CRISPR screening identifies M1AP as a new MYC regulator with a promoter-reporter system

MYC promoter-reporter system and lentivirus plasmids

The Polymerase chain reaction (PCR) primers were designed across the proximal MYC promoter at −3.1 kb from the MYC transcriptional start site (TSS), and PCR was performed with KOD-FX (Toyobo). The primers were 5′-TTAGCTAGCGAGGGTT- TTCTTTGAGGGGC-3′ (forward) and 5′-AACGGATCCCGGAGATTAGCGAGAGAGGA-3′ (reverse).

A plasmid of Dendra2 (pDendra2) was purchased from Takara-Clontech. The cytomegalovirus immediate early promoter (PCMVIE) was removed with restriction enzymes (Afl III and Nhe I), and the multi-cloning site was replaced with Nhe I and BamH I. Amplified fragments of −3.1 kb of the MYC promoter region with Nhe I and BamH I restriction enzyme sites were integrated into pDendra2.The new reporter system with Dendra2 and the −3.1 kb MYC promoter region (pMYC-promoter-Dendra2) was used for further experiments. The double-stranded DNA fragments of M1AP with attB sites were purchased from gBlocks^® (IDT) and generated into the pENTR221 vector using Gateway^® BP clonase. Then they were transferred to the TripZ-TRE-DEST-IRES-GFP-EF1A-rtTA vector (Kurata et al., 2016) using Gateway^® LR clonase.

Cell culture and SAM library screening

Human embryonic kidney cells (HEK 293T) were kindly provided by Dr. David Laraespada, Masonic Cancer Center, University of Minnesota, MN, USA. and maintained in Dulbecco’s modified Eagle’s medium (DMEM, Fuji Film) with 10% fetal bovine serum and penicillin/streptomycin with 5% CO₂ at 37 °C . To transduce the lentivirus, HEK 293T cells (5 × 10⁵ cells/well) were seeded in a six-well cell culture plate one day before the transfection. The next day, 3 µg of lentivirus plasmid was transfected with 1 µg of pMDG and 2 µg of pCMV using Lipofectamine™ 3000 Reagent (Invitrogen) according to previous work (Kurata et al., 2016). Twelve hours after transfection, the medium was changed to fresh DMEM. The virus supernatant was harvested at 48 h after transfection, then filtered with Millex™ -HP 0.45 µm (Millipore). The target HEK 293T cells (5 × 10⁵ cells/well) were seeded in a six-well plate one day prior to transduction and were transduced with this lentivirus supernatant with 5 µg/ml polybrene (Sigma). The used lentiviral plasmids were lentiMPHv2 (Addgene, #89308), lentidCAS-VP64_blast (Addgene, #61425), and lentiSAMv2 (Addgene, #61597). At first, lentiMPHv2 was transduced into HEK 293T and treated with Hygromycin B for 2 weeks. Then, lentidCAS9-VP64 was induced in MPH-expressing HEK 293T cells and treated with Blasticidin for 2 weeks. The MPH-dCas9-VP64-expressing HEK 293T applied to further experiments. The control and MYC activation gRNA (Fig. S1C) were induced in MPH-dCas9-VP64-expressing HEK 293T cells. The CRISPR activation library, lentiSAMv2, was also transduced into MPH-dCas9-VP64-expressing HEK 293T cells and then treated with Zeocin (300 µg/ml, Invivogen) for 2 weeks.

Collection of Dendra2-positive cells and identification of gRNAs

After 2 weeks of selection with Zeocin, pMYC-promoter-Dendra2 was transfected into SAM library lentivirus-transduced cells with Lipofectamine™ 3000 Reagent (Invitrogen). HEK 293T cells transfected with pMYC-promoter-Dendra2 were evaluated for Dendra2-positive cells by fluorescence microscope (Keyence). Cells were collected at 72 h and washed with phosphate-buffered saline (PBS) and applied to the cell sorter. The Dendra2-positive cells were collected into a 96-well plate with MoFlo XDP (Beckman Coulter).

A total of 169 cells were collected, and 12 clones were successfully established. Collected clones were centrifuged and washed with PBS solution and pelletized. Lysis buffer solution with proteinase K (20 mg/ml) was added to cells and left for 30 min at 60 °C. After treatment at 94 °C for 10 min, the lysate was centrifuged for 15 min, and the supernatant was collected as genome DNA (gDNA) containing fractions. Using extracted gDNA of each collected clone, PCR was conducted with GoTaq^® qPCR Master Mix (Promega) with the following primers: 5′-GAGGGCCTATTTCCCATGAT -3′ (forward) and 5′-GAGGGCCTATTTCCCATGAT-3′ (reverse). PCR product containing the gRNA region was employed for TA cloning with pGEM^®-T Vector Systems (Promega) and gRNAs were sequenced by Sanger-sequencing.

Preparation of CRISPR backbone plasmid with target gRNAs and transcriptional activation for MYC using the CRISPR activation system

The oligonucleotides for each target gRNA sequence were designed using the list from the SAM library and integrated into the pENTR221/U6/stuffer/MS2 vector (Kurata et al., 2018a). The MPH-dCas9-VP64-expressing HEK 293T cells (1 × 10⁵ cells/ml) were seeded in a 24-well cell culture plate. The next day, 2.5 µg of individual gRNA vector and pMYC- promoter-Dendra2 were transfected with Lipofectamine ™ 3000 Reagent. At 48 h after the transfection, the MYC transcriptional activation was evaluated by Dendra2-positivity by Keyence.

Evaluation of MYC expression

A pTRIPZ-M1AP-GFP was transduced into the HEK 293T cells. These transduced cells were selected with puromycin (1 µg/ml, Invitrogen) for 2 weeks. Seventy-two hours after the induction of Doxycycline (DOX) (1 µg/ml), the collected cells were employed for RNA isolation.

RNA was isolated using an RNeasyR Mini Kit (Qiagen) according to the manufacturer’s instructions. Complementary DNA (cDNA) was generated from RNA with ReverTra AceR qPCR RT Master Mix (Toyobo). Beta-ACTIN was used as an endogenous control. Using the ABI Prism 7900HT (Applied Biosystems), quantitative PCR (qPCR) analysis was performed to quantify the RNA level using SYBR Mix. The sequences for the PCR primers used in gene expression were as follows: MYC : 5′-CGACTCTGAGGAGGAACAAGAA-3′ (forward) and 5′-CAGCAGAAGGTGATCCAGACT-3′ (reverse), β-ACTIN: 5′-CACAGAGCCTCGCCTTTGCC-3′ (forward) and 5′-CACAGAGCCTCGCCTTTGCC-3′ (reverse). The mRNA level of the targeted gene was analyzed by comparison with the standard calibration curve. Western blot analysis was performed according to standard procedures (Kurata et al., 2016). The primary antibodies used were anti-c-Myc (Abcam ab32072) and anti-GAPDH (Santa Cruz sc47724), each at a dilution of 1:1000. The primary antibodies were detected with horseradish peroxidase (HRP)-conjugated secondary antibodies at 1:50000. Protein bands were visualized with Clarity^TM Western ECL Substrate (Bio-Rad).

Dual-Luciferase reporter assay

The fragments of the MYC promoter was integrated into a pGL4 vector (Promega). A pT3.5-CAG-M1AP was generated with pENTR221-M1AP and pT3.5-CAG-DEST (Kurata et al., 2018a) using Gateway^® LR clonase. The luciferase reporter vectors, the pT3.5-CAG-M1AP vector, and the pRL Renilla luciferase reporter vector were co-transfected in the HEK 293T cells. The samples were harvested 48 and 72 h after the transfection. Each assay was biologically triplicated and repeated. Luciferase activities were measured using the dual-luciferase reporter assay system (Promega) and Lumat LB9507 (Perkin Elmer). RNA was harvested at the same time points and the expression of MYC mRNA was analyzed by qPCR.

M1AP and MYC expressions in databases

For the purpose of comparing gene expression between cancer and non-cancerous tissue, the GEPIA web server was employed (Tang et al., 2017) (http://gepia2.cancer-pku.cn/).

Datasets for AML samples were used in BloodSpot database, that database was used to retrieve gene expression data from curated human and murine microarray and RNA-sequencing (RNA-Seq) datasets (Bagger et al., 2016). GSE42519, GSE13159,  GSE15434,  GSE61804,  GSE14468, and The Cancer Genome Atlas (TCGA) datasets were used for analysis. In addition, following database were used; The Human Protein Atlas (https://www.proteinatlas.org/) and cBioPortal (https://www.cbioportal.org) using TCGA data (Cancer Genome Atlas Research Network et al., 2013; Gao et al., 2013; Cerami et al., 2012).

Statistical analysis

The data were statistically analyzed using the EZR version 1.36 software. A value of p < 0.05 was considered statistically significant for all analyses.

Development of the MYC promoter-reporter system for library screening

Using PCR primers designed across the proximal MYC promoter at −3.1 kb (−3134 to −40) from the MYC TSS, PCR was performed. The MYC promoter region was integrated upstream of Dendra2. The pDendra2 harboring the 3.1 kb MYC promoter region was called “pMYC-promoter-Dendra2” for further experiments (Fig. S1A). Plasmids of pMYC-promoter-Dendra2 and gRNA of MYC or HPRT (as a negative control) were transfected in HEK 293T cells using the CRISPR activation system. We confirmed that Dendra2 showed positivity only in the cells with gRNA for MYC  (Figs. S1B, S1C). Therefore, pMYC-promoter-Dendra2 was utilized as a MYC promoter-reporter analysis tool in this study.

MYC transcriptional activator screening Using the CRISPR activation library

To screen transcriptional activators working on the MYC promoter region, the CRISPR activation library was employed. The SAM library lentivirus was infected into MPH-dCas9-VP64-expressing HEK 293T cells (1 × 10⁷ cells). After selection, pMYC- promoter-Dendra2 was transfected into these HEK 293T cells. The multiplicity of infection (MOI) ranged from 1.4 to 5.2.

At 72 h after transfection, cells were proceeded to Dendra2-positive cell collection using flow cytometry (positive ratio: 0.017%). Sorted single cells were collected into a 96-well plate and incubated for 2 weeks. Among the collected 169 cells, 12 clones were successfully established (Fig. 1).

Second screening to identify candidate genes as MYC activators

For the second screening, pMYC-promoter-Dendra2 was again transfected into the 12 clones. Using fluorescence microscope, clones 3 and 5 were shown to be truly Dendra2 positive (Figs. 2A–2L), and the photoswitchable attribute of Dendra2 was useful for distinguishing it from cells with non-specific autofluorescence (Fig. S2). Using extracted gDNA of clones 3 and 5, PCR was performed to amplify the gRNA containing region, which was introduced by the CRISPR activation library. Then, the PCR product was employed for TA cloning, followed by colony PCR. By this method, four candidate genes—PLEKHO2 , MICU, MBTPS1, and M1AP were identified (Table 1).

Each candidate gRNA was transfected to the HEK 293T cells using the CRISPR activation system accompanied by pMYC-promoter-Dendra2. Because Dendra2 was positive only in the cells transfected with M1AP gRNA (Figs. 3A–3D), M1AP was considered as the most promising gene that would increase MYC transcriptional activity. Then, double-stranded DNA fragments coding the full length of M1AP induced into Doxycycline-inducible lentiviral vector (pTRIPZ-M1AP-GFP, Fig. 4A). The pTRIPZ-M1AP-GFP was introduced into HEK 293T cells. Seventy-two hours after the induction M1AP with DOX (1 µg/ml), high MYC expression was observed in DOX-induced cells by RT qPCR (t-test, p < 0.05, Fig. 4B). Furthermore, in the western blots, an increase of MYC expression was also observed in DOX-induced cells (Fig. 4C).

Additionally, a dual-luciferase assay conducted with a reporter containing the MYC promoter region showed a significant increase of promoter activity in M1AP-overexpressing cells at both 48 and 72 h (Figs. 4D and S3). On the other hand, the MYC mRNA, which was isolated at the same time points, showed mildly higher in M1AP-overexpressing cells at 48 h but not at 72 h (Figs. 4E and Fig. S3 ). Taken together, although M1AP was confirmed to activate the MYC promoter strongly, the MYC mRNA itself might be regulated also by other factors.

Correlation between MYC and M1AP in the database

According to the GEPIA web server (Tang et al., 2017) (http://gepia2.cancer-pku.cn/), high M1AP expression is observed in cervical carcinoma, acute myeloid leukemia (AML), ovarian carcinoma, thyroid carcinoma, uterine endometrial carcinoma, and uterine carcinosarcoma when compared to normal tissue (Fig. 4SA). Among these carcinomas, a correlation analysis using cBioPortal revealed a mild positive correlation between MYC and M1AP only in AML (Fig. 5A, Figs. S4B- S4F). Therefore, we focused on hematopoietic cells, and M1AP expression was analyzed using the BloodSpot database. As a result, M1AP was significantly higher in AML tissue compared to normal bone marrow tissue (Fig. 5B). Interestingly, M1AP was highly expressed in AML, especially with t (8;21) (Fig. 5C).