The CARM1-p300-c-Myc-Max (CPCM) transcriptional complex regulates the expression of CUL4A/4B and affects the stability of CRL4 E3 ligases in colorectal cancer

The transcription factor c-Myc and two cullin family members CUL4A/4B function as oncogenes in colorectal cancer. Our recent publication reveals that c-Myc specifically activates the expression of CUL4A/4B through binding to their promoters. However, the underlying mechanism of how c-Myc actions in this process is still unknown. Using mass spectrometry and immunoprecipitation assays, we identified c-Myc formed a transcriptional complex with its partner Max (Myc-associated factor X), a histone acetyltransferase p300 and a coactivator associated arginine methyltransferase 1 (CARM1) in the present study. Knockdown or overexpression of the components of CARM1-p300-c-Myc-Max (CPCM) complex resulted in a decrease or increase of CUL4A/4B levels, respectively. Individual knockdown or inhibition of CPCM components decreased cell proliferation, colony formation, and cell invasion. Biochemically, knockdown or inhibition of CPCM components decreased their occupancies on the promoters of CUL4A/4B and resulted in their downregulation. Importantly, inhibition of CPCM components also caused a decrease of CRL4 E3 ligase activities and eventually led to an accumulation of ST7 (suppression of tumorigenicity 7), the specific substrate of CRL4 E3 ligases in colorectal cancer. Moreover, the in vivo tumor formation results indicated that knockdown or inhibition of CPCM components significantly decreased the tumor volumes. Together, our results suggest that the CPCM complex mediates explicitly the expression of CUL4A/4B, and thus affects the stability of CRL4 E3 ligases and the ubiquitination of ST7. These results provide more options by targeting the CPCM components to inhibit tumor growth in the therapy of colorectal cancer.


Introduction
Colorectal cancer (CRC) is one of the highest incidences and mortality of cancer [1,2]. According to the clinical data from the North American Association of Central Center Registries (NAACCR), the incidence and mortality of CRC in the United States are 0.04% and 0.016%, respectively [1,2]. In the past decades, the molecular mechanisms regarding CRC tumorigenesis have been extensively investigated and current evidence recognizes that genomic instability, genetic factors, inflammatory microenvironment, aberrant expression of tumor suppressors and oncogenes, and differentially expressed noncoding RNAs [e.g., microRNAs (miRNAs) and long noncoding RNAs (lncRNAs)] are the major contributors of CRC pathogenesis [3][4][5][6]. At least three distinct pathways involved in genomic instability have been reported, and they include chromosomal instability (with an incidence of 65%-70% in sporadic colorectal cancers),
To explore the mechanism of how c-Myc activates the expression of CUL4A/4B in CRC cells, we immunoprecipitated c-Myc-associated complex and applied it to mass spectrometry analysis. After coimmunoprecipitation assay, we discovered that c-Myc dimerized with its partner protein Max, and directly interacted with a histone acetyltransferase p300, which further recruited CARM1 (Coactivator Associated Arginine Methyltransferase 1) to assemble a transcriptional complex known as CARM1-p300-c-Myc-Max (CPCM). We then focused our studies on evaluating the contribution of CPCM components to the expression of CUL4A/4B and CRL4 DCAF4 E3 ligase activities.

Cell transfection
Cells were seeded into 6-well plates and incubated overnight to reach a density of 50% confluence. Specific siRNAs including sip300 (Thermo Fisher Scientific, Waltham, MA, USA, assay ID:106444), sic-Myc (assay ID: 103828), siCARM1 (assay ID: 112501), and siMax (assay ID: 143519) for gene knockdown or plasmids for gene overexpression were mixed with Lipofectamine 2000 (Thermo Fisher Scientific, #11668019) to form siRNA (or plasmid) duplex-Lipofectamine, which were then added into 6-well plates containing cells. Three replicates were performed for each siRNA or plasmid. After mixing gently, cells were further incubated 24 h at 37°C in a CO 2 incubator. The resulting cells were subjected to RNA and protein extraction.

Immunoprecipitation and mass spectrometry
Cells (1×10 8 ) expressing pCDNA3-2×Flag-3×HA-c-Myc or pCDNA3-2×Flag-3×HA (Empty vector, control) were lysed in RIPA lysis and extraction buffer (Thermo Fisher Scientific, #89900) containing 1 × protease inhibitor cocktail (Sigma-Aldrich, #P8340). Cell extracts were centrifuged at 14,000 rpm for 30 min and the supernatant was incubated with anti-Flag agarose (Sigma-Aldrich, #A2220) to pull down Flag-associated proteins at 4°C overnight. The resulting Flag-associated proteins were washed five times with RIPA lysis and extraction buffer and then eluted with 100 µg/mL Flag peptide (Sigma-Aldrich, #F4799). The obtained protein complex was further incubated with anti-HA agarose (Thermo Fisher Scientific, #26181) to pull down HA-associated proteins at 4°C overnight. The resulting HA-associated proteins were washed five times with RIPA lysis and extraction buffer, followed by loading onto a 10% SDS-PAGE gel for separation. The gel was subsequently performed sliver staining with a kit (Thermo Fisher Scientific, #24612). Protein bands were cut into small slices and then digested with a Trypsin Kit (Thermo Fisher Scientific, #60109101). Mass spectrometry analysis was performed to determine c-Myc-associated proteins following a protocol as described previously [27].

Cell proliferation assay
Cell viability was determined using a CellTiter 96 non-radioactive cell proliferation kit (Promega, Madison, WI, USA, #G4000). Briefly, HT29 and HCT-116 cells were transfected with sip300, sic-Myc, siCARM1, and siMax to generate their corresponding knockdown cells. These cells were plated onto 96-well plates and cell viability was determined every day for five days according to the manufacturer's method. In addition, HT29 and HCT-116 cells were grown in DMEM and the same medium containing 5 µM sAJM, 20 µM CARM-IN-1 or 50 nM C646. Cell viability was also determined every day for five days.

Colony formation assay
The c-Myc-knockdown (KD), p300-KD, CARM1-KD and Max-KD cells in HT-29 and HCT-116 backgrounds were seeded onto 6-well plates with a density of 10 3 cells per well. Cells were incubated at 37°C for two weeks with a medium change every three days. For colony formation assay in cells treated with CPCM component inhibitors, the HT-29 and HCT-116 were seed to 6-well plates with a density of 10 3 cells per well. Cells were cultured in a 37°C incubator to adhere for 16 h, followed by treatment with 5 µM sAJM, 20 µM CARM-IN-1 or 50 nM C646 for two weeks with a medium change every three days. Colonies were fixed with 70% ethanol for 10 min and stained with 0.2% crystal violet and the 6-well plates were photographed.

Cell invasion assay
The knockdown cells (5×10 4 ) of CPCM components and cells treated with individual CPCM component inhibitor were suspended into 100 µL serum-free DMEM medium and plated on the top filter membrane in a Boyden chambers insert (Millipore, Burlington, MA, USA, #P18P01250). The lower chamber was filled with DMEM medium containing 10% FBS. After incubation at 37°C for 24 h, cells on the lower chamber were fixed with 70% ethanol for 10 min and then were stained with 0.1% crystal violet, followed by a photograph.

Total RNA isolation and quantitative real-time PCR (qRT-PCR) analysis
Total RNA was isolated from cultured cells using a TRIZOL reagent (Thermo Fisher Scientific, #15596026) according to the method provided by the manufacturer.
The purified RNA was reverse-transcribed into cDNA using an M-MuLV reverse transcriptase kit (New England Biolabs, Beijing, China, #M0253S). After dilution 10-fold, cDNAs were applied to qRT-PCR using an SYBR Green Kit (Bio-Rad, Shanghai, China, #1725150) to quantify the expression of genes with primers listed in Supplementary Table 1. The PCR procedures included: 95°C for 2 min, followed by 40 cycles of 30 seconds at 95°C and 20 seconds at 68°C. The individual gene expression level was normalized to β-actin.

In vivo ubiquitination assay
The in vivo ubiquitination of ST7 was performed following a previous method [27]. Briefly, HCEC-1CT cells (5×10 7 ) expressing pCDNA3-2×Flag-ST7 and HA-ubiquitin were treated with 5 µM sAJM, 20 µM CARM-IN-1 or 50 nM C646 for 6 h, followed by lysing RIPA lysis and extraction buffer containing 1 × protease inhibitor cocktail. After centrifuging at 14,000 rpm for 30 min, the supernatant was incubated with anti-Flag-agarose to pull down Flag-ST7-associated proteins. The ubiquitination of ST7 was detected with an anti-HA antibody.

Chromatin immunoprecipitation (ChIP) assay
Cells (1×10 8 ) under 80% confluence were washed twice with cold PBS buffer (Thermo Fisher Scientific, #20012027), and then were crosslinked with 1% formaldehyde (Polysciences, Warrington, PA, USA, #18814) at 23°C for 15 min. The crosslinked cells were applied to a ChIP assay using a high-sensitivity ChIP kit (Abcam, #ab185913) according to a protocol provided by the manufacturer. The antibodies used in this assay included anti-c-Myc, anti-p300, anti-CARM1 and anti-Max. The information of these antibodies was the same as described in western blotting assay. The purified DNA samples were applied to qRT-PCR analyses using an SYBR green kit (same as described in mRNA detection) with the primers listed in Supplementary Table 2. The relative enrichment of individual CPCM components on the promoters of CUL4A and CUL4B were normalized to the input.

In vivo tumor formation and growth inhibition
The Athymic nu/nu mice were sourced from Shanghai SLAC Laboratory Animal Co. Ltd (Shanghai, China) and were maintained in accordance with a guideline approved by the Institutional Animal Care and Use Committee (IACUC) of Sichuan University. The HT29, c-Myc-KD, p300-KD, CARM1-KD and Max-KD cells (5×10 6 each) in 100 μL PBS were mixed with Matrigel (1:1 ratio, v/v) (BD Biosciences, San Jose, CA, USA, #354234). Cells were subcutaneously injected into mice to establish tumor xenografts and tumors were measured with fine calipers at 5-day intervals. In addition, mice injected with HT29 cells were randomly assigned to four groups (n = 5 in each group), followed by injecting with PBS, sAJM, CARM-IN-1 or C646 every five days. Tumor volumes were determined with a formula: Volume =(Length×Width 2 )/2.

Statistical analysis
The mean ± standard deviation (SD) in each experiment represented three independent replicates. Data were analyzed using a two-sided Student's t test.

c-Myc associated with Max, p300, and CARM1 to assemble the CPCM complex in vitro and in vivo
As mentioned earlier, our recent findings reported that the amplified c-Myc in CRC cells specifically bound to the promoters of CUL4A/4B and to activate their expression [27]. The induced CUL4A and CUL4B separately formed a CRL4 E3 ligase to ubiquitinate ST7, resulting in tumorigenesis [27]. To reveal how c-Myc cooperates with other proteins to assemble a transcriptional complex in this process, we constructed a c-Myc overexpression vector pCDNA3-2×Flag-3×HA-c-Myc. After transfecting it into HT29 cells, we performed a two-step purification using anti-Flag agarose and anti-HA agarose to enrich c-Myc-associated proteins ( Figure 1A). We then applied this complex to mass spectrometry assay to identify proteins. The results identified a total of 33 proteins in this complex (Supplementary Table 3). After analyzing the results, we found Max was a well-known partner of c-Myc and they could form a heterodimer, which could directly bind to DNA [41]. Based on this notion, we thought it was not necessary to determine the interaction between c-Myc and Max in our study. Moreover, we also found two other proteins including p300 and CARM1 have been previously reported to form a transcriptional complex with other transcription factors such as NF-κB (Nuclear Factor Kappa B) [42], Runx2 (Runt-related Transcription Factor 2) and bHLH [43,44]. Given the conserved assembly mechanism of transcription factors with coactivators and corepressors, we speculated that c-Myc might interact with p300 and CARM1. To verify this hypothesis, we performed Co-IP assays to determine the direct interactions of c-Myc-p300, c-Myc-CARM1, and CARM1-p300. Accordingly, we transfected the Myc-tag and Flag-tag vectors of these three proteins into HT29 cells. After immunoprecipitation using both anti-Flag agarose and anti-Myc agarose in each combination of plasmids, the output proteins were subjected to immunoblots to determine protein interactions. Our results indicated that c-Myc could directly interact with p300 instead of CARM1 ( Figure 1B), and p300 could directly interact with CARM1 ( Figure 1C). These in vitro results suggested that p300 functioned as an adaptor protein to connect c-Myc-Max heterodimer and CARM1, forming the CPCM complex. We next aimed to determine if these three proteins formed a complex in vivo. For this purpose, we performed in vivo immunoprecipitation using anti-c-Myc antibody in the cancerous tissue from an advanced colitis-associated cancer (CAC, a subtype of CRC) patient under stage IV. Immunoblot detection results using the purified protein complex indicated that c-Myc could pull down Max, CARM1, and p300 ( Figure 1D). These in vitro and in vivo results demonstrated that c-Myc-Max heterodimer recruited p300 and CARM1 to assemble the CPCM complex.

The components of CPCM complex were upregulated in cancerous tissues of CAC patients and cultured CRC cells
Our previous publication has reported that c-Myc is overexpressed in 48 cancerous tissues from CAC patients in comparison to their adjacent noncancerous tissues [27]. To determine the expression levels of other CPCM components in the same RNA samples of CAC cancerous tissues, we performed qRT-PCR analyses to measure mRNA levels of Max, p300 and CARM1. Our results showed that all of these three CPCM components were upregulated in 48 cancerous tissues compared to their adjacent noncancerous tissues (Figures 2A-2C). Meanwhile, we also detected their protein levels in five CAC cancerous tissues (n=1 in each TNM grade). Consistent with their mRNA levels, the CPCM member protein levels were gradually increased in the CAC tumor tissues with the severity of TNM stages (Figures 2D and 2E). To determine if the overexpression of CPCM components happens in CRC cells, we measured their mRNA levels in seven human CRC cell lines including HT29, HT55, HCT-15, HCT-116, HCA-24, SW620 and T84. The qRT-PCR results indicated that these seven cell lines exhibited varying mRNA levels of CPCM components ( Figure  2F). Of these cell lines, HT29 exhibited the highest mRNA levels of c-Myc (~6.5-fold), Max (~5.8-fold), p300 (~3.5-fold), and CARM1 (~5.6-fold), followed by HCT-116, HCT-15, HCA-24, HT55, SW620 and T84 ( Figure 2F). These results suggested that the overexpression of CPCM components in CAC cancerous tissues and cultured CRC cells was a universal phenomenon. Based on the higher expression levels of the CPCM components in HT29 and HCT-116 cells, we carried out the following experiments in these two cell lines unless otherwise specified.

Knockdown of CPCM components resulted in downregulation of CUL4A/4B
Both CUL4A and CUL4B are direct targets of c-Myc and knockdown of c-Myc leads to the downregulation of CUL4A/4B [27]. Thus, we next sought to determine the effects of other CPCM components on CUL4A/4B expression. For this purpose, we transfected CPCM component siRNAs or their overexpression vectors into HT29 and HCT-116 cells to specifically knock down or overexpress CPCM members. After determining their successful downregulation or overexpression ( Figures 3A-3D), we determined the expression of CUL4A/4B in these cells. As expected, our qRT-PCR results showed that both CUL4A and CUL4B were significantly downregulated in all CPCM knockdown cells compared to controls ( Figures 3E). Conversely, the After two-step purification, the resulting protein complexes were loaded onto an SDS-PAGE gel for separation. The protein gel was stained with a sliver-staining kit. The positions of IgG, CARM1, and HA-c-Myc were shown. (B) c-Myc interacted directly with p300 but not CARM1. The HT29 cells cotransfected with different combinations of plasmids including pCDNA3-2×Flag + pCDNA3-6×Myc, pCDNA3-2×Flag + pCDNA3-6×Myc-p300, pCDNA3-2×Flag + pCDNA3-6×Myc-CARM1, pCDNA3-2×Flag-c-Myc + pCDNA3-6×Myc, pCDNA3-2×Flag-c-Myc + pCDNA3-6×Myc-p300, and pCDNA3-2×Flag-c-Myc + pCDNA3-6×Myc-CARM1. The resulting cells were applied for Co-IP analyses with anti-Flag and anti-Myc resins. The input and out proteins were detected using anti-Flag and anti-Myc antibodies, respectively. (C) p300 interacted directly with CARM1. The HT29 cells cotransfected with different combinations of plasmids including pCDNA3-2×Flag + pCDNA3-6×Myc, pCDNA3-2×Flag + pCDNA3-6×Myc-p300, pCDNA3-2×Flag-CARM1 + pCDNA3-6×Myc, and pCDNA3-2×Flag-CARM1 + pCDNA3-6×Myc-p300. The resulting cells were applied for Co-IP analyses with anti-Flag and anti-Myc resins. The input and out proteins were detected using anti-Flag and anti-Myc antibodies, respectively. (D) c-Myc associated with Max, p300 and CARM1 in vivo. One cancerous tissue from a CAC patient under stage IV was applied to immunoprecipitation analysis using anti-c-Myc and anti-IgG antibodies, respectively. The purified protein complexes were applied to western blotting assays to examine c-Myc, Max, p300 and CARM1. expression of CUL4A/4B was markedly upregulated in all CPCM overexpression cells compared to control cells ( Figure 3F). In addition, we also examined the protein levels of CUL4A/4B and ST7 in these knockdown and overexpression cells. Consistent with their mRNA levels, we also observed a significant decrease or increase in CUL4A/4B protein levels in CPCM knockdown or overexpression cells, respectively (Supplementary Figure 1). In contrast, the protein level of ST7 was accumulated or decreased in CPCM knockdown or overexpression cells, respectively (Supplementary Figure 1). These results clearly supported that the CPCM complex was responsible for the regulation of CUL4A/4B expression. Since CARM1, p300, c-Myc and Max could assemble a complex, knocking down or overexpressing any two of them simultaneously should cause a similar effect on CUL4A/4B expression as in cells only knocking down or overexpressing a single member. To verify this hypothesis, we simultaneously knocked down or overexpressed c-Myc and CARM1 in HT29 cells, and then examined the mRNA and protein levels of CUL4A/4B. As expected, our results showed that the mRNA and protein levels of CUL4A/4B in cells knocking down or overexpressing c-Myc+CARM1 were similar to cells knocking down or overexpressing c-Myc or CARM1 alone (Supplementary Figure 2).

Proinflammatory cytokines activated the expression of CPCM components
We previously showed that proinflammatory cytokines including interleukin-6 (IL-6) and tumor necrosis factor alpha (TNF-α) could induce the expression of c-Myc, CUL4A and CUL4B [27]. Thus, we next aimed to determine if treatments with proinflammatory cytokines also affect the expression of other CPCM components. Accordingly, we treated HCEC-1CT cells with a series of concentrations of IL-6 (0, 40, 80, 120, 160 and 200 ng/mL) and TNF-α (0, 4, 8, 12, 16 and 20 ng/mL), followed by examining mRNA levels of CPCM components. The qRT-PCR results showed that the expression of CPCM components was gradually induced with the increase of both IL-6 and TNF-α concentrations (Figures 4A and 4B). Specifically, treatments with 200 ng/mL IL-6 and 20 ng/mL TNF-α resulted in the induction of c-Myc (~4.7-fold and ~5.5-fold, respectively), Max (~4.3-fold and ~4.8-fold, respectively), p300 (~2.7-fold and ~2.9-fold, respectively), and CARM1 (~6.1-fold and ~6.7-fold, respectively) ( Figures 4A and 4B). Consistent with previous results, we also examined CUL4A/4B mRNA levels in these treatments and found that they were gradually induced with the increase of IL-6 and TNF-α concentrations ( Figures 4C  and 4D). In addition, we also examined the protein levels of CPCM components and CUL4A/4B. The immunoblot results indicated that these proteins shared similar patterns to their corresponding mRNA levels ( Figures 4E and 4F). These results consistently supported that intracellular inflammatory status was responsible for the upregulation of CPCM components and their target genes CUL4A and CUL4B.

Knockdown or inhibition of the CPCM components caused oncogenic phenotype defects
Our previous results have shown that knockdown of c-Myc can inhibit oncogenic phenotypes of colorectal cells [27]. We next aimed to determine if knockdown of the other CPCM components also had similar effects. For this purpose, we subjected CPCM knockdown cells to evaluate their oncogenic phenotypes. The cell proliferation results indicated knockdown either Max, CARM1 or p300 significantly inhibited the growth of colorectal cancer cells to a comparable level in c-Myc-knockdown cells (Figures 5A and 5B). The colony formation results also indicated that the downregulation of CPCM components caused decreased colonies ( Figure 5C and Supplementary Figure 3A). Moreover, we also observed a significant decrease in invading cells in these knockdown cells compared to controls ( Figure  5D and Supplementary Figure 3B). Since knockdown of CPCM components inhibited cancer cell growth, we speculated that inhibition of this complex by the inhibitors that specifically targeted CPCM components should also cause similar effects. To verify this hypothesis, we treated colorectal cells with a c-Myc inhibitor (sAJM-589), a CARM1 inhibitor (CARM1-IN-1) and a p300 inhibitor (C646), respectively. After these treatments, we primarily measured mRNA and protein levels of CPCM components. The results indicated that all these treatments could not change mRNA and protein levels of these CPCM components ( Supplementary  Figures 4A and 4B). However, these treatments caused the downregulation of CUL4A/4B and the accumulation of ST7 (Supplementary Figure 4B), which suggested that oncogenic phenotypes might be inhibited following these inhibitor treatments. To verify this hypothesis, we also evaluated cell proliferation, colony formation and cell invasion abilities under CPCM inhibitor treatments. Similar to the results in their knockdown cells, we also observed the significant repression of cell proliferation (70% deduction at the 5-day point) (Supplementary Figures  5A-5C), colony numbers (75% deduction at the 5-day point) (Supplementary Figures  5D-5F, and Supplementary Figure 6A) and invading cell numbers (70% deduction at the 5-day point) (Supplementary Figures 5G-5I, and Supplementary Figure 6B) in comparison to non-treatment cells.

Inhibition of the CPCM components impaired their bindings on the promoters of CUL4A/4B
To determine if the CPCM complex regulated CUL4A/4B expression through binding to their promoters, we carried out ChIP assays using antibodies that recognized CPCM components. Firstly, we performed ChIP assays in HT29, c-Myc-KD, CARM1-KD and p300-KD cells without any treatment. The qRT-PCR results showed that the occupancies of CPCM components on the promoters of CUL4A/4B were significantly decreased (~60-70% deduction) in c-Myc-KD, CARM1-KD and p300-KD cells compared to HT19 cells (Figures 6A and 6B). Besides, we also evaluated the occupancies of CPCM components in cells treated with IL-6 (200 ng/mL) alone and in cells treated with both CPCM component inhibitors and IL-6. The results showed that IL-6 treatment significantly increased (~40-fold) the occupancies of CPCM components on the promoters of CUL4A/4B compared to controls (Figures 6C and  6D). The combined treatments of CPCM component inhibitors and IL-6 significantly decreased their occupancies from ~40-fold to ~4-fold ( Figures 6C and  6D). At the same time, we also examined mRNA levels of CUL4A/4B in cells treated with CPCM component inhibitors and IL-6. The results showed that the expression of CUL4A/4B was significantly induced (~8.3-fold) in cells treated with IL-6 alone compared to HT29 control cells, while their expression was only slightly induced (~1.5-fold) by IL-6 after CPCM component inhibitor treatments (Supplementary Figure 7). These data suggested that the CPCM complex bound explicitly to the CUL4A/4B promoters and activated their expression under IL-6 treatment. or 50 nM C646 for 6 h, followed by treatment with 200 ng/mL IL-6 for another 6 h. The resulting cells were subjected to ChIP assays using anti-c-Myc, anti-CARM1, anti-p300 or IgG for immunoprecipitation. The purified DNA was used to examine the enrichment of CPCM components on the promoters of CUL4A (C) and CUL4B (D). **P < 0.01 and ***P < 0.001.

Inhibition of the CPCM components repressed the ubiquitination of ST7
One possibility for the reason that caused the defects of oncogenic phenotypes in cells with CPCM component knockdown or inhibition was that CRL4 E3 ligase activities were repressed. To very this possibility, we measured protein levels of ST7 and its ubiquitination under the conditions of inhibitor treatments. As expected, the results indicated that the ST7 protein level was significantly accumulated ( Figure 7A). Consistent with the results in CPCM knockdown cells, we also observed inhibition of the CPCM components caused the decrease of CUL4A and CUL4B protein levels, while these inhibitors could not change the protein levels of CPCM components ( Figure 7A). To evaluate the ubiquitination level of ST7 in the treatments of CPCM component inhibitors, we primarily cotransfected pCDNA3-2×Flag-ST7 with pCDNA3-2×HA-Ubiquitin into HCEC-1CT cells, followed by treated with sAJM-589, CARM1-IN-1 or C646. After immunoprecipitation with anti-Flag resin, we detected ST7 ubiquitination level. The results indicated that CPCM component inhibitors significantly decreased the ST7 ubiquitination level to a similar pattern ( Figure 7B). These results suggested that the inhibition of the CPCM components repressed the ubiquitination of ST7.

Knockdown or inhibition of CPCM components decreased the tumor formation in vivo
Our above in vitro results showed that knockdown or inhibition of CPCM components decreased the growth of CRC cells. To evaluate their in vivo effects, we injected HT29, c-Myc-KD, p300-KD, and CARM1-KD cells into nude mice to establish tumor xenografts. Our results indicated that the tumors volumes in mice injected with c-Myc-KD, p300-KD, and CARM1-KD cells were similar and they were much smaller than tumors from mice injected with HT29 cells ( Figure 8A). To determine the in vivo effects of CPCM component inhibitors, we primarily injected mice with HT-29 cells and then these mice were randomly assigned to four groups (n = 5 in each group). These four-group mice were subsequently injected with PBS, sAJM, CARM-IN-1 or C646 every five days, respectively. The measurement of tumor volumes indicated that these inhibitors significantly inhibited tumor growth in vivo, and there was no significant difference in mice injected different CPCM component inhibitors ( Figure 8B). In addition, we also examined in vivo protein level changes in tumors derived from different group mice. As shown in Figure 8C, knockdown of CPCM components caused the decrease of CUL4A/4B protein levels but the increase of ST7 ( Figure 8C). Similarly, we also observed CPCM inhibitor treatments resulted in the deduction of CUL4A/4B protein levels but the increase of ST7 ( Figure 8D). These results suggested that the impaired CPCM complex decreased CUL4A/4B protein levels, which led to the deduction of ST7 ubiquitination and resulted in its accumulation. The accumulated ST7 functioned as a tumor suppressor to inhibit tumor growth.

Discussion
Transcription factors regulate gene expression through coordinating with other proteins such as coactivators and corepressors, and recruiting RNA polymerase II [45]. Our recent publication found that c-Myc induced the expression of CUL4A/4B in CRC cells [27]. To explore the transcriptionally regulatory mechanism of c-Myc in the regulation of CUL4A/4B expression, we purified c-Myc-coupled complex and identified several interesting partners including Max, p300, and CARM1 in this study. We then determined the interactions of these proteins and revealed how these four proteins assembled to a CPCM complex. We also evaluated the effects of knockdown and inhibition of CPCM components on CUL4A/4B expression, ST7 ubiquitination, oncogenic phenotypes, and in vivo tumor growth. Our results supported a model in which the CPCM complex specifically bound to the promoters of CUL4A/4B and activated their expression. The amplified CUL4A/4B assembled two separate CRL4 E3 ligases with DDB1, RBX1 and DCAF4, thereby promoting the ubiquitination of ST7 and leading to its degradation. The degraded ST7 lost its role in preventing tumor cell growth and resulted in the occurrence of CRC ( Figure  9).   In CRC cells, c-Myc is amplified and it dimerizes with Max. The c-Myc-Max heterodimer recruits p300 and CARM1 to assemble the CPCM complex, which specifically binds to the promoters of CUL4A/4B to activate their expression. The overexpressed CUL4A/4B recruit RBX1, DDB1 and DCAF4 to assemble two independent CRL4A/4B DCAF4 E3 ligases, which ubiquitinate a tumor suppressor ST7 and cause its degradation, leading to tumorigenesis.