]. To mimic the HIF1α activity in vitro, we applied an HRE-luciferase reporter. Compared with the WT HIF1α, K214R decreased activity and displayed less nuclear localization under normoxic condition in CRC cells (Fig. 8E, F).

Fig. 8: Lysine 214 of HIF1α is ubiquitinated by TRIM1 and is essential for HIF1α’s activity.

A Mass spectrometric detection of modified peptides of HIF1α. Extracted ion chromatograms are shown with peak intensities indicating the relative amounts of either the modified or unmodified peptides of HIF1α in the presence/absence of TRIM1. The intensity of peptides of HIF1α is listed in Supplementary Table S1. B The protein motif and 3D structure visualization of K214 in HIF1α. The complete structure of HIF1α is currently unavailable and predicted by the AlphaFold online tool (https://alphafold.ebi.ac.uk/). C The interaction between TRIM1 and the WT or K214R HIF1α by Co-IP assay. SW480 cells were co-transfected with the indicated plasmids. Samples lysed were immunoprecipitated with anti-Flag antibody, and the input and immunoprecipitated samples were detected by immunoblotting with the indicated antibodies. D Validation of K214 as a ubiquitination modification site of HIF1α by TRIM1. Flag-TRIM1 expressed-SW480 cells were transfected with the empty control vector or a plasmid expression WT HIF1α or K214R HIF1α in the presence of the WT HA-ubiquitin. At 18 h post-transfection, GFP-HIF1α was immunoprecipitated followed by immunoblotting analysis with the corresponding antibodies. E Effect of K214 mutation on the HIF1α activity. The plasmid construct for WT or K214R GFP HIF1α was co-transfected with the plasmid constructs for HRE-Luc into SW480 cells. HIF1α activity was determined using a luciferase reporter assay. The expression was shown below. *p < 0.05. F Effect of K214 mutation on the nucleus distribution of the HIF1α. The plasmid construct for WT or K214R GFP HIF1α was transfected into SW480 cells under normoxic conditions for 18 h. Shown are representative images showing cellular localization of HIF1α (green) and nucleus (blue). The percentage of GFP-HIF1α in the cytosol for each sample is indicated. At least 50 cells were counted for samples from experiments done in triplicate. Scale bar, 10 μm.

TRIM1 promotes HIF1α activity by accelerating its nuclear translocation

Then, we sought to determine the consequences of HIF1α ubiquitination by TRIM1. Over-expression of the WT but not the enzymatically active TRIM1 significantly elevated the HRE activity (Fig. 9A). Conversely, TRIM1 knockdown by siRNA oligonucleotides results in an attenuated HRE activity induced by DMOG (a HIF1α activator) and hypoxic treatment (Fig. 9B, C). Knockdown of HIF1α significantly decreased HRE activity induced by TRIM1 and DMOG (Fig. 9D, E). Besides, our transcriptome results showed the increased expression of HIF1α-downstream genes in the TRIM1-transfection sample (Fig. 9F), confirming TRIM1-mediated HIF1α activation. Although several E3 ligases have been reported to regulate HIF1α’s activity via alteration of its expression level or protein stability, our results showed that TRIM1 expression or silencing did not alter the HIF1α mRNA level (Figs. 9F and 6C). Chase experiments with cycloheximide (CHX) showed that TRIM1 expression also did not affect the protein stability of HIF1α (Supplementary Fig. S4). Interestingly, WT TRIM1 overexpression led to the nucleus translocation of endogenous HIF1α (Fig. 9G, H) after nucleus and cytoplasmic fractionation. These results suggest that TRIM1 activates HIF1α signaling by accelerating its nucleus translocation instead of altering its expression.

Fig. 9: TRIM1 promotes HIF1α activity by accelerating its nuclear translocation.

A Effects of TRIM1 over-expression on the HIF1α activities. The plasmid construct for WT GFP-TRIM1, ΔRING GFP-TRIM1, or GFP was co-transfected with the plasmid constructs for HRE-Luc into SW480 cells. HIF1α activity was determined using a luciferase reporter assay. B Effects of TRIM1 knockdown on the DMOG-induced HIF1α activities. After transfection of TRIM1 siRNA for 48 h, SW480 cells were transfected with the HRE-Luc plasmid in the presence or absence of DMOG. C Effects of TRIM1 knockdown on the hypoxia induced HIF1α activities. After transfection of TRIM1 siRNA for 48 h, SW480 and LoVo cells were transfected with the HRE-Luc plasmid. After 10 h, cells were subjected to the hypoxic treatment (1% O2, 5% CO2, 95% humidity) for another 18 h. HIF1α activity was determined using a luciferase reporter assay. D, E Effects of HIF1α knockdown on the TRIM1-mediated HRE promoter activity. D The silencing efficiency of HIF1α siRNA was determined by immunoblotting. E After transfection of 4# HIF1α siRNA for 48 h, HRE-Luc plasmid was co-transfected with a plasmid construct for GFP-TRIM1 or GFP into SW480 cells. DMOG treatment acted as the positive control. F Effects of TRIM1 expression on the mRNA expression of HIF1α and the HIF1α-responsive genes. The figure was generated from our transcriptome data. G, H Effects of TRIM1 over-expression on the nucleus distribution of the endogenous HIF1α. G SW480 cells were transfected with a plasmid for WT Flag- TRIM1, ΔRING Flag-TRIM1 or GFP for 18 h. Total nucleus (N) and cytosol proteins (C) were fractionated and immunoblotted with the indicated antibodies. H The nuclear distribution of HIF1α was quantitated by determining the ratio of band signal intensity for HIF1α/ H3 in (G) with Image J software. I Quantification of the TRIM1-mediated HBPs in SW480 cells. GFP-HIF1α were co-expressed with Flag-tagged WT TRIM1, ΔRING TRIM1 or pVec into SW480 cells for 18 h. Lysates were subjected to IP with GFP-specific antibody. The precipitates were further separated by SDS-PAGE before in-gel digestion with trypsin and LC-MS/MS analyses. Scatter plots of protein ratios as a function of their relative abundance (denoted by MS/MS spectral counts). The ratio is calculated as spectral counts in FL TRIM1 transfected samples divided by those in controls. Higher ratios indicate increased binding efficiency with HIF1α. Red dots correspond to the potential HBPs involved in nuclear import, the green dots correspond to immunoprecipitated HIF1α, and the yellow dots correspond to TRIM1. Results are as means ± SD from three independent experiments. *p < 0.05, **p < 0.01. J Schematic diagram of this work. TRIM1 expression promotes the proliferation and migration of colorectal cancer cells and predicts poor prognosis for CRC patients. Mechanistically, TRIM1 interacts with HIF1α, catalyzes its K63-linked ubiquitination, and promotes its nuclear translocation. HIF1α in the nuclear then binds the HRE region in the promoter, initiates the expression of downstream genes, promotes cellular metabolism, and attenuates immune response.

TRIM1 facilitates the association of HIF1α with nucleus transport proteins

To investigate the molecular mechanisms underlying TRIM1-mediated HIF1α nucleus translocation, we used mass spectrometry to analyze proteins pulled down by HIF1α with and without TRIM1 (Supplementary Fig. S5A). HIF1α interactome analyses showed that the HIF1α-binding proteins (HBPs) were more abundant in TRIM1-transfected samples. Among these, 308 differential HBPs overlapped in TRIM1/ΔRING and TRIM1/pVec groups (Supplementary Fig. S5B). KEGG analysis of the putative HBPs revealed several enriched metabolic pathways (Supplementary Fig. S5C). Interestingly, seven proteins are involved in the nucleocytoplasmic transport pathway (in red), including DDX39B, SEH1L, EEF1A2, NUP153, NUP214, NUP98, and TNPO1 (Fig. 9I). Consistent with this notion, we also determined the effects of the modification site (K214) of HIF1α on the HBPs and obtained similar HIF1α interactome results. When K214 was mutated, there was also a drastic reduction in the redundancy of these HBPs (Supplementary Fig. 5D, E). Therefore, we speculate that TRIM1 potentially enhances the interaction of HIF1α with nuclear transport proteins through ubiquitination at K214 of HIF1α, increasing its probability of entering the nucleus.