Human cancer cell lines HCT116 (colorectal adenocarcinoma) and H1299 (non-small cell lung cancer) were purchased from American Type Cell Collection (Rockville, MD, USA). HCT116 p53^+/+ and HCT116 p53^−/− cells were gifts from Dr B. Vogelstein (Johns Hopkins University, Baltimore, MD, USA). Human embryonic kidney cells (HEK293T: a subclone of HEK293 cells engineered to express the SV40 large T antigen) were purchased from Riken Cell Bank (Ibaraki, Japan). Cells were incubated in a 37°C incubator containing 5% CO₂. HEK293T cells were transfected with Fugene6 (Promega, Madison, WI, USA). small interfering RNA oligonucleotides were purchased from Sigma Genosis (Woollands, TX, USA) and transfected with Lipofectamine RNAiMAX reagent (Invitrogen). We generated and purified replication-deficient recombinant viruses expressing p53 (Ad-p53) or LacZ (Ad-LacZ), as previously described (13). H1299 (p53-null) cells were infected with viral solutions at various multiplicities of infection (MOIs) and incubated at 37°C until being harvested. HCT116 cells were treated with 2 µg/ml ADR for 2 h to induce genotoxic stress.

HCT116 p53^+/+ or p53^−/− cells were harvested at 12, 24, 48, or 72 h after adriamycin (ADR) treatment or no treatment as control. Cells were lysed in buffer [8 M urea, 50 mM HEPES-NaOH, pH 8.0], and proteins were reduced with 10 mM Tris (2-carboxyethyl) phosphine (Sigma-Aldrich, St. Louis, MO, USA) at 37°C for 30 min, then subjected to alkylation with 50 mM iodoacetamide (Sigma-Aldrich) at 25°C in the dark for 45 min. The proteins were then digested with immobilized trypsin (Thermo Fisher Scientific, Bremen, Germany) at 37°C for 6 h. The resulting peptides were desalted with solid phase extraction (Oasis HLB µ-elution plate, Waters Corp., Milford, MA, USA) and analyzed by mass spectrometer (Linear Trap Quadropole Orbitrap Velos mass spectrometry, Thermo Fisher Scientific) combined with liquid chromatography (UltiMate 3000 RSLC nano-flow HPLC system, DIONEX Corporation, Sunnyvale, CA, USA). The LC/MS spectra were searched against the Homo sapiens protein sequence database in SwissProt using proteome analysis software (Proteome Discoverer 1.4 software, Thermo Fisher Scientific); a false discovery rate of 1% was set for both peptide and protein identification filters. Differential peptide quantification analysis (label-free quantification analysis) for 10 samples was performed on the Expressionist server platform (Genedata AG, Basel, Switzerland), as previously described (14). The fold change after ADR treatment was calculated with the following formula: f1 = Median peak intensity in [(p53^+/+ with ADR at 12, 24, 48, 72 h) / (Maximum peak intensity in (p53^+/+ and p53^−/− without ADR and p53^−/− with ADR at 12, 24, 48, 72 h) +1].

For the proteome analysis, we analyzed 47,534 peaks (derived from 22,276 genes) on the basis of the following criteria: fold induction >2 and p<0.05. Candidates demonstrating at least one peptide satisfying the criteria were included.

Gene expression analysis was carried out using microarray kit (Sure print G3 Human GE 8×60 K microarray, Agilent Technology, Santa Clara, CA, USA) according to the manufacturer's protocol. Briefly, HCT116 p53^+/+ and p53^−/− cells were treated with 2 µg/ml ADR and incubated at 37°C until being harvested. Total RNA was isolated from the cells through standard protocols. Each sample was labeled and hybridized to array slides.

The fold change after ADR treatment was calculated with the following formula: f2 = Median expression of probe in (p53^+/+ with ADR at 12, 24, 48 h)/Maximum expression of probe in (p53^+/+ and p53^−/− without ADR and p53^−/− with ADR at 12, 24, 48 h).

In the transcriptome analysis, we filtered 62,976 peaks (derived from 24,220 genes) on the basis of the following criteria: fold induction >2 and p<0.05. Genes demonstrating at least one probe satisfying the criteria were included.

MICALL1 expression and p53 mutation status in colorectal tumor samples were obtained from the Cancer Genome Atlas (TCGA) project by using cBioPortal. The expression levels of four sample categories, normal tissues (n=41), tumor tissues (n=389), tumor tissues with wild-type p53 (n=149), and tumor tissues with mutant p53 (n=114), were compared using Mann-Whitney U tests. We filtered candidates on the basis of the following criteria: i) a median expression level of normal tissues higher than the median expression level of tumor tissues (p<0.05); ii) a median expression level of p53 wild-type tumor tissues higher than the median expression level of p53 mutant tumor tissues (p<0.05). All analyses were performed with the EZR program (15).

cDNA fragments of MICALL1 or CD2AP amplified with KOD-plus DNA polymerase (Toyobo, Osaka, Japan) were cloned into the pCAGGS expression vector. The primers used for amplification are shown in Table I.

Total RNA from human cell lines was extracted with RNeasy Plus Mini kit (Qiagen, Valencia, CA, USA) according to the manufacturer's protocol. Complementary DNAs were synthesized with superscript III reverse transcriptase (Invitrogen). Quantitative real-time PCR was performed with SYBR Green Master Mix and a Light Cycler 96 (Roche, Basel, Switzerland). Primer sequences are shown in Table I. The PCR cycling condition were as follows: initial melting at 95°C for 5 sec, followed by 45 cycles at 55°C for 10 sec and 72°C for 10 sec. Analysis of the melting curve for the primers was conducted to confirm the specificity of the PCR product.

Cells were harvested and lysed with RIPA buffer containing 1 mM PMSF, 0.1 mM DTT, and 0.1% Calbiochem Protease Inhibitor Cocktail set III, EDTA-free (Merck Millipore, Darmstadt, Germany). The lysed samples were sonicated with a 30 sec on/30 sec off cycle for 15 min with a Bioruptor UCD-200 (Cosmobio, Tokyo, Japan). The cell lysates were centrifuged at 14,000 × g for 15 min and boiled after addition of SDS sample buffer (Bio-Rad, Hercules, CA, USA). Each sample was separated by SDS-PAGE and transferred to nitrocellulose membranes that were blocked using 5% skim milk (198-1-605, wako Pure Chemical Industries, Tokyo, Japan). Membranes were incubated at 4°C overnight with a mouse monoclonal anti-MICALL1 antibody (1:100, sc-398397, Santa Cruz Biotechnology, Santa Cruz, CA, USA), a rabbit monoclonal anti-CD2AP antibody (1:100, sc-9137, Santa Cruz Biotechnology), a rabbit monoclonal anti-FLAG antibody (1:2000, F7425, Sigma-Aldrich), a rat monoclonal anti-HA antibody (1:2000, 3F10, Sigma-Aldrich), a mouse monoclonal anti-p53 antibody (1:1000, DO-1, Santa Cruz Biotechnology), a mouse monoclonal anti-p21 antibody (1;100, OP64, Merck Millipore), or a mouse monoclonal anti-β-actin antibody (1:1000, AC-15, Abcam, Cambridge, MA, USA). Membranes were washed and probed with the secondary antibody conjugated to horseradish peroxidase against mouse, rabbit, or rat (all 1:30000, Santa Cruz Biotechnology, ref sc-2005, sc-2004 and sc-2006, respectively). Finally, proteins were visualized by chemiluminescent detection (Amersham ECL Western Blotting Detection Reagent (GE Health Care, Freiburg, Germany).

Protein-protein interactions were investigated by co-immunoprecipitation experiments. HA-tagged MICALL1 proteins and FLAG-tagged CD2AP proteins were expressed in HEK293T cells with Fugene6 (E2692, 30 µl per dish, Promega). The cells were harvested after 36 h of transfection and lysed for one hour with 0.5% NP40 buffer containing 50 mM Tris pH 7.4, 150 mM NaCl and protease inhibitor cocktail. Whole cell extracts were subjected to immunoprecipitation overnight with anti-Flag (Sigma-Aldrich) and anti-HA (Sigma-Aldrich) antibodies cross-linked to agarose beads at 4°C. Endogenous protein interactions in HCT116 p53^+/+ cells were also assessed by immunoprecipitation experiments using an anti-MICALL1 antibody (Santa Cruz Biotechnology). HCT116 p53^+/+ cells were incubated in 100-mm dishes and treated with 2 µg/ml ADR. After incubation for 48 h, the cells were harvested and lysed for one hour with the same buffer cocktail described above. The samples were incubated overnight with a mouse anti-MICALL1 antibody at 4°C. The exogenous and endogenous immunoprecipitates were separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotted using antibodies against MICALL1 and CD2AP.

A DNA fragment including two potential p53 binding sequences of MICALL1 was amplified and cloned into the pGL4.24 vector (Promega). For mutagenesis, point mutations were introduced with site-directed mutagenesis at the 4th and 14th nucleotides (C to T mutations) and the 7th and 17th nucleotides (G to T mutations) within the consensus p53 binding sequence (16). Reporter assays were performed using a Dual Luciferase Assay system (pGL3-promoter) (Promega) or Dual-Glo Luciferase Assay System (pGL4.24) (Promega), as previously described (17). The sequences of the primers used for amplification and site-directed mutagenesis are shown in Table I. Cells were not treated with Adriamycin in this experiment.

We carried out a Chromatin immunoprecipitation (ChIP) assay with an EZ-Magna ChIP G Chromatin Immunoprecipitation kit (Merck Millipore), per the manufacturer's protocol. H1299 cells (1×10⁶ cells per samples) infected with Ad-p53 or Ad-LacZ at a MOI of 10 were cross-linked with 1% form-aldehyde for 10 min, washed with PBS, and lysed using nuclear lysis buffer. The sample lysates were sonicated with a Bioruptor UCD-200 (CosmoBio), and DNA was sheared to 200–1000 base pair. Each sample was immunoprecipitated with an anti-p53 antibody (OP-40; Merck Millipore) or normal mouse IgG (sc-2025; Santa Cruz Biotechnology). Column-purified DNA was quantified by qPCR. Cells were not treated with Adriamycin in this experiment.

HCT116 p53^−/− or p53^+/+ cells were grown on cover glasses, transfected with Lipofectamine RNAiMAX (Thermo Fisher Scientific) for 36 h and fixed with 4% paraformaldehyde (163-20145-500 ml, Wako) for 10 min. The fixed cells were permeabilized with 0.2% Triton X100 in PBS (both from Sigma-Aldrich, ref X100-6X500ML and P4417-100TAB, respectively) for 5 min. Cells were blocked with blocking buffer [0.2% Triton X-100 and 3% BSA (bovine serum albumin, 001-000-162, Jackson Immunoresearch Laboratories, Inc., West Grove, PA, USA), in PBS 1X] for 1 h. After that, cells were incubated with mouse monoclonal anti-MICALL1 antibody (1:50, sc-398397, Santa Cruz Biotechnology) and rabbit monoclonal anti-CD2AP antibody (1:50, sc-9137, Santa Cruz Biotechnology) or rabbit monoclonal anti-RAB8A (1:200, #6975, Cell Signaling Technology) in staining solution (0.2% Triton X-100 and 3% BSA in PBS) for 2 h at room temperature. After washes with PBS, cells were incubated with a goat anti-mouse-Alexa Fluor 594- and a goat anti-rabbit-Alexa Fluor 488-labeled secondary antibodies (both 1:2000, from Thermo Fisher Scientific, ref A-11008 and A-11005), prepared in staining solution for 1 h at room temperature. Nuclei were stained with DAPI (H1200, Vector Laboratories, Youngstown, OH, USA) for 25 min and visualized with FluoView FV1000 confocal microscope (Olympus, Tokyo, Japan).

HCT116 cells were plated at 1×10⁴ cells per 35-mm culture well one day before transfection with MICALL1. On the second day after transfection, the culture medium was replaced by one containing 0.5 mg/ml G418 (G-418 Sulfate, 070-05183, Wako) every 2 days. Surviving colonies were fixed with methanol and stained with 1% crystal violet, and visible colonies were counted.

To identify novel p53 targets that play important roles in colorectal carcinogenesis, we previously performed transcriptome and proteome analyses using HCT116 p53^+/+ and p53^−/− cells (18,19). After treatment with 2 µg/ml ADR for 2 h, cells were harvested at different time points; i.e., 12, 24, 48 and 72 h (72 h only for samples used for mass spectrometry). Cells without ADR treatment were used as controls.

We identified 124 candidates on the basis of the proteome analysis at f1 >2 (p<0.05, Table II) and 523 candidates on the basis of the transcriptome analysis at f2 >2 (p<0.05, Table III). Ultimately, we identified 28 genes through mass spectrometry and microarray screening.

We then further screened these 28 genes by using a TCGA dataset consisting of 624 colorectal tumor samples. We selected genes as follows; i) significantly (p<0.05) higher expression in normal colon tissues than in tumor tissues and ii) significantly (p<0.05) higher expression in tumor tissues with wild-type p53 than in tissues with mutant p53 (Fig. 1).

As a result, two novel p53 targets (MICALL1 and APOBEC3C) and three known p53 targets (CDKN1A, FAS, TP53I3) were identified in our screening (Fig. 1B). Among the two novel candidates, induction of MICALL1 was validated by both qPCR and western blotting (Fig. 2A and B). Therefore, we selected MICALL1 for further analysis. We also found that MICALL1 was induced in H1299 cells infected by Ad-p53 but not in Ad-LacZ-infected cells (Fig. 2C).

To evaluate the effect of p53 activation on MICALL1 expression, we analyzed the expression of MICALL1 using HCT116 cells treated with Nutlin-3a (inhibitor of MDM2) that activates p53. As shown in Fig. 2D, MICALL1 expression was induced by Nutrin-3a only in HCT116 p53^+/+ cells. These results clearly indicate that MICALL1 is induced by p53 activation.

We then performed immunocytochemical analysis using HCT116 p53^+/+ and p53^−/− cells under ADR treatment (Fig. 3). Without ADR treatment, expression of MICALL1 was very low in both types of cells. However, in response to ADR treatment, MICALL1 was induced at the cytoplasmic tubular structure only in HCT116 p53^+/+ cells.

To investigate whether MICALL1 is a direct target of p53, we searched for the p53 binding motif (16) within the MICALL1 locus and found two potential binding sequences (p53BS1, p53BS2) in the approximately 3000 base pair of 5′ flanking sequence (Fig. 4A). A 151-base pair DNA fragment (p53BS1+2) including two p53 binding sequences was amplified and cloned upstream of the minimal promoter in the pGL4.24 vector (pGL4.24/p53BS1+2). The reporter assays showed increased luciferase activity in H1299 cells transfected with pGL4.24/p53Bs1+2 in the presence of a plasmid expressing wild-type p53 (Fig. 4B). However, base substitutions in p53BS (pGL4.24/p53BSmut1, pGL4.24/p53BSmut2) decreased the observed enhanced luciferase activity.

To verify whether p53 directly bound to p53BS, we performed a ChIP assay using H1299 cells infected with either Ad-p53 or Ad-LacZ. After precipitation with an anti-p53 antibody, the DNA fragment containing p53BS1 was quantified by qPCR, which showed that p53 specifically bound to p53BS1 in cells infected with Ad-p53 (Fig. 4C). Thus, we concluded that p53 regulates MICALL1 expression through p53BS in the 5′ flanking region of the MICALL1 gene.

To further investigate the role of MICALL1 as a p53 downstream target, we screened MICALL1-interacting proteins by LC-MS analysis. MICALL1 was immunoprecipitated from cell lysates of HEK293T cells transfected with the MICALL1 expression plasmid. The protein complex including immunopurified MICALL1 was analyzed by SDS-PAGE and subsequent silver staining. A protein band at approximately 75 kDa, which was abundant in the protein complex including MICALL1, was subjected to LC-MS analysis. The results indicated that CD2AP is likely to bind MICALL1 (Fig. 5A and B). Interaction between CD2AP and MICALL1 was confirmed by western blotting (Fig. 5C). We also confirmed the interaction by using lysates from HEK293T cells overexpressing FLAG-MICALL1 and HA-CD2AP. Moreover, this interaction between endogenous MICALL1 protein and CD2AP was observed in ADR-treated HCT116 cells (Fig. 5D).

MICALL1 has been reported to be a marker of TREs, which are essential for the recycling of receptors and lipids to the plasma membrane (20). MICALL1 has also been reported to recruit RAB8A, a protein related to vesicle-mediated transport, to TREs (21). Since MICALL1 was found at tubular structure in the cytoplasm after DNA damage, we investigated the subcellular localization of CD2AP and RAB8A in HCT116 p53^+/+ cells (Fig. 6A and B).

Initially, we determined the localization of CD2AP and RAB8A after DNA damage. HCT116 p53^+/+ cells were grown on coverslips and treated with 2 µg/ml ADR; 48 h later, CD2AP and RAB8A exhibited a tubular-like distribution with MICALL1 (Fig. 6A-1, 2, and B-1, 2). Thus, CD2AP and RAB8A exhibited altered cytoplasmic localization to TREs after DNA damage.

Next, we tested whether p53 or MICALL1 knockdown changed the localization of CD2AP and RAB8A after DNA damage (Fig. 6C). At 48 h after ADR treatment, cells were fixed, and MICALL1 and CD2AP or RAB8A was identified with antibodies. According to the results, their localization change did not occur in p53 or MICALL1-knockdown cells (Fig. 6A-4, 5, 6, and B-4, 5, 6). According to the above results, MICALL1 is involved in TRE regulation in the response to DNA damage.

To investigate the role of MICALL1 in cell proliferation, we performed a colony formation assay using HCT116 cells and found a significant decrease in colony counts for cells transfected with MICALL1 compared with cells mock transfected with the vector (Fig. 6D). Furthermore, ectopic expression of CD2AP also repressed cell growth like MICALL1 expressing cells (Fig. 6D). These results indicated that MICALL1 would suppress tumor cell growth through the regulation of CD2AP.