The 5'-untranslated region of p16INK4a melanoma tumor suppressor acts as a cellular IRES, controlling mRNA translation under hypoxia through YBX1 binding.

CDKN2A/p16INK4a is an essential tumor suppressor gene that controls cell cycle progression and replicative senescence. It is also the main melanoma susceptibility gene. Here we report that p16INK4a 5'UTR mRNA acts as a cellular Internal Ribosome Entry Site (IRES). The potential for p16INK4a 5'UTR to drive cap-independent translation was evaluated by dual-luciferase assays using bicistronic and monocistronic vectors. Results of reporters' relative activities coupled to control analyses for actual bicistronic mRNA transcription, indicated that the wild type p16INK4a 5'UTR could stimulate cap-independent translation. Notably, hypoxic stress and the treatment with mTOR inhibitors enhanced the translation-stimulating property of p16INK4a 5'UTR. RNA immunoprecipitation performed in melanoma-derived SK-Mel-28 and in a patient-derived lymphoblastoid cell line indicated that YBX1 can bind the wild type p16INK4a mRNA increasing its translation efficiency, particularly during hypoxic stress. Modulation of YBX1 expression further supported its involvement in cap-independent translation of the wild type p16INK4a but not a c.-42T>A variant. RNA SHAPE assays revealed local flexibility changes for the c.-42T>A variant at the predicted YBX1 binding site region. Our results indicate that p16INK4a 5'UTR contains a cellular IRES that can enhance mRNA translation efficiency, in part through YBX1.


Bicistronic reporter plasmid
pRuF bicistronic vector containing the whole p16 INK4a 5′UTR was generated as previously described [1]. Exploiting EcoRI and NdeI restriction sites within pRuF-empty backbone and PCR primer tails, p16 INK4a 5′UTR fragments were cloned using T4 Ligase enzyme (New England BioLabs, EuroClone, Milan, Italy) and XL1Blue bacterial cells. As positive control, the c-MYC 5′UTR region containing the previously identified IRES site was PCR amplified from MCF7 RNA using the primers pair described in previous studies [2] modified to include tails containing the EcoRI and NdeI restriction sites, and cloned in pRuF reporter vector as described above. With a similar approach two negative controls were developed inserting the 5′UTR of β-globin and β-actin [3]. As an additional negative control, the reverse complement p16 INK4a 5′UTR sequence (named p16INV) was cloned in pRuF plasmid as negative control and we also generated two deletion constructs, denominated Redux 180 and Redux 90, where the proximal 90 nucleotides and the distal 180 nucleotides of the p16 INK4a 5′UTR were deleted respectively. The correct insertions of the desired fragments were confirmed by EcoRI and NdeI endonucleases double digestion and direct sequencing (BMR Genomics, Padua, Italy). pCMV6-Entry Myc-DDK tagged vector to overexpress YBX1 was purchased from Origene (Tema Ricerca, Bologna, Italy). Monocistronic vectors based on the pGL3-promoter plasmid, containing the c-MYC, p16 wild-type or p16 c.-42T>A variant 5′UTR upstream of the Firefly luciferase were constructed as described in [1]. The pRL-SV40 plasmid was used as a control for transfection together with the monocistronic constructs. Transfection grade plasmids were obtained using the Pure Yield Midiprep system according to the manufacturer's instructions (Promega, Milan, Italy).

Dual luciferase assays
Luciferase assays with the cloned pRuF vectors were performed in 24-well plates with MCF7, HCT116, MCF7vector or MCF7shp53 cells, as previously described [1]. Briefly, cells were transiently transfected with Fugene HD transfection reagent (Promega) and 400 ng of pRuF constructs or 350 ng of pGL3-promoter + 50 ng of pRL-SV40 plasmids. When appropriate, 100ng of the pCMV6-empty or pCMV6-YBX1 expression vector were co-transfected along with the different pRuF reporter vectors. Twenty-four hours after transfection cells where lysed with Passive Lysis Buffer (PLB) 1X and Firefly and Renilla luciferase activity was measured with the Infinite M200 multi-plate reader (Tecan, Milan, Italy). In the experiments that needed cap-dependent inhibition, 8 hours after transfection MCF7 cells were treated with 50 nM mTOR inhibitors or cultured in hypoxia for additional 16 hours prior to the luciferase assay.

High-content imaging analysis of global protein translation
To investigate the impact of treatments with mTOR inhibitors or hypoxia on global protein sysntesis, SK-Mel-28 cells were exposed to 50 nM Rapamycin or 50 nM Torin1 or cultured in hypoxic chamber for a total of 16 hours. 3 hours prior to sample analysis, the culture medium was removed and replaced with fresh, methionine free medium supplemented with an amino acid analog of methionine containing an alkyne moiety, following the procedure of the Click-iT ® HPG Alexa Fluor ® 488 Protein Synthesis Assay Kit (Molecular Probes, Invitrogen, Life Technologies). Relative quantification of protein synthesis was performed by acquiring images by the PerkinElmer Operetta ® High Content Imaging System, using DAPI staining to visualize nuclei. Images were analyzed and quantified using Columbus software.

RNA extraction and qPCR on total RNA
To extract RNA, cells were washed once with phosphate buffer saline (PBS). Total RNA was extracted using the RNAeasy mini Kit (Qiagen, Milan, Italy) according to the manufacturer's instructions. In-column DNAse treatment (Qiagen) was performed to remove DNA contamination during total RNA extraction. Purity and concentration of RNA was evaluated using a NanoDrop ND1000 spectrophotometer. cDNA was generated starting from 1 μg of RNA by using the RevertAid TM First Strand cDNA Synthesis Kit (ThermoFisher, Milan, Italy). Quantitative Real-time PCR (q-PCR) was performed using a cDNA aliquot equivalent to 25 ng of converted RNA using a CFX96 qPCR Thermal cycler (BioRad, Milan, Italy) and the 2X KAPA SYBRGreen FAST qPCR Master Mix (Kapa Biosystems, Resnova, Ancona, Italy). In the transiently transfected MCF7 cells the regions spanning the Renilla and the Firefly luciferase mRNA and the portion in between the two reporter genes were amplified using specific primers (Eurofins MWG Operon, Ebersberg, Germany -all primers' sequences utilized in this work are available upon request). The relative mRNA abundance was quantified as explained in a previous study [1]. Endogenous p16 INK4a , c-MYC and p53 relative mRNA levels were measured from SK-Mel-28 treated cells. The relative quantitation of endogenous mRNAs from treated SK-Mel-28 was obtained using the comparative Ct method (∆∆Ct), taking into account the efficiency of cDNA synthesis by the quantification of the glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and β-2microglobulin (B2M) reference genes.

RNA extraction and qPCR on subpolysomal and polysomal RNA fractions
The protocol for sucrose gradient fractionation was previously described [1]. Fractions corresponding to subpolysomal peaks were pooled together and the same was done for peaks corresponding to polysomal fractions (see Figure 3B for example of profiles). Prior to RNA extraction by organic solvent and salt/ethanol precipitation 15 ng of spike-in Luciferase control RNA (Promega) was added to both pools. RNA was resupended in DEPC water, quantified at nanodrop and processed for qPCR analysis as described above. The sequences of all primers were taken from previous publications or designed using primer blast and are available upon request.

Ribonucleoprotein ImmunoPrecipitation (RIP) assays
RIP was performed as previously described [7]. Briefly, 12 x 10 6 cells were lysed in 500 μl of NT2 buffer (50 mM Tris-HCl pH7.7, 150 mM NaCl, 1 mM MgCl2; 0.05%NP40, 1U/ul RNase IN, 20 mM DTT, 1% BSA, Protease inhibitor cocktail) pre-chilled at 4°C and syringed using an U100 insulin needle. Lysates were centrifuged at 10000g for 10 minutes then the supernatant was pre-cleared by interaction with protein-A-coated agarose beads (equilibrated in NT2 buffer) overnight at 4°C in constant shaking (100 μl slurry beads / 500 μl lysate). 150 μl of the pre-cleared lysate were mixed with protein G coated agarose beads with a specific primary antibody (e.g. YBX1, Ab 59-Q, Santa Cruz) (or control IgG) conjugated for 6 hours at 4°C then washed twice in NT2 buffer. 20 μl Protein-G-coated slurry agarose beads were conjugated with 4 μg antibody at room temperature for 2 hours, washed and equilibrated in NT2 lysis buffer before use. RNA was isolated from the different samples (immunoprecipitated anti-YBX1, IgG and pre-cleared input) by TriZol, as recommended by the manufacturer, retrotranscribed into cDNA by RevertAid TM First Strand cDNA Synthesis Kit (ThermoFisher) and used as template for PCR or for a q-PCR analysis that was conducted as described above.

In Vitro Transcription and Acylation of RNA
RNA was transcribed from amplified inserts using T7 Megascript kit from Ambion, following manufacturer's protocol. In a typical in vitro modification protocol, RNA was heated in metal-free water for two minutes at 95°C. The RNA was then flash-cooled on ice. The RNA 3X SHAPE buffer (333 mM HEPES, pH 8.0, 20 mM MgCl 2 , 333 mM NaCl) was added and the RNA was allowed to equilibrate at 37°C for ten minutes. To this mixture, 1 μL of 10X 2-methylnicotinic acid imidazolide electrophile (NAI) stock in DMSO or DMSO only was added. The reaction was permitted to continue until the desired time. Reactions were extracted once with acid phenol:chloroform (pH 4.5 ± 0.2) and twice with chloroform. RNA was precipitated with 40 μL of 3M sodium acetate buffer (pH 5.2) and 1 μL of glycogen (20 μg/μL). Pellets were washed twice with 70% ethanol and resuspended in 10 μL RNase-free water.

Reverse Transcription of modified RNA
32 P-end-labeled DNA primer (reverse primer above) was annealed to 3 μg of total RNA by incubating at 95°C for two minutes followed by a step-down cooling (2 deg/sec) to 4°C. To the reaction, first-strand buffer, DTT and dNTPs were added, pre-incubated at 52 °C for one minute, followed by Superscript III (2units/μL final concentration) addition. Extensions were performed for ten minutes. To the reaction, 1 μL of 4M NaOH was added and allowed to react for 5 minutes at 95°C. 10 μL of Gel Loading Buffer II (Ambion, Inc.) was then added, and cDNA extensions were resolved on 8% denaturing (7 M Urea) polyacrylamide gels (29:1 acrylamide:bisacrylamide, 1X TBE). Sequencing lanes are from DMSO control treated cells.

Characterization of reverse transcription stops
cDNA extensions were visualized by phosphorimaging (STORM, Molecular Dynamics). cDNA bands were integrated with SAFA [8] SHAPE reactivities were normalized to a scale spanning 0 to 1.5, where 1.0 is defined as the mean intensity of highly reactive nucleotides [9]. RNA secondary structures were predicted using RNA structure software, incorporating the SHAPE structure probing as a restraint [10].