Insulin-like growth factor 2 mRNA-binding protein-1 (IMP1) has high affinity for KRAS mRNA, and it can regulate KRAS expression in cells. We first characterized the molecular interaction between IMP1 and KRAS mRNA. Using IMP1 variants with a point mutation in the GXXG motif at each KH domain, we showed that all KH domains play a critical role in the binding of KRAS RNA. We mapped the IMP1-binding sites on KRAS mRNA and show that IMP1 has the highest affinity for nts 1–185. Although it has lower affinity, IMP1 does bind to other coding regions and the 3′-UTR of KRAS mRNA. Eight antisense oligonucleotides (AONs) were designed against KRAS RNA in the nts 1–185 region, but only two, SM6 and SM7, show potent inhibition of the IMP1–KRAS RNA interaction in vitro. To test the activity of these two AONs in SW480 human colon cancer cells, we used 2′-O-methyl-modified versions of SM6 and SM7 in an attempt to down-regulate KRAS expression. To our surprise, both SM6 and SM7 had no effect on KRAS mRNA and protein expression, but significantly inhibited IMP1 protein expression without altering IMP1 mRNA level. On the other hand, knockdown of IMP1 using siRNA lowered the expression of KRAS. Using Renilla luciferase as a reporter, we found that IMP1 translation is significantly reduced in SM7-treated cells with no change in let-7a levels. The present study shows that the regulation of KRAS expression by IMP1 is complex and may involve both the IMP1 protein and its mRNA transcript.

Introduction

The insulin-like growth factor 2 mRNA-binding protein 1 (IGF2BP1), also known as IMP1 or CRD-BP (coding region determinant-binding protein), belongs to a small conserved family of RNA-binding proteins, which also include IMP2 and IMP3 [1,2]. Mammals have all three IMPs with two N-terminal RNA-recognition motifs (RRMs) followed by four C-terminal KH [hnRNP (heterogeneous nuclear ribonucleoprotein) K-homology] domains [1]. In vitro studies of IMP1 and its paralogues have all concluded that the KH domains, but not the two RRMs, are important for direct RNA binding [1,2]. Although our understanding on the physical interaction between IMP1 and its target RNA is incomplete, it does appear that the KH34 di-domain plays an essential role [35]. Site-directed mutagenesis studies have also shown the critical RNA-binding role played by the KH1 and KH2 domains [68].

IMP1 is overexpressed in various human cancers which include cancers of the breast, colon, brain, lung, testicular, skin, ovarian, chorion and liver [1,9]. The oncogenic role of IMP1 is shown in two animal cancer models. Mice genetically engineered to carry targeted expression of IMP1 in mammary glands developed mammary adenocarcinoma [10]. The role of IMP1 in the development of colorectal cancer was suggested by the following findings: (i) in colorectal cancer cell xenograft model, overexpression of IMP1 in the colorectal cancer cell lines promoted xenograft tumour growth and their dissemination into the blood [11], (ii) deletion of intestinal IMP1 significantly reduced the number of tumours in a mouse model specifically engineered to develop intestinal tumours and (iii) depletion of IMP1 in human colorectal cancer cell line xenografts resulted in reduced numbers of circulating tumour cells [11].

In contrast with its oncogenic role, there is a growing body of evidence for the tumour suppressive role of IMP1 under certain circumstances. For example, the loss of IMP1 led to leukaemia [12] and breast cancer cell proliferation and migration [13,14]. Three recent in vivo studies provided further evidence for the tumour suppressive role of IMP1 [4,15,16]. In a mouse xenograft model, IMP1 suppressed breast tumour formation and lung metastasis derived from the growth of implanted human MDA231 breast cancer cells [4]. In another study using mice conditionally expressing IMP1, the IMP1 transgene showed little or no metastases to the lungs from the primary tumour, which was in contrast with control mice not expressing IMP1 [15]. In a mouse model of colon cancer, loss of IMP1 in stromal cells led to a microenvironment that promoted tumourigenesis [16]. As postulated [15,16], it is likely that the tumour suppressive role of IMP1 is manifested only during metastasis as the consequence of loss of the protein in a specific cell type. While the exact mechanism whereby IMP1 promotes tumour growth and metastasis has not been elucidated, cumulative evidence suggests that its ability to physically associate with RNAs, either to stabilize mRNAs or to influence miRNA activity or both, is one important criterion. IMP1 was first discovered by its ability to physically associate with a specific coding region of c-myc mRNA to influence its stability in a cell-free system [17]. Subsequent studies show that IMP1 can regulate c-myc mRNA half-life in cells [1820]. Various studies have now showed such function of IMP1; the ability to bind and stabilize subsets of mRNAs for genes implicated in cancer. For instance, IMP1 binds to 3′-untranslated region (UTR) of CD44 mRNA and stabilizes the transcript, leading to cell adhesion, cytoplasmic spreading and invadopodia formation [21]. IMP1 also binds to βTrCP1 and GLI1 mRNAs and its overexpression leads to stabilization of those transcripts and proliferation of colorectal cancer cells [19,22]. Using the enhanced cross-linking and immuno-precipitation technology, the 3′UTRs of integrin ITGB5 and BCL2 mRNAs were recently identified as binding sites of IMP1 [23]. Depletion of IMP1 led to destabilization of ITGB5 mRNA, decreased expression of ITGB5 protein and decreased adherence in human embryonic stem cells [23]. As in human embryonic stem cells, IMP1 depletion leads to decreased BCL2 mRNA and protein expression as well as increased cell death [23].

The KRAS mRNA is another oncogenic target of IMP1. Using the method of immuno-precipitation coupled with UV cross-linking, it was shown that IMP1 has affinity for the coding region and the 3′-UTR of KRAS mRNA [24]. Additional experiments showed that overexpression of IMP1 led to increased KRAS expression and colon cancer cell proliferation, while loss of IMP1 led to the inhibition of KRAS expression [24]. In addition to revealing the mechanistic oncogenic function of IMP1, a better understanding of IMP1–KRAS mRNA interaction could lead to the development of a new therapeutic approach against KRAS. This is particularly important because there is currently no clinically approved medication against KRAS despite tremendous efforts [25].

In this study, we set out to understand the molecular interaction between IMP1 and KRAS RNA. We mapped 3′-UTR and coding region of KRAS mRNA to identify the smallest region that has affinity for IMP1. Using IMP1 KH variants with point mutation in the GXXG motif, we investigated the role of each KH domains critical for binding to KRAS RNA. We designed and assessed eight specific antisense oligonucleotides (AONs) against KRAS RNA and found two that are potent inhibitors of IMP1–KRAS RNA interaction. To our surprise, these 2′-O-methyl AONs inhibited IMP1 but not KRAS expression in cells. We further show that such inhibition of IMP1 expression is through control of the translation of its mRNA. Overall, we show that the control of KRAS expression by IMP1 is complex and may involve more than just protein–RNA interaction.

Materials and methods

Oligonucleotides and primers

Table 1 shows sequences of AONs (SM1 to SM8) used in the present study. For work in cells, a different set of AONs (SM2, SM6 and SM7) with RNA bases were synthesized as 2′-O-methyl derivatives with a phosphodiester backbone. Table 2 shows the sequences of primers used to amplify the DNA template for synthesizing all the KRAS RNA fragments described in Figure 1. Table 2 also shows the sequences of primers used to amplify fragments of 3′-UTR of IMP1 to generate the psiCHECK2-IMP1-3′-UTR-luciferase reporter plasmids. All oligonucleotides and primers were synthesized by Integrated DNA Technologies (IDT) Inc. (Coralville, Iowa).

Characterizing the IMP1–-KRAS RNA interaction.

Figure 1.
Characterizing the IMP1–-KRAS RNA interaction.

(A) Schematic representation of the entire KRAS transcript and its corresponding six RNA fragments assessed for binding to IMP1. (B) The six [32P] RNA fragments corresponding to KRAS RNA nts 1–185, 175–401, 388–610, 568–793, 772–988 and 971–1155 were generated as described in the Materials and methods section. Each of the RNA was analysed by EMSA for binding to various concentrations of IMP1. The positions of the protein-bound RNA and free RNA are shown. (C) EMSA showing binding of IMP1 (0–720 nM) to [32P] KRAS RNA nts 1–185. The intensity of bound and free RNA was quantified to determine the % bound RNA at each concentration of IMP1 to obtain a saturation binding curve shown on the right. Three separate experiments were combined to generate the binding curve.

Figure 1.
Characterizing the IMP1–-KRAS RNA interaction.

(A) Schematic representation of the entire KRAS transcript and its corresponding six RNA fragments assessed for binding to IMP1. (B) The six [32P] RNA fragments corresponding to KRAS RNA nts 1–185, 175–401, 388–610, 568–793, 772–988 and 971–1155 were generated as described in the Materials and methods section. Each of the RNA was analysed by EMSA for binding to various concentrations of IMP1. The positions of the protein-bound RNA and free RNA are shown. (C) EMSA showing binding of IMP1 (0–720 nM) to [32P] KRAS RNA nts 1–185. The intensity of bound and free RNA was quantified to determine the % bound RNA at each concentration of IMP1 to obtain a saturation binding curve shown on the right. Three separate experiments were combined to generate the binding curve.

Table 1
Sequences of AONs used in the competitive EMSA against KRAS RNA
Name Sequences (5′ to 3′) 
SM1 ACCACAAGTTTATATTCAGTCAT 
SM2 TGCCTACGCCACCAGCTCCAACT 
SM3 TAGCTGTATCGTCAAGGCACTCT 
SM4 TCGTCCACAAAATGATTCTGAAT 
SM5 AATCCTCTATTGTTGGATCATAT 
SM6 AATTACTACTTGCTTCCTGTAGG 
SM7 TCCAAGAGACAGGTTTCTCCATC 
SM8 TCATGACCTGCTGTGTCGAGAATA 
Name Sequences (5′ to 3′) 
SM1 ACCACAAGTTTATATTCAGTCAT 
SM2 TGCCTACGCCACCAGCTCCAACT 
SM3 TAGCTGTATCGTCAAGGCACTCT 
SM4 TCGTCCACAAAATGATTCTGAAT 
SM5 AATCCTCTATTGTTGGATCATAT 
SM6 AATTACTACTTGCTTCCTGTAGG 
SM7 TCCAAGAGACAGGTTTCTCCATC 
SM8 TCATGACCTGCTGTGTCGAGAATA 
Table 2
Sequences of primers used for amplifying KRAS DNA fragments and for generating fragments used to construct psicheck2-IMP1-3′-UTR-luciferase reporter plasmids
Name Sequences (5′ to 3′) 
KRAS 1–185 Forward: GGATCCTAATACGACTCACTATAGGATGACTGAATATAAACTT
Reverse: TCATGACCTGCTGTGTCG 
KRAS 175–401 Forward: GGATCCTAATACGACTCACTATAGG GCAGGTCATGAGGAGTAC
Reverse: GCTAAGTCCTGAGCCTGT 
KRAS 388–610 Forward: GGATCCTAATACGACTCACTATAGGGCTCAGGACTTAGCAAGA
Reverse: CCACTTGTACTAGTATGC 
KRAS 568–793 Forward: GGATCCTAATACGACTCACTATAGGATACAATTTGTACTTTTT
Reverse: CACAGGCATTGCTAGTTC 
KRAS 772–988 Forward: GGATCCTAATACGACTCACTATAGGTTTTGAACTAGCAATGCC
Reverse: CCAATTAGAAGGTCTCAA 
KRAS 971–1155 Forward: GGATCCTAATACGACTCACTATAGGTTGAGACCT TCTAATTGG
Reverse: CATCATCAGGAAGCCCAT 
To generate F1
(nts 2069–3586) 
Forward: ACCA CTCGAG CCAGCCCCTCCCTGTCCCTT
Reverse: ACCA GCGGCCGCCATATG TAGAGAGGCCTT 
To generate F2
(nts 3587–5401) 
Forward: ACCA CATATG GAAAAGCCCATG
Reverse: ACCA GCGGCCGC TCACTCTTCTCAAATCAT 
To generate F3
(nts 5402–7149) 
Forward: ACCA CTCGAG ATGATTTGAGAAGAGTGA
Reverse: ACCA GCGGCCGCCATATG CAGCCCAATCGATATGG 
To generate F4
(nts 7150–8708) 
Forward: ACCA CATATG GGAGGGCTGGCCATGAGG
Reverse: ACCA GCGGCCGC TGGCTGAAACAACCTTCA 
Name Sequences (5′ to 3′) 
KRAS 1–185 Forward: GGATCCTAATACGACTCACTATAGGATGACTGAATATAAACTT
Reverse: TCATGACCTGCTGTGTCG 
KRAS 175–401 Forward: GGATCCTAATACGACTCACTATAGG GCAGGTCATGAGGAGTAC
Reverse: GCTAAGTCCTGAGCCTGT 
KRAS 388–610 Forward: GGATCCTAATACGACTCACTATAGGGCTCAGGACTTAGCAAGA
Reverse: CCACTTGTACTAGTATGC 
KRAS 568–793 Forward: GGATCCTAATACGACTCACTATAGGATACAATTTGTACTTTTT
Reverse: CACAGGCATTGCTAGTTC 
KRAS 772–988 Forward: GGATCCTAATACGACTCACTATAGGTTTTGAACTAGCAATGCC
Reverse: CCAATTAGAAGGTCTCAA 
KRAS 971–1155 Forward: GGATCCTAATACGACTCACTATAGGTTGAGACCT TCTAATTGG
Reverse: CATCATCAGGAAGCCCAT 
To generate F1
(nts 2069–3586) 
Forward: ACCA CTCGAG CCAGCCCCTCCCTGTCCCTT
Reverse: ACCA GCGGCCGCCATATG TAGAGAGGCCTT 
To generate F2
(nts 3587–5401) 
Forward: ACCA CATATG GAAAAGCCCATG
Reverse: ACCA GCGGCCGC TCACTCTTCTCAAATCAT 
To generate F3
(nts 5402–7149) 
Forward: ACCA CTCGAG ATGATTTGAGAAGAGTGA
Reverse: ACCA GCGGCCGCCATATG CAGCCCAATCGATATGG 
To generate F4
(nts 7150–8708) 
Forward: ACCA CATATG GGAGGGCTGGCCATGAGG
Reverse: ACCA GCGGCCGC TGGCTGAAACAACCTTCA 

T7 promoter sequences are bolded.

Restriction enzyme sites are underlined.

Generation and purification of recombinant IMP1 and its variants

The plasmid pET28b(+)-IMP1, which contains the mouse IMP1 cDNA, was used to generate recombinant WT IMP1. The generation of various KH point mutation variants was accomplished using the PCR-based site-directed mutagenesis method and has been previously described [5]. Recombinant CRD-BP was purified from Escherichia coli BL21 (DE3) using a 1 ml bed volume of nickel-NTA (QIAGEN) column under denaturing conditions. Proteins eluted from the column at pH 5.4 were subjected to three steps of dialysis. The first step was for 24 h in pH 7.4 buffer containing 200 mM NaCl, 20 mM Tris–HCl, 1 mM reduced glutathione, 0.1 mM oxidized glutathione, 10% (v/v) glycerol, 2 M urea and 0.01% (v/v) Triton X-100. Proteins were then dialysed twice, each for 2 h in the same buffer as above, but without urea and the glutathiones. Following dialysis, samples were spun at 13 200 rpm for 30 min to remove any precipitated proteins. The purified protein solutions were then quantified using Quick Start Bradford 1x Dye Reagent (Bio-Rad, Mississauga, Ontario) or BCA Protein Assay Kit (ThermoFisher, Ottawa, Ontario), and analysed for purity using Coomassie brilliant blue-stained 12% SDS–PAGE.

Generation of DNA templates and radiolabelled in vitro transcription

The human KRAS cDNA clone, which contains the full coding sequence and a partial 3′-UTR (accession # NM_033360), was purchased from Origene Technologies Inc. (Rockville, MD, U.S.A.). It was used as a template for the generation of PCR-amplified DNA fragments. The different sets of forward and reverse primers used to amplify the different regions of KRAS cDNA are shown in Table 2. PCR-amplified DNA templates were used directly for in vitro transcription by T7 RNA polymerase. One microgram of DNA template was incubated for 1 h at 37°C in a 20-µl reaction containing 1x transcription buffer (Promega, Madison, Wisconsin), 10 mM dithiothreitol, 1 unit RNasin (Promega), 0.5 mM ATP, 0.5 mM CTP, 0.5 mM GTP, 12.5 µM UTP, 1.5 units T7 RNA polymerase (Promega, Madison, WI) and 40 µCi [α-32P] UTP (3000 Ci/mmol). Following incubation, 3 units of RNase-free DNase I (Promega) were added, and the reaction was further incubated for 10 min at 37°C. Upon addition of 10 µl Stopping dye (9 M urea, 0.01% bromophenol blue, 0.01% xylene cyanol FF, 0.01% phenol), the entire sample was electrophoresed on a 8% polyacrylamide/7M urea gel. The band containing internally radiolabelled RNA was excised and eluted with elution buffer (10 mM Tris–HCl pH 7.5, 0.1M NaCl, 1 mM EDTA, 0.01% SDS) at 45°C for 6 h. The purified, radiolabelled RNA was then phenol/chloroform extracted followed by ethanol precipitation. Specific activity of the RNA was then determined by scintillation counting.

Generation of unlabelled RNA

For the synthesis of unlabelled KRAS RNA nts 1–185 for use in competitive electrophoretic mobility shift assay (EMSA), the following reactions and procedure were followed. Each in vitro transcription reaction contained 5 µg of gel-purified DNA template, 10 µl 10x T7 buffer (400 mM Tris–HCl pH 7.6, 240 mM MgCl2, 20 mM spermidine, 0.1% Triton X-100), 5 µl each of 100 mM ATP, CTP, GTP and UTP, 40 units of RNasin, 5 units of T7 RNA polymerase to a total volume of 100 µl. The reaction was incubated at 37°C overnight followed by treatment with DNase I to remove the DNA template. Phenol–chloroform extraction was performed to remove protein impurities followed by passing through G-50 spin column to remove unincorporated nucleotides. RNA was ethanol precipitated and re-suspended in DEPC-treated water. Purity of RNA was checked by visualization on 2.5% agarose gel, and their concentration was determined using spectrophotometer.

Electrophoretic mobility shift assay

The EMSA was conducted in essentially the same manner as previously described [5]. EMSA competition assays involved the pre-incubation between competitor molecules (oligonucleotides or RNA) and 540 nM IMP1 for 10 min at 35°C. Following the pre-incubation, [32P]-labelled KRAS RNA nts 1–185 (20 000 cpm/reaction) was added to the reaction. This was followed by the standard EMSA protocol. The molar excess concentrations of AONs SM1 to SM8 over the [32P]-labelled KRAS RNA are shown in Figure 3.

Cell maintenance and transfection with siRNA and oligonucleotide inhibitors

SW480 human colon cancer cell line was purchased from the American Type Culture Collection (ATCC) (Rockville, Maryland) and maintained in Minimal Essential Media (Life Technologies, Burlington, Ontario) supplemented with 10% fetal bovine serum in a humidified incubator at 37°C supplied with 5% CO2. Upon receipt of cells from ATCC, cells were immediately expanded and frozen in liquid nitrogen in large batches and used for up to 10 passages only (∼3 months). Cells were routinely maintained in 25 or 75 cm2 tissue culture flasks and harvested by 0.25% trypsin/0.02% EDTA treatment when they were in the logarithmic phase of growth for various experiments. Cells were plated in 6-well plates at a density of 1.5 × 105 cells/ml. After 20 h, cells were transfected with 40 nM of double-stranded Dicer substrate RNAi (dsRNAi) which was directed against IMP1 mRNA (Integrated DNA Technologies Inc.) using Lipofectamine 2000 reagent (Life Technologies Inc.). To ensure an efficient knocked down, we transfected cells with dsRNAi again after another 24 h incubation. The sense and antisense sequences were: si-IMP1-1, 5′-GGAAAAGACCUACAGCAGCCAGG-3′ and 3′-CCUUUUCUGGAUGUCGUCGGUCC-5′; si-IMP1-2, 5′-CUUGGGACCAAAGAAGUUCUCCUCCUU-3′ and 3′-GAACCCUGGUUUCUUCAAGAGGAGG-5′. As a negative control, cells were transfected with the negative control duplex scrambled-negative (SN) (Integrated DNA Technologies Inc.). After a total of 72 h incubation, cells from duplicate wells were subjected to cell lysate or RNA isolation as described below. For treatment with oligonucleotide inhibitors, SW480 cells were transfected with 2′-O-methyl AONs SM2, SM6 and SM7 at a concentration of 125 nM using Lipofectamine 2000 reagent. As with dsRNAi, cells were transfected twice with the AONs. For inhibition of let-7a, cells were transfected with 100 nM 2′-O-methyl hsa-let-7a inhibitor 5′-AAC UAU ACA ACC UAC UAC CUC A-3′ or with 2′-O-methyl Negative control 5′-CAG UAC UUU UGU GUA GUA CAA-3′ [26].

RNA isolation and quantitative real-Time PCR

Total RNA was extracted from cells using the mirVana™ miRNA Isolation Kit (Ambion Inc.) according to the manufacturer's instructions. After removing DNA using DNA Free kit (Ambion Inc.), the first strand of cDNA synthesis was performed using iScript cDNA synthesis kit (Bio-Rad, Mississauga, Ontario) on 1 µg of total RNA. The resulting cDNA (1 μl) was amplified in 25 μl of iQ SYBR Green Supermix (Bio-Rad) or iQ Supermix (Bio-Rad). The following primer sets for IMP1 [27] and KRAS [28] were used in iQ Supermix: IMP1, forward 5′-AACCCTGAGAGGACCATCACT-3′, reverse 5′-AGCTGGGAAAAGACCTACAGC-3′ and Taqman probe 5′/56-FAM/-TGTTGCAGGGCCGAGCAGGA/3BHQ_1/-3′; KRAS, forward 5′-CGAATATGATCCAAC AATAGAG-3′, reverse 5′-ATGTACTGGTCCCTCATT-3′ and Taqman probe 5′/56-FAM/-TACTCCTCTTGACCTGCTGTG/3BHQ_1/-3′. The following primer sets were used in iQ SYBR Green Supermix: β-actin, forward 5′-TTGCCGACAGGATGCAGAAGGA-3′ and reverse 5′-AGGTGGACAGCGAGGCCAGGAT-3′; IGF-2, forward 5′-ACACCCTCCAGTTCGTCTGT-3′, reverse 5′-GGGGTATCTTGGGGAAGTTGT-3′; H19, forward 5′-TGCTGCACTTTACAACCACTG-3′, reverse 5′-ATGGTGTCTTTGATGTTGGGC-3′. The primers used for qRT-PCR to detect c-myc, CD44, GLI1 and β-TrCP1 mRNAs had been previously described [7,29]. For the detection of let-7a and U6 RNA, the following primers were used: let-7a ST-primer, 5′-GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACAACTA-3′; let-7a-F, 5′-GCCGCTGAGGTAGTAGGTTGTA-3′; let-7a-R, 5′-GTGCAGGGTCCGAGGT-3′; U6-F, 5′-CTCGCTTCGGCAGCACA-3′; U6-R, 5′-AACGCTTCACGAATTTGCGT-3′ [30,31]. The comparative cycle-threshold (CT) method was used, and β-actin served as an endogenous control. For instance, the CT of target gene (e.g. KRAS) was first normalized to that of reference gene β-actin, for the test sample (e.g. si-IMP1-1) and the calibrator sample (e.g. SN) using the following formula: ΔCT(test) = CT(target,test) CT(reference,test); ΔCT(calibrator) = CT(target,calibrator) CT (reference,calibrator). Then, the CT of the test sample was normalized to the CT of the calibrator sample using ΔΔCT =ΔCT(test) − ΔCT(calibrator). Finally, the normalized expression ratio was calculated using . For the measurement of let-7a, the stem-loop RT-PCR method was used [30]. Let-7a RNA level was normalized to the reference gene U6 RNA and then expressed relative to the values of the negative control using the above formula.

Preparation of cell lysates, immunoblot analysis and antibodies

SW480 cells were seeded in 6-well plates and transfected with siRNAs or SM oligonucleotides as essentially as described above. Cell lysates were prepared as previously described [5]. For immunoblot analysis, protein samples were resolved on a 10% SDS–PAGE and transferred onto a nitrocellulose membrane. The following primary antibodies were used: anti-IMP1 (sc-166344, 1:1,000, Santa Cruz Biotechnology), anti-KRAS (sc-30, 1:500, Santa Cruz Biotechnology), anti-IGF-II (H-103, sc-5622, 1:1,000, Santa Cruz Biotechnology), anti-HGFα (H-145, sc-7949, 1:1,000, Santa Cruz Biotechnology), anti-thioredoxin (ab26320, 1:1,000, Abcam) and anti-β-actin (clone AC-15, 1:4,000, Sigma). Anti-mouse IgG-HRP (1:4,000, Promega) and anti-rabbit IgG-HRP (1:2,000, Promega) were used as secondary antibodies. All Western blots were visualized with the FluorChem Q system (ProteinSimple) and analysed using the AlphaView Q software (ProteinSimple).

Construction of psiCHECK2-IMP1-3′UTR-luciferase reporter plasmids and luciferase assay in SW480 cells

Luciferase reporter plasmids containing the 3′UTR of IMP1 were generated using different regions of IMP1 3′UTR (NM_006546.3). To generate these constructs, four sets of primers (Table 2) were designed to amplify the 3′UTR of IMP-1 into four fragments (F1– F4) by PCR using genomic DNA isolated from HCT116 as the DNA template. In the first step, only F1 and F3 were individually inserted at the 3′ end of the Renilla luciferase reporter gene in the empty psiCHECK-2 vector using Xho1 and Not1 sites to generate psiCHECK-F1 and psiCHECK-F3 plasmids, respectively. Subsequently, F2 and F4 fragments were inserted at the 3′ end of F1 and F3, using Nde1 and Not1 sites, to generate psiCHECK-F1 + F2 and psiCHECK-F3 + F4 plasmids, respectively. An endogenous Nde1 site at the 3′ end of F1 was used, while an additional Nde1 site at the 3′ end of F3 was created in the F3 reverse primer. Sequences of all plasmid constructs were confirmed by DNA sequencing. For the luciferase reporter experiment, SW480 cells were plated onto 96-well plates at a density of 3 × 104 cells/well one day before transfection with SM2, SM6 or SM7. Twenty-four hours after transfection with the AONs, the cells were transfected using Lipofectamine 2000 with the luciferase plasmids. Another 24 h later, luciferase expression was assessed using the Dual-Luciferase Reporter Assay System (Promega) and Bio-Tek's Synergy 2 multi-plate reader. The psiCHECK-2 plasmid has Renilla luciferase reporter gene upstream of the 3′-UTR of IMP1, as well as a Firefly luciferase reporter gene used for normalization of transfection efficiency. In this system, the two luciferase cistrons are expected to be in separate mRNAs.

Results

Mapping the KRAS RNA regions that bind IMP1

IMP1 has high affinity for the coding region and 3′-UTR of KRAS mRNA [24]. However, the region(s) of KRAS mRNA responsible for the binding has not been determined. To map the smallest region(s) of KRAS mRNA which can bind IMP1, we generated three fragments corresponding to the coding region of KRAS mRNA called fragments A (nts 1–185), B (nts 175–401) and C (nts 388–610), and three fragments corresponding to the 3′-UTR of KRAS mRNA called fragments D (nts 568–793), E (nts 772–988) and F (nts 971–1155) (Figure 1A). Using EMSA, all fragments were assessed for their ability to bind to IMP1. Out of the six fragments, only fragments A, C, D and F yielded a higher molecular mass band, suggesting the formation of RNA-protein complexes, when incubated with an increasing concentration of recombinant IMP1 on a non-denaturing gel (Figure 1B). The binding data for all six fragments were analysed by densitometry and Kd plots were generated by fitting to the Hill equation for quantitative comparison of fragments A, D and F (data not shown). Fragment C was not included in the calculation due to repeated inability to achieve greater than 70% bound fraction. Our results show that fragment A (nts 1–185) has the highest apparent affinity for IMP1, with a Kd of 171 ± 18 nM, for IMP1 (Figure 1C). The Kd values for fragments D and F were 220 ± 21 and 278 ± 64 nM, respectively. Based on these findings, we chose fragment A in subsequent experiments for the study of IMP1–KRAS mRNA interaction. It is important to point out that it is not feasible to directly compare affinity of the 1200 nt full-length KRAS mRNA with the KRAS RNA fragments described above. From our experiences performing EMSA with IMP1, size appears to be a factor in binding, with larger RNA generally having higher apparent affinity. This may reflect the existence of multiple binding sites.

Assessing the role of the KH domains of IMP1 in binding KRAS RNA

The highly conserved GXXG motif located within the flexible loop in a KH domain plays a critical role in the RNA-binding function of several proteins [3235]. The glycine residues in the GXXG motif are essential because of their conformational flexibility and small steric size [32]. Indeed, mutating the first glycine in the GXXG motif to an aspartate has been shown to severely impede RNA-binding function of many proteins containing KH domains [3235], including IMP1 [5,36]. Using site-directed mutagenesis to mutate the first glycine to an aspartate in the GXXG motif of KH domains, we recently showed that at least two KH domains of IMP1 participate in binding c-myc and CD44 RNAs [5]. To determine whether IMP1 has similar requirements for binding to KRAS mRNA, we used single point-mutation KH variants (KH1, KH2, KH3 and KH4) and double point-mutation KH variants (KH1–2, KH1–3, KH1–4, KH2–3, KH2–4 and KH3–4) as previously described [5]. The double point-mutation KH variants have point mutations in the GXXG motifs of two KH domains. For example, the KH1–2 variant has the point mutations at the GXXG motifs of both the KH1 and KH2 domains. Figure 2 shows representative results from EMSA in assessing these variants for binding to [32P]-labelled KRAS RNA nts 1–185. c-myc RNA nts 1705–1886 which has high affinity to IMP1 [37] was used as a positive control. As previously described [37], IMP1 probably binds c-myc RNA cooperatively as a dimer as evidenced by two distinct bands (two upper images, Figure 2). In contrast, only one upper band was observed with the KRAS RNA suggesting that IMP1 dimer may have higher affinity for KRAS RNA. KH1 variant could bind KRAS RNA although at much reduced ability as compared with the WT IMP1 (upper panel, Figure 2). The other KH single variants, KH2, KH3 and KH4, had even lower ability to bind KRAS RNA as evidenced by the lower intensity of the RNA-protein complex. For all the KH single variants, determination of Kd values was not possible due to very low % bound fraction. All the di-domain variants, including the KH3–4 variant, were completely ineffective in binding KRAS RNA (bottom three images, Figure 2). Taken together, these results indicate all four KH domains are important for binding to KRAS RNA and removal of any one KH domain substantially reduces binding to KRAS RNA.

KRAS RNA binding to various IMP1 KH domain variants.

Figure 2.
KRAS RNA binding to various IMP1 KH domain variants.

[32P] KRAS RNA nts 1–185 was subject to EMSA with various concentrations of IMP1 KH domain variants as indicated. Samples within each panel indicate results from the same gel generated from the same experiment. Data shown are representatives from at least three experiments using at least two separately prepared recombinant proteins. The positions of protein-bound and free RNA are indicated.

Figure 2.
KRAS RNA binding to various IMP1 KH domain variants.

[32P] KRAS RNA nts 1–185 was subject to EMSA with various concentrations of IMP1 KH domain variants as indicated. Samples within each panel indicate results from the same gel generated from the same experiment. Data shown are representatives from at least three experiments using at least two separately prepared recombinant proteins. The positions of protein-bound and free RNA are indicated.

Assessment of a panel of AONs for ability to inhibit the IMP1–KRAS RNA interaction.

Figure 3.
Assessment of a panel of AONs for ability to inhibit the IMP1–KRAS RNA interaction.

(A) The predicted secondary structure of KRAS RNA (nts 1–185) was generated using the Mfold programme [47]. The solid lines (black, blue, and red) indicate regions where each of the eight SM AONs hybridize. (B) Purified recombinant IMP1 (540 nM) was incubated with [32P] KRAS RNA nts 1–185 in the presence of AONs SM1 to SM8 at 10- or 50-fold molar excess. As a positive control, 10- and 50-fold molar excess of unlabelled KRAS-A (lanes 3 and 4) was used. (C) Data from (B) and two additional experiments were combined (n = 3) to obtain the average % complex reduction as compared with the absence of inhibitor. One-way ANOVA was performed as statistical analysis. *P < 0.02.

Figure 3.
Assessment of a panel of AONs for ability to inhibit the IMP1–KRAS RNA interaction.

(A) The predicted secondary structure of KRAS RNA (nts 1–185) was generated using the Mfold programme [47]. The solid lines (black, blue, and red) indicate regions where each of the eight SM AONs hybridize. (B) Purified recombinant IMP1 (540 nM) was incubated with [32P] KRAS RNA nts 1–185 in the presence of AONs SM1 to SM8 at 10- or 50-fold molar excess. As a positive control, 10- and 50-fold molar excess of unlabelled KRAS-A (lanes 3 and 4) was used. (C) Data from (B) and two additional experiments were combined (n = 3) to obtain the average % complex reduction as compared with the absence of inhibitor. One-way ANOVA was performed as statistical analysis. *P < 0.02.

Identification of SM6 and SM7 as potent AON inhibitors of IMP1–KRAS RNA interaction

We next investigated the possibility of using AONs to inhibit IMP1–KRAS RNA interaction. We designed eight 23-nt AONs, termed SM1 to SM8 (Table 1), which correspond to the entire sequence of KRAS RNA nts 1–185 and assessed their ability to inhibit the binding of 32P-labelled KRAS RNA to IMP1. Figure 3A shows the region targeted by the AONs and Figure 3B shows a representative result of the competitive EMSA. The quantitative data presented as the average % reduction in bound RNA when compared with a control with no inhibitor are shown in Figure 3C. At 50-fold molar excess, the positive control unlabelled KRAS RNA nts 1–185 successfully inhibited IMP1–KRAS interaction with ∼35% complex reduction (Figure 3B,C). Among the eight AONs, SM6 and SM7 appeared to be the most effective inhibitor with 55 and 45% complex reduction, respectively (Figure 3B,C). Using a wider concentration range of the oligonucleotides, the effectiveness of SM6 and SM7 as inhibitors of IMP1–KRAS RNA interaction was confirmed (Figure 4). Such inhibition is specific since SM2 oligonucleotide had no effect on IMP1–KRAS RNA interaction even up to 4000-fold molar excess (Figure 4A).

Inhibition of IMP1–KRAS RNA interaction by SM6 and SM7 AONs.

Figure 4.
Inhibition of IMP1–KRAS RNA interaction by SM6 and SM7 AONs.

(A) Purified recombinant IMP1 (540 nM) was incubated with [32P] KRAS RNA nts 1–185 in the presence or absence of various concentrations of SM2, SM6 and SM7. The quantified data from (A) and two additional experiments for SM6 and SM7 were combined to generate the binding curves as shown in (B).

Figure 4.
Inhibition of IMP1–KRAS RNA interaction by SM6 and SM7 AONs.

(A) Purified recombinant IMP1 (540 nM) was incubated with [32P] KRAS RNA nts 1–185 in the presence or absence of various concentrations of SM2, SM6 and SM7. The quantified data from (A) and two additional experiments for SM6 and SM7 were combined to generate the binding curves as shown in (B).

Assessing the effect of SM6 and SM7 oligonucleotides on KRAS and IMP1 expression in cells

Based on the observation that both SM6 and SM7 can inhibit IMP1–KRAS RNA interaction in vitro, we next assessed whether these oligonucleotides are effective inhibitors of KRAS expression in cells. To determine whether KRAS expression can be controlled by IMP1, we used SW480 human colon cancer cell line in which IMP1 is overexpressed [24]. IMP1 was knocked down using two different siRNAs, si-IMP1-1 [24] and si-IMP1-2 [27], previously shown to be effective. Our results show that siRNA could reduce IMP1 as well as KRAS protein expression (Figure 5A), consistent with a previous report [24]. To determine whether the reduced KRAS protein level is related to the reduced mRNA level, KRAS mRNA levels were determined using qRT-PCR. To our surprise, while IMP1 mRNA decreased to ∼30–40% in siRNAs-transfected cells, we did not see any changes in KRAS mRNA levels (Figure 5B). These results suggest that IMP1 positively regulates the translation of KRAS rather than controlling the stability of KRAS mRNA.

Effect of IMP1 knockdown on IMP1 and KRAS expression.

Figure 5.
Effect of IMP1 knockdown on IMP1 and KRAS expression.

(A) SW480 human colon cancer cells were transfected with 40 nM of si-IMP1-1, si-IMP1-2 or SN for 48 h after which cell lysates were isolated and subjected to Western blot analysis. The levels of IMP1 and KRAS in si-IMP1-treated cells were normalized to β-actin and expressed relative to SN-treated cells as described in the Materials and methods section. The bar graph shown is the averaged data taken from two biological replicates. (B) SW480 cells were transfected as in (A) and total RNAs extracted were subject to IMP1 and KRAS mRNAs quantification using qRT-PCR. Results shown were averaged from four biological replicates ± SEM. For (A) and (B), one-way ANOVA was performed as statistical analysis. *P <0.001; **P <0.01; ***P <0.05.

Figure 5.
Effect of IMP1 knockdown on IMP1 and KRAS expression.

(A) SW480 human colon cancer cells were transfected with 40 nM of si-IMP1-1, si-IMP1-2 or SN for 48 h after which cell lysates were isolated and subjected to Western blot analysis. The levels of IMP1 and KRAS in si-IMP1-treated cells were normalized to β-actin and expressed relative to SN-treated cells as described in the Materials and methods section. The bar graph shown is the averaged data taken from two biological replicates. (B) SW480 cells were transfected as in (A) and total RNAs extracted were subject to IMP1 and KRAS mRNAs quantification using qRT-PCR. Results shown were averaged from four biological replicates ± SEM. For (A) and (B), one-way ANOVA was performed as statistical analysis. *P <0.001; **P <0.01; ***P <0.05.

When SW480 cells were treated with SM6 and SM7, KRAS protein level remained unchanged, while IMP1 protein level reduced significantly (60%; Figure 6A). Although the KRAS level appeared to increase in the blot shown in Figure 6A, experiments from three biological replicates confirm that there was no significant effect (bar graph in Figure 6A). The reduced IMP1 protein expression did not correlate with a decrease in mRNA levels as shown by qRT-PCR analysis (Figure 6B). These results suggest that reduced IMP1 protein expression may be a result of reduced IMP1 translation in SM6- and SM7-treated cells.

Effect of SM2, SM6 and SM7 on IMP1 and KRAS expression.

Figure 6.
Effect of SM2, SM6 and SM7 on IMP1 and KRAS expression.

(A) SW480 human colon cancer cells were transfected with 125 nM of SM6, SM7 or SM2 for 48 h after which cell lysates were isolated and subjected to Western blot analysis. The levels of IMP1 and KRAS in SM6- and SM7-treated cells were normalized to β-actin and expressed relative to SM2-treated cells as described in the Materials and methods section. One-way ANOVA was performed as statistical analysis. *P < 0.001. (B) SW480 cells were transfected as in (A) and total RNAs extracted were subject to IMP1 and KRAS mRNAs quantification using qRT-PCR. The bar graphs shown are averaged data taken from four biological replicates ± SEM.

Figure 6.
Effect of SM2, SM6 and SM7 on IMP1 and KRAS expression.

(A) SW480 human colon cancer cells were transfected with 125 nM of SM6, SM7 or SM2 for 48 h after which cell lysates were isolated and subjected to Western blot analysis. The levels of IMP1 and KRAS in SM6- and SM7-treated cells were normalized to β-actin and expressed relative to SM2-treated cells as described in the Materials and methods section. One-way ANOVA was performed as statistical analysis. *P < 0.001. (B) SW480 cells were transfected as in (A) and total RNAs extracted were subject to IMP1 and KRAS mRNAs quantification using qRT-PCR. The bar graphs shown are averaged data taken from four biological replicates ± SEM.

Effect of SM6 and SM7 oligonucleotides on the expression of other IMP1 targets in cells

To determine whether SM6- and SM7-mediated down-regulation of IMP1 has any consequences on the expression of IMP1-targeted mRNAs, we assessed the RNA levels of c-myc, CD44, IGF-II, HGF-α, GLI1, β-TrCP1 and H19, and protein levels of c-myc, CD44, HGF-α and IGF-II. Both the mRNA and protein levels of c-myc, CD44 and HGF-α were extremely low or undetectable in SW480 cells (data not shown). As shown in Figure 7A, transfection of cells with SM6 or SM7 led to a significant increase in IGF-II mRNA levels, while there was a significant decrease in H19 RNA level. SM6 and SM7 had no significant effect on GLI1 and β-TrCP1 mRNA levels. In contrast, cells transfected with SM6 or SM7 show significantly reduced IGF-II protein level (Figure 7B).

Effect of SM6 and SM7 on gene expression in SW480 cells.

Figure 7.
Effect of SM6 and SM7 on gene expression in SW480 cells.

(A) SW480 cells were transfected with 125 nM of SM6, SM7 or SM2 for 48 h after which RNAs were isolated and subjected to qRT-PCR analysis as described in Materials and methods section. Results shown were averaged from three (n = 3 for IGF-II GLI1, and β-TrCP1 mRNAs) or four separate biological replicates (n = 4 for H19 RNA) ± SEM. (B) SW480 cells were transfected as in (A) and isolated cell lysate were subjected to Western blot analysis as shown on the left panel. The bar graph shown on the right panel was taken from three biological replicates (n = 3). For (A) and (B), one-way ANOVA was performed as statistical analysis. *P < 0.001; **P < 0.05.

Figure 7.
Effect of SM6 and SM7 on gene expression in SW480 cells.

(A) SW480 cells were transfected with 125 nM of SM6, SM7 or SM2 for 48 h after which RNAs were isolated and subjected to qRT-PCR analysis as described in Materials and methods section. Results shown were averaged from three (n = 3 for IGF-II GLI1, and β-TrCP1 mRNAs) or four separate biological replicates (n = 4 for H19 RNA) ± SEM. (B) SW480 cells were transfected as in (A) and isolated cell lysate were subjected to Western blot analysis as shown on the left panel. The bar graph shown on the right panel was taken from three biological replicates (n = 3). For (A) and (B), one-way ANOVA was performed as statistical analysis. *P < 0.001; **P < 0.05.

Effect of SM6 and SM7 oligonucleotides on the translation of IMP1

Lower IMP1 protein expression without a concomitant decrease in the mRNA level suggests that the translation of IMP1 is reduced in SM6- and SM7-treated cells. The 3′-UTR of IMP1 contains many microRNA target sites and is therefore subjected to translational regulation [3841]. To investigate whether translational repression is induced by SM6 and SM7 in SW480 cells, we used the luciferase reporter system where Renilla luciferase gene is fused with the 3′-UTR of IMP1. Different regions of the 3′-UTR of IMP1, namely nts 2069–3586 (F1), nts 2069–5401 (F1 + F2) and nts 5402–8708 (F3 + F4) were used to construct three different reporter plasmids as illustrated in Figure 8A. These reporter plasmids also contain the firefly luciferase gene for transfection efficiency normalization. We first determined whether the 3′-UTR alone has any effect on the translation of IMP1 in SW480 cells. When transfected into the cells, normalized Renilla luciferase activity was reduced by ∼50% when it was fused to the F1 and F1 + F2 regions and was reduced by ∼15% when fused to the F3 + F4 region (Figure 8B). These results suggest that the major translational repression site of IMP1 in SW480 cells lies within nts 2069–3586 because both the F1 and F1 + F2 regions decrease the relative luciferase activity to almost the same extent (Figure 8B). We next determined whether SM6 and SM7 treatment can influence the translation of luciferase fused with the different regions of 3′-UTR of IMP1. We compared the normalized Renilla luciferase activity to that measured in SM2-treated control cells. Figure 8C shows that SM6 and SM7 had no significant effect on the luciferase activity when F1 + F2 was fused to the luciferase gene. However, both SM6 and SM7 caused a decrease in the luciferase activity, with significant effect shown by SM7, when F3 + F4 region was placed downstream of the luciferase gene. In summary, using Renilla luciferase reporter system, translation under the regulation of the 3′-UTR of IMP1 is indeed repressed in SM6- and SM7-treated cells.

Effect of SM6 and SM7 on the translation of IMP1.

Figure 8.
Effect of SM6 and SM7 on the translation of IMP1.

(A) Schematic representation of the 3′-UTR of IMP1 RNA and the regions sub-cloned into psiCHECK-2 luciferase reporter plasmid. (B) Luciferase reporter plasmids containing different regions of IMP1 3′-UTR, namely F1, F1 + F2, F3 + F4, or were transfected into SW480 cells. Renilla luciferase activity was normalized against Firefly luciferase activity for transfection efficiency and then expressed as a ratio relative to empty psiCHECK-2 vector as a control. (C) PsiCHECK-2 empty vector and luciferase reporter plasmids containing F1 + F2 or F3 + F4 region of IMP1 3′-UTR were transfected into SW480 cells previously transfected with SM2, SM6 or SM7. Normalized Renilla luciferase activity was first expressed as a ratio relative to psiCHECK-2 for each treatment group and then expressed as a ratio relative to SM2. Results shown in (B) and (C) were averaged from three biological replicates ± SEM. For (B) and (C), one-way ANOVA was performed as statistical analysis. *P < 0.001; **P < 0.01. (D) Binding of SM6 and SM7 to IMP1 3′-UTR as predicted by RNAHybrid 2.2 software. Watson-Crick and wobble (G-U) base pairing are shown by solid and dashed lines, respectively. Red coloured nucleotides in SM6 and SM7 are in the loop region as shown in (E). (E) Secondary structure of SM6 and SM7 as predicted by Mfold programme [47].

Figure 8.
Effect of SM6 and SM7 on the translation of IMP1.

(A) Schematic representation of the 3′-UTR of IMP1 RNA and the regions sub-cloned into psiCHECK-2 luciferase reporter plasmid. (B) Luciferase reporter plasmids containing different regions of IMP1 3′-UTR, namely F1, F1 + F2, F3 + F4, or were transfected into SW480 cells. Renilla luciferase activity was normalized against Firefly luciferase activity for transfection efficiency and then expressed as a ratio relative to empty psiCHECK-2 vector as a control. (C) PsiCHECK-2 empty vector and luciferase reporter plasmids containing F1 + F2 or F3 + F4 region of IMP1 3′-UTR were transfected into SW480 cells previously transfected with SM2, SM6 or SM7. Normalized Renilla luciferase activity was first expressed as a ratio relative to psiCHECK-2 for each treatment group and then expressed as a ratio relative to SM2. Results shown in (B) and (C) were averaged from three biological replicates ± SEM. For (B) and (C), one-way ANOVA was performed as statistical analysis. *P < 0.001; **P < 0.01. (D) Binding of SM6 and SM7 to IMP1 3′-UTR as predicted by RNAHybrid 2.2 software. Watson-Crick and wobble (G-U) base pairing are shown by solid and dashed lines, respectively. Red coloured nucleotides in SM6 and SM7 are in the loop region as shown in (E). (E) Secondary structure of SM6 and SM7 as predicted by Mfold programme [47].

Assessing the possible role of let-7a

Let-7a is a miRNA known to target the 3′-UTR of many genes, including IMP1 [24,38] and KRAS [42]. To assess let-7a as a possible mediator of SM7-induced translational repression of IMP1, we performed the following experiments. First, we employed a let-7a inhibitor previously shown to be effective in inhibiting let-7a expression [26] to investigate the effect on IMP1 translation in SW480 cells using the luciferase reporter assay. Figure 9A shows that treatment with 100 nM let-7a inhibitor led to a reduction in let-7a level by ∼50% in SW480 cells as compared with the negative control oligonucleotide. However, let-7a inhibition had no significant effect on the relative Renilla luciferase activity when F3 + F4 region of 3′-UTR of IMP1 was placed downstream of the luciferase gene (Figure 9B). Such finding together with the results in Figure 8C suggests that let-7a is not a mediator in SM7-induced translational repression of IMP1. Secondly, we measured the level of let-7a in SM6- and SM7-treated cells. As shown in Figure 9C, both oligonucleotides had no effect on let-7a level as compared with SM2. This further supports the notion that let-7a is not involved in SM7-mediated translational repression of IMP1.

Assessing let-7a as a possible mediator in the control of IMP1 expression.

Figure 9.
Assessing let-7a as a possible mediator in the control of IMP1 expression.

(A) SW480 cells were transfected with 100 nM let-7a inhibitor for 48 h after which RNAs were isolated and subjected to qRT-PCR analysis. Results shown were averaged from two separate biological replicates. (B) PsiCHECK-2 empty vector and luciferase reporter plasmids containing F3 + F4 region of IMP1 3′-UTR were transfected into SW480 cells previously transfected with let-7a inhibitor or the Negative control oligonucleotide. Normalized Renilla luciferase activity was first expressed as a ratio relative to psiCHECK-2 for each treatment group and then expressed as a ratio relative to the Negative control. Results shown were averaged from two biological replicates ± SD. (C) SW480 cells were transfected with 125 nM SM2, SM6 or SM7 for 48 h after which RNAs were isolated and subjected to qRT-PCR analysis to determine the let-7a and U6 RNA levels. Results shown were averaged from two separate biological replicates.

Figure 9.
Assessing let-7a as a possible mediator in the control of IMP1 expression.

(A) SW480 cells were transfected with 100 nM let-7a inhibitor for 48 h after which RNAs were isolated and subjected to qRT-PCR analysis. Results shown were averaged from two separate biological replicates. (B) PsiCHECK-2 empty vector and luciferase reporter plasmids containing F3 + F4 region of IMP1 3′-UTR were transfected into SW480 cells previously transfected with let-7a inhibitor or the Negative control oligonucleotide. Normalized Renilla luciferase activity was first expressed as a ratio relative to psiCHECK-2 for each treatment group and then expressed as a ratio relative to the Negative control. Results shown were averaged from two biological replicates ± SD. (C) SW480 cells were transfected with 125 nM SM2, SM6 or SM7 for 48 h after which RNAs were isolated and subjected to qRT-PCR analysis to determine the let-7a and U6 RNA levels. Results shown were averaged from two separate biological replicates.

Discussion

IMP1 has affinity for the coding region as well as the 3′-UTR of KRAS mRNA [24]. Consistent with this notion, IMP1 was able to influence the expression of KRAS in colon cancer cells [24]. Based on these findings, we set out to characterize the molecular interaction between IMP1 and KRAS RNA. It is anticipated that a better understanding of this interaction could improve our knowledge on the oncogenic mechanism of IMP1 leading to new therapeutic approach against IMP1 and/or KRAS.

Using KH variants with a point mutation in the GXXG motif at each of the KH domains, we previously showed that the single point-mutation in each of the four KH domains had no major impact on the ability of IMP1 to bind to c-myc and CD44 RNAs [5]. Here, we show that the all single point-mutation KH variants had significantly reduced the ability to bind KRAS RNA (Figure 2), suggesting that the GXXG motif at all four KH domains does play an important role in associating with KRAS RNA. All the KH variants with double mutations, including KH3–4 variant, showed complete abolition in binding KRAS RNA (Figure 2). We found similar results with GLI1 RNA [7] but not with the c-myc, CD44 [5] and MITF [8] RNAs where the KH3–4 variant was equally as effective as the wild-type IMP1 in binding. Results from the present study support the notion that IMP1 utilizes its different KH domains to bind distinct RNAs and suggest that any future development of IMP1 inhibitors need to focus on specific RNA target in question.

In the present study, we mapped both the coding region and 3′-UTR of KRAS mRNA to determine the smallest region for interaction with IMP1. We found regions spanning nts 1–185 and nts 388–610 of coding region, and regions spanning nts 568–793 and 971–1155 at the 3′-UTR to have affinity for IMP1 (Figure 1). By far, IMP1 has the highest affinity for nts 1–185 at the coding region of KRAS RNA. This information is important because it has allowed us to develop a fluorescence polarization method to study IMP1–KRAS RNA interaction in a rapid and convenient manner (unpublished). Furthermore, we were able to design AONs targeting nts 1–185 of KRAS RNA to disrupt IMP1–KRAS RNA as described below.

Based on the knowledge that nts 1–185 of KRAS RNA interacts with IMP1 with high affinity, we designed eight AONs against IMP1–KRAS RNA. We found two AONs, SM6 and SM7, as effective inhibitors of IMP1–KRAS RNA interaction. We hypothesized that AONs SM6 and SM7 should be able to decrease KRAS expression in cells where IMP1 has been proposed to bind and shield KRAS mRNA from decay. We chose the SW480 human colon cancer cell where such relationship exists [24]. To first confirm this, we knocked down IMP1 using two different siRNAs targeting different regions of IMP1 mRNA. Both siRNAs were equally effective in decreasing IMP1 mRNA and protein levels (Figure 5). In agreement with the previous report [24], we found inhibition of KRAS protein expression in SW480 cells transfected with siRNAs against IMP1 (Figure 5A). However, to our surprise, we consistently did not observe any reduction in KRAS mRNA in cells where IMP1 was reduced (Figure 5B). In the previous study, the effect on KRAS mRNA expression upon knockdown of IMP1 was not reported [24]. The decrease in KRAS protein, but not its mRNA, upon the loss of IMP1 suggests that IMP1 regulates KRAS expression at the level of translation.

Here, we found that both SM6 and SM7 AONs that inhibited IMP1–KRAS RNA interaction in vitro had no effect on KRAS mRNA and protein expression in SW480 cells (Figure 6). Based on the in vitro findings that other regions of KRAS mRNA are also able to bind IMP1 (Figure 1), it is therefore not surprising that SM6 and SM7 are not able to completely block IMP1–KRAS mRNA interaction in cells. It is also possible that the binding sites of SM6 and SM7 on KRAS mRNA are buried in complex tertiary or quaternary structures in cells and are not easily accessible. What was surprising to us was the finding that both SM6 and SM7 can significantly inhibit IMP1 protein expression but not its mRNA level (Figure 6), suggesting the inhibition is at the translation level. This was indeed confirmed when a reporter was used to test this hypothesis. We generated plasmids containing the Renilla luciferase gene fused to different regions of the 3′-UTR of IMP1. Indeed, SM6 and more significantly SM7 inhibited the luciferase activity when the 3′ end region (nts 5402–8708) of the 3′-UTR of IMP1 was fused with the luciferase gene (Figure 8C). Given that the 3′-UTR of IMP1 has many let-7 target sites [38], we tested the hypothesis that let-7a is a mediator in SM7-induced translational repression of IMP1. We found that let-7a inhibitor had no effect on the relative Renilla luciferase activity when the gene was fused to 3′-UTR of IMP1 (Figure 9B). Furthermore, both SM6 and SM7 had no effect on let-7a level (Figure 9C). Our overall results suggest that let-7a does not play a role in mediating SM7-induced translational repression of IMP1. However, we do not rule out other let-7 family members as possible mediators [38].

We propose an unexpected direct effect of SM6 and SM7 on IMP1 mRNA leading to the inhibition of IMP1 translation. As predicted by RNAHybrid 2.2 software, SM6 and SM7 can hybridize with nts 7285–7308 and nts 7301–7324 of the 3′-UTR of IMP1 mRNA, respectively (Figure 8D). Interestingly, the predicted secondary structure of SM7 with a stem-loop structure (Figure 8E) resembles a molecular beacon [43] containing a single-stranded region with the exact nucleotides involved in hybridization to the target site of 3′-UTR of IMP1. To this end, RNA molecules with a stem-loop structure, like a molecular beacon, are better able to present their inhibitory sequence in the single-stranded region making it a strong miRNA inhibitor [43]. In the present study, both SM6 and SM7 are likely to be able to block the interaction between IMP1 and nts 1–185 of KRAS mRNA in cells. However, as shown in vitro (Figure 1), the other regions in KRAS mRNA are also able to bind IMP1. Therefore, such an overall effect could not occlude IMP1 from binding to KRAS mRNA in SM6- and SM7-treated cells. Additionally, SM6 and SM7 could bind strongly to the 3′-UTR of IMP1 due to its complementary region in the single-stranded loop region (Figure 8D,E), functioning like miRNAs binding to the 3′-UTR of target genes to mediate translational suppression. It is also interesting to note that only 3 out of 5 nucleotides in the loop region of SM6 (Figure 8E) are predicted to hybridize to the 3′-UTR of IMP1 (Figure 8D). Such difference may explain the lower potency of SM6 in targeting the 3′-UTR of IMP1 leading to less inhibitory effect on the translation of luciferase gene. SM2, on the other hand, does not form a stem-loop structure, coinciding with its role as a negative control.

Another novel finding from the present study is that the F1 3′-UTR region (nts 2069–3586) serves as a strong translational repression site of IMP1 in SW480 human colon cancer cells. Interestingly, we found that this is not the case in other cell types (unpublished). This could be due to the different levels of selective miRNAs present in SW480 cells. Although the F1 region of the 3′-UTR of IMP1 confers the strongest translational regulation in SW480 cells, we found that this is not the region responsible for the translation repression in SM6- and SM7-treated cells, further suggesting that the mechanism of the AONs on IMP1 is not via the miRNAs.

It is also interesting to note that the decrease in IMP1 protein level caused by SM6 or SM7 (Figure 6) did not result in the reduction in KRAS protein level. This finding is contradictory to the hypothesis that IMP1 protein controls the level of KRAS expression. However, when IMP1 was knocked down using siRNA, KRAS expression was reduced (Figure 5). The only difference achieved by the two methods is in the level of IMP1 mRNA which may play a more important role than previously perceived. Since siRNA is the only method used in knocking down the expression of a gene in all other studies, the current study provides the first experimental evidence that the level of IMP1 mRNA is important for controlling the expression of other genes. We propose that this is due to the role of IMP1 mRNA in controlling the availability of miRNAs which, in turn, regulate other genes such as KRAS. Thus, when IMP1 mRNA levels in SM6-/SM7-treated cells remained high, miRNAs such as let-7 are less available to bind KRAS mRNA to inhibit its translation.

We also examined the expression of several RNAs known to be targets of IMP1 upon treatment with the AONs. Among the RNAs and proteins, we could only detect IGF-II mRNA and protein levels as well as GLI1, β-TrCP1 and H19 RNA levels in SW480 cells (Figure 7). Both SM6 and SM7 significantly decrease H19 RNA level. Given that IMP1 has been shown to have strong affinity for H19 RNA [44,45] and participates in its localization [44], we propose that the down-regulation of H19 RNA level by SM6 and SM7 is mediated by the decrease in IMP1 protein (Figure 6A). We also found that both SM6 and SM7 significantly increased IGF-II mRNA levels (Figure 7A). Interestingly, significant increase in IGF-II mRNA level has also been observed in MCF-7 human breast cancer cells [18] and K562 human leukemic cells [12] upon knocked down of IMP1 using siRNA. This may be the result of the decrease in H19 RNA. The reciprocal expression of H19 RNA and IGF-II mRNA has been observed and it was proposed that H19 RNA acts as an antagonist of IGF-II mRNA expression [46]. Similar to KRAS, the AONs had no effect on GLI1 and β-TrCP1 expression, probably due to unchanged IMP1 mRNA level. In contrast with IGF-II mRNA, we found that both SM6 and SM7 significantly suppressed IGF-II protein expression (Figure 7B). Like IMP1, we propose that such an inhibition could be a direct effect of the AONs on the translation of IGF-II. Indeed our analysis using RNAHybrid 2.2 software shows that all five nucleotides of the single-stranded loop region of SM6 (Figure 8E) are predicted to target the coding region of IGF-II mRNA, while five out of six nucleotides of SM7 (Figure 8E) are perfectly complementary to the 3′-UTR of IGF-II mRNA (data not shown).

In conclusion, the present study has provided novel insights into the molecular interaction between IMP1 and KRAS mRNA. We show that all four KH domains of IMP1 are important for binding to KRAS RNA. We systematically mapped the KRAS mRNA region that binds IMP1 and show that IMP1 has the highest affinity for nts 1–185 at the coding region. We found two AONs, SM6 and SM7, as effective inhibitors of IMP1–KRAS RNA interaction in vitro. To our surprise, both AONs had no effect on KRAS expression in cells where IMP1–KRAS mRNA relationship exists. Instead, we found that both AONs inhibited IMP1 protein expression. We further show that this inhibition is through translational repression and hence these AONs function as translational inhibitors. The design of 2′-O-methyl AONs against specific transcript, as shown here with the SM6 and SM7 AONs, should be used with caution. We recommend that upon examination of the predicted secondary structure of designed AON for possible stem-loop-like structure, the single-stranded loop region should be further analysed for possible complementary sequences in the database. Together with siRNAs, these AONs which can function as specific translational inhibitors, may be useful tools for studying the functional role of proteins versus their corresponding mRNA transcripts.

Abbreviations

     
  • AON

    antisense oligonucleotide

  •  
  • cDNA

    complementary DNA

  •  
  • dsRNAi

    dicer substrate RNAi

  •  
  • EMSA

    electrophoretic mobility shift assay

  •  
  • IGF-II

    insulin-like growth factor-II

  •  
  • IMP1

    insulin-like growth factor 2 mRNA-binding protein 1

  •  
  • K

    K-homology

  •  
  • KRAS

    Kirsten rat sarcoma

  •  
  • mRNA

    messenger RNA

  •  
  • nts

    nucleotides

  •  
  • qRT-PCR

    quantitative real-time polymerase chain reaction

  •  
  • RRMs

    RNA-recognition motifs

  •  
  • siRNA

    small interfering RNA

  •  
  • SN

    scrambled-negative

  •  
  • UTR

    untranslated region

Author Contribution

S.M. and C.W. performed the experiments and data analysis. W.-M.L. provided expert advice on plasmid construction, data analysis and interpretation of experimental results. C.H.L. supervised the research and wrote the manuscript. All authors reviewed drafts of the manuscript.

Funding

This work was supported by the Natural Sciences & Engineering Research Council Discovery Grant [grant number 227158 to C.H.L.].

Acknowledgments

We thank all members in the Lee Laboratory for their insightful scientific discussion and valuable input.

Competing Interests

The Authors declare that there are no competing interests associated with the manuscript.

References

References
1
Bell
,
J.L.
,
Wächter
,
K.
,
Mühleck
,
B.
,
Pazaitis
,
N.
,
Köhn
,
M.
,
Lederer
,
M.
et al. 
(
2013
)
Insulin-like growth factor 2 mRNA-binding proteins (IGF2BPs): post-transcriptional drivers of cancer progression?
Cell. Mol. Life Sci.
70
,
2657
2675
2
Degrauwe
,
N.
,
Suvà
,
M.-L.
,
Janiszewska
,
M.
,
Riggi
,
N.
and
Stamenkovic
,
I.
(
2016
)
IMPs: an RNA-binding protein family that provides a link between stem cell maintenance in normal development and cancer
.
Genes Dev.
30
,
2459
2474
3
Chao
,
J.A.
,
Patskovsky
,
Y.
,
Patel
,
V.
,
Levy
,
M.
,
Almo
,
S.C.
and
Singer
,
R.H.
(
2010
)
ZBP1 recognition of β-actin zipcode induces RNA looping. ZBP1 recognition of β-actin zipcode induces RNA looping
.
Genes Dev.
24
,
148
158
4
Wang
,
G.
,
Huang
,
Z.
,
Liu
,
X.
,
Huang
,
W.
,
Chen
,
S.
,
Zhou
,
Y.
et al. 
(
2016
)
IMP1 suppresses breast tumor growth and metastasis through the regulation of its target mRNAs
.
Oncotarget
7
,
15690
15702
5
Barnes
,
M.
,
Rensburg
,
G.V.
,
Li
,
W.-M.
,
Mehmood
,
K.
,
Mackedenski
,
S.
,
Chan
,
C.-M.
et al. 
(
2015
)
Molecular insights into the coding region determinant-binding protein-RNA interaction through site-directed mutagenesis in the heterogeneous nuclear ribonucleoprotein-K-homology domains
.
J. Biol. Chem.
290
,
625
639
6
Wächter
,
K.
,
Köhn
,
M.
,
Stöhr
,
N.
and
Hüttelmaier
,
S.
(
2013
)
Subcellular localization and RNP formation of IGF2BPs (IGF2 mRNA-binding proteins) is modulated by distinct RNA-binding domains
.
Biol. Chem.
394
,
1077
1090
7
Mehmood
,
K.
,
Akhtar
,
D.
,
Mackedenski
,
S.
,
Wang
,
C.
and
Lee
,
C.H.
(
2016
)
Inhibition of GLI1 expression by targeting the CRD-BP-GLI1 mRNA interaction using specific oligonucleotide
.
Mol. Pharmacol.
89
,
606
617
8
Rensburg
,
G.V.
,
Mackedenski
,
S.
and
Lee
,
C.H.
(
2017
)
Characterizing the coding region determinant-binding protein (CRD-BP)-microphthalmia-associated transcription factor (MITF) mRNA interaction
.
PLoS ONE
12
,
e0171196
9
Gutschner
,
T.
,
Hämmerle
,
M.
,
Pazaitis
,
N.
,
Bley
,
N.
,
Fiskin
,
E.
,
Uckelmann
,
H.
et al. 
(
2014
)
Insulin-like growth factor 2 mRNA-binding protein 1 (IGF2BP1) is an important protumorigenic factor in hepatocellular carcinoma
.
Hepatology
59
,
1900
1911
10
Tessier
,
C.R.
,
Doyle
,
G.A.
,
Clark
,
B.A.
,
Pitot
,
H.C.
and
Ross
,
J.
(
2004
)
Mammary tumor induction in transgenic mice expressing an RNA-binding protein
.
Cancer Res.
64
,
209
214
11
Hamilton
,
K.E.
,
Noubissi
,
F.K.
,
Katti
,
P.S.
,
Hahn
,
C.M.
,
Davey
,
S.R.
,
Lundsmith
,
E.T.
et al. 
(
2013
)
IMP1 promotes tumor growth, dissemination, and a tumor-initiating cell phenotype in colorectal cancer cell xenografts
.
Carcinogenesis
34
,
2647
2654
12
Liao
,
B.
,
Patel
,
M.
,
Hu
,
Y.
,
Charles
,
S.
,
Herrick
,
D.J.
and
Brewer
,
G.
(
2004
)
Targeted knockdown of the RNA-binding protein CRD-BP promotes cell proliferation via an insulin-like growth factor II-dependent pathway in human K562 leukemia cells
.
J. Biol. Chem.
279
,
48716
48724
13
Gu
,
W.
,
Pan
,
F.
and
Singer
,
R.H.
(
2009
)
Blocking beta-catenin binding to the ZBP1 promoter represses ZBP1 expression, leading to increased proliferation and migration of metastatic breast cancer cells
.
J. Cell Sci.
122
,
1895
1905
14
Gu
,
W.
,
Katz
,
Z.
,
Wu
,
B.
,
Park
,
H.Y.
,
Li
,
D.
Lin
,
S.
et al. 
(
2012
)
Regulation of local expression of cell adhesion and motility-related mRNAs in breast cancer cells by IMP1/ZBP1
.
J. Cell Sci.
125
,
81
91
15
Nwokafor
,
C.U.
,
Sellers
,
R.S.
and
Singer
,
R.H.
(
2016
)
IMP1, an mRNA binding protein that reduces the metastatic potential of breast cancer in a mouse model
.
Oncotarget
7
,
72662
72671
16
Hamilton
,
K.E.
,
Chatterji
,
P.
,
Lundsmith
,
E.T.
,
Andres
,
S.F.
,
Giroux
,
V.
,
Hicks
,
P.D.
et al. 
(
2015
)
Loss of stromal IMP1 promotes a tumorigenic microenvironment in the colon
.
Mol. Cancer Res.
13
,
1478
1486
17
Bernstein
,
P.L.
,
Herrick
,
D.J.
,
Prokipcak
,
R.D.
and
Ross
,
J.
(
1992
)
Control of c-myc mRNA half-life in vitro by a protein capable of binding to a coding region determinant
.
Genes Dev.
6
,
642
654
18
Ioannidis
,
P.
,
Mahaira
,
L.G.
,
Perez
,
S.A.
,
Gritzapis
,
A.D.
,
Sotiropoulou
,
P.A.
,
Kavalakis
,
G.J.
et al. 
(
2005
)
CRD-BP/IMP1 expression characterizes cord blood CD34+ stem cells and affects c-myc and IGF-II expression in MCF-7 cancer cells
.
J. Biol. Chem.
280
,
20086
20093
19
Noubissi
,
F.K.
,
Elcheva
,
I.
,
Bhatia
,
N.
,
Shakoori
,
A.
,
Ougolkov
,
A.
,
Liu
,
J.
et al. 
(
2006
)
CRD-BP-mediates stabilization of βTrCP1 and c-myc mRNA in response to β-catenin signaling
.
Nature
441
,
898
901
20
Weidensdorfer
,
D.
,
Stohr
,
N.
,
Baude
,
A.
,
Lederer
,
M.
,
Kohn
,
M.
,
Schierhorn
,
A.
et al. 
(
2008
)
Control of c-myc mRNA stability by IGF2BP1-associated cytoplasmic RNPs
.
RNA
15
,
104
115
21
Vikesaa
,
J.
,
Hansen
,
T.V.O.
,
Jønson
,
L.
,
Borup
,
R.
,
Wewer
,
U.M.
,
Christiansen
,
J.
et al. 
(
2006
)
RNA-binding IMPs promote cell adhesion and invadopodia formation
.
EMBO J.
25
,
1456
1468
22
Noubissi
,
F.K.
,
Goswami
,
S.
,
Sanek
,
N.A.
,
Kawakami
,
K.
,
Minamoto
,
T.
,
Moser
,
A.
et al. 
(
2009
)
Wnt signaling stimulates transcriptional outcome of the hedgehog pathway by stabilizing GLI1 mRNA
.
Cancer Res.
69
,
8572
8578
23
Conway
,
A.E.
,
Nostrand
,
E.L.V.
,
Pratt
,
G.A.
,
Aigner
,
S.
,
Wilbert
,
M.L.
,
Sundararaman
,
B.
et al. 
(
2016
)
Enhanced CLIP uncovers IMP protein-RNA targets in human pluripotent stem cells important for cell adhesion and survival
.
Cell Rep.
15
,
666
679
24
Mongroo
,
P.S.
,
Noubissi
,
F.K.
,
Cuatrecasas
,
M.
,
Kalabis
,
J.
,
King
,
C.E.
,
Johnstone
,
C.N.
et al. 
(
2011
)
IMP-1 displays cross-talk with K-Ras and modulates colon cancer cell survival through the novel proapoptotic protein CYFIP2
.
Cancer Res.
71
,
2172
2182
25
Bates
,
S.E.
(
2015
)
Targeting RAS: the elusive prize
.
Clin. Cancer Res.
21
,
1796
26
Liu
,
K.
,
Zhang
,
C.
,
Li
,
T.
,
Ding
,
Y.
,
Tu
,
T.
,
Zhou
,
F.
et al. 
(
2015
)
Let-7a inhibits growth and migration of breast cancer cells by targeting HMGA1
.
Int. J. Oncol.
46
,
2526
2534
27
Kato
,
T.
,
Hayama
,
S.
,
Yamabuki
,
T.
,
Ishikawa
,
N.
,
Miyamoto
,
M.
,
Ito
,
T.
et al. 
(
2007
)
Increased expression of insulin-like growth factor-II messenger RNA-binding protein 1 is associated with tumour progression in patients with lung cancer
.
Clin. Cancer Res.
13
,
434
442
28
Cogoi
,
S.
,
Zorzet
,
S.
,
Rapozzi
,
V.
,
Géci
,
I.
,
Pedersen
,
E.B.
and
Xodo
,
L.E.
(
2013
)
MAZ-binding G4-decoy with locked nucleic acid and twisted intercalating nucleic acid modifications supresses KRAS in pancreatic cancer cells and delays tumor growth in mice
.
Nucleic Acids Res.
41
,
4049
4064
29
King
,
D.T.
,
Barnes
,
M.
,
Thomsen
,
D.
and
Lee
,
C.H.
(
2014
)
Assessing specific oligonucleotides and small molecule antibiotics for the ability to inhibit the CRD-BP-CD44 RNA interaction
.
PLoS ONE
9
,
e91585
30
Chen
,
C.
,
Ridzon
,
D.A.
,
Broomer
,
A.J.
,
Zhou
,
Z.
,
Lee
,
D.H.
,
Nquyen
,
J.T.
et al. 
(
2005
)
Real-time quantification of microRNAs by stem-loop RT-PCR
.
Nucleic Acids Res.
33
,
e179
31
Shen
,
W.
,
Han
,
Y.
,
Huang
,
B.
,
Qi
,
Y.
,
Xu
,
L.
,
Guo
,
R.
et al. 
(
2015
)
MicroRNA-483-3p inhibits extracellular matrix production by targeting Smad4 in human trabecular meshwork cells
.
Invest. Opthalmol. Vis. Sci.
56
,
8419
8427
32
Lewis
,
H.A.
,
Musunuru
,
K.
,
Jensen
,
K.B.
,
Edo
,
C.
,
Chen
,
H.
,
Darnell
,
R.B.
et al. 
(
2000
)
Sequence-specific RNA binding by a Nova KH domain: implications for paraneoplastic disease and the fragile X syndrome
.
Cell
100
,
323
332
33
Lin
,
Q.
,
Taylor
,
S.J.
and
Shalloway
,
D.
(
1997
)
Specificity and determinants of Sam68 RNA binding
.
J. Biol. Chem.
272
,
27274
27280
34
Zhou
,
Y.
,
Mah
,
T.-F.
,
Greenblatt
,
J.
and
Friedman
,
D.I.
(
2002
)
Evidence that the KH RNA-binding domains influence the action of the E. coli NusA protein
.
J. Mol. Biol.
318
,
1175
1188
35
Paziewska
,
A.
,
Wyrwicz
,
L.C.
,
Bujnicki
,
J.M.
,
Bomsztyk
,
K.
and
Ostrowski
,
J.
(
2004
)
Cooperative binding of the hnRNP K three KH domains to mRNA targets
.
FEBS Letts.
577
,
134
140
36
Hollingworth
,
D.
,
Candel
,
A.M.
,
Nicastro
,
G.
,
Martin
,
S.R.
,
Briata
,
P.
,
Gherzi
,
R.
et al. 
(
2012
)
KH domains with impaired nucleic acid binding as a tool for functional analysis
.
Nucleic Acids Res.
40
,
6873
6886
37
Sparanese
,
D.
and
Lee
,
C.H.
(
2007
)
CRD-BP shields c-myc and MDR-1 RNA from endonucleolytic attack by a mammalian endoribonuclease
.
Nucleic Acid Res.
35
,
1209
1221
38
Boyerinas
,
B.
,
Park
,
S.-M.
,
Shomron
,
N.
,
Hedegaard
,
M.M.
,
Vinther
,
J.
,
Andersen
,
J.S.
et al. 
(
2008
)
Identification of let-7-regulated oncofetal genes
.
Cancer Res.
68
,
2587
2591
39
Ohdaira
,
H.
,
Sekiguchi
,
M.
,
Miyata
,
K.
and
Yoshida
,
K.
(
2012
)
MicroRNA-94 suppresses cell proliferation and induces senescence in A549 lung cancer cells
.
Cell Prolif.
45
,
32
38
40
Rebucci
,
M.
,
Semeus
,
A.
,
Leonard
,
E.
,
Delaive
,
E.
,
Dieu
,
M.
,
Fransolet
,
M.
et al. 
(
2015
)
miRNA-196b inhibits cell proliferation and induces apoptosis in HepG2 cells by targeting IGF2BP1
.
Mol. Cancer
14
,
79
41
Kamran
,
F.
,
Andrade
,
A.C.
,
Nella
,
A.A.
,
Clokie
,
S.J.
,
Rezvani
,
G.
,
Nilsson
,
O.
et al. 
(
2015
)
Evidence that up-regulation of microRNA-29 contributes to postnatal body growth deceleration
.
Mol. Endocrinol.
29
,
921
932
42
Johnson
,
S.M.
,
Grosshans
,
H.
,
Shingara
,
J.
,
Byrom
,
M.
,
Jarvis
,
R.
,
Cheng
,
A.
et al. 
(
2005
)
RAS is regulated by the let-7 microRNA family
.
Cell
120
,
635
647
43
Li
,
W.M.
,
Chan
,
C.-M.
,
Miller
,
A.L.
and
Lee
,
C.H.
(
2017
)
Dual functional roles of molecular beacon as microRNA detector and inhibitor
.
J. Biol. Chem.
292
,
3568
3580
44
Runge
,
S.
,
Nielsen
,
F.C.
,
Nielsen
,
J.
,
Lykke-Andersen
,
J.
,
Wewer
,
U.M.
and
Christiansen
,
J.
(
2000
)
H19 RNA binds four molecules of insulin-like growth factor II mRNA-binding protein
.
J. Biol. Chem.
275
,
29562
29569
45
Nielsen
,
J.
,
Kristensen
,
M.A.
,
Willemoes
,
M.
,
Nielsen
,
F.C.
and
Christiansen
,
J.
(
2004
)
Sequential dimerization of human zipcode-binding protein IMP1 on RNA: a cooperative mechanism providing RNP stability
.
Nucleic Acids Res.
32
,
4368
4376
46
Li
,
Y.-M.
,
Franklin
,
G.
,
Cui
,
H.-M.
,
Svensson
,
K.
,
He
,
X.-B.
,
Adam
,
G.
et al. 
(
1998
)
The H19 transcript Is associated with polysomes and May regulate IGF2 expression in trans
.
J. Biol. Chem.
273
,
28247
28252
47
Zuker
,
M.
(
2003
)
Mfold web server for nucleic acid folding and hybridization prediction
.
Nucleic Acids Res.
15
,
3406
3415

Author notes

*

These authors contributed equally to this work.