1
Supplementary information for Materials and Methods
Human Breast Tissue Microarray Construction.Specimens from paraffin-embedded formalin-fixed blocks were placed on two tissue microarray slides, each containing node-negative and node-positive cores, as well as normal breast controls. Follow-up was available for 95% of the patients, from 2.4 months to 53.8 years, with a median follow-up time of 9.0 years. Complete treatment information was not available for the entire cohort; however most patients were treated with local irradiation. None of the node-negative patients were given adjuvant systemic therapy. Among the node-positive patients, approximately 15% were given chemotherapy and approximately 27% were given tamoxifen (post 1978).
Cell Culture and Protein Fractionation. BT20, SKBR3, and MDA-MB-231 human breast carcinoma cells, which constitutively express c-fms, were maintained in MEM (ATCC) supplemented with 0.1 mM non-essential amino acids, 2 mM L-glutamine, 1 mM sodium pyruvate, 1.5 g/L sodium bicarbonate, and 10% fetal calf serum (Invitrogen, Carlsbad, CA) in5% CO2 at 37°C. When the cultures reached 75% confluence, they were starved in serum-free DMEM/ F12 Ham medium (Sigma-Aldrich, St. Louis, MO) for 72 h, washed twice in Dulbecco’s phosphate-buffered saline (Invitrogen, Carlsbad, CA) and collected with a cell scraper. Nuclear and cytoplasmic fractions were prepared from the cells using NE-PER Reagent (Pierce, Rockford, IL) or by the Dignam et al. method (Dignam et al., 1983), aliquoted and stored at -80˚C. Total cellular protein extract was prepared from cells using 50 mM Tris-HClpH 7.4, 100 mM NaCl, 2 mM EDTA, 1% Igepal (Sigma-Aldrich, St. Louis, MO), and Protease Inhibitor Cocktail Set 1 at 1:100 dilution (Calbiochem, Gibbstown, NJ). Protein concentrations were determined by BCA assay (Pierce, Rockford, IL), using BSA as the standard.
Subcellular Location of HuR in Breast Cancer Cells. Twenty or 40 μg of proteins extracted from BT20, SKBR3, and MDA-MB-231 cells were loaded per lane on a 4-15% or 4-20% SDS-PAGE gel, electrophoresed, and transferred to a Immune-Blot PVDF Membrane (Bio-Rad, Hercules, CA). Membranes were probed with HuR mAb (19F12, Santa Cruz Biotechnology, Santa Cruz, CA),α-tubulin mAb (TU-02, Santa Cruz Biotechnology, Santa Cruz, CA), HDAC1 mAb (10E2, Santa Cruz Biotechnology, Santa Cruz, CA), and Pan actin mAb (ACTN05, Lab Vision, Fremont, CA). Immunoblot processing and ECL protein detection were performed according to the manufacturer’s instructions using SuperSignal West Pico (Pierce, Rockford, IL) and HRP-conjugated 2˚ Abs (Santa Cruz Biotechnology, Santa Cruz, CA).Signal intensity of the bands on X-ray film wasscanned by Molecular ImagerChemiDoc XRS System(Bio-Rad, Hercules, CA).Data was interpreted by Quantity One software.Signal intensity of HuR was normalized by the actin signal.
Immunohistochemical Analysis of HuR.Tissue microarray slides were incubated at 65˚C overnight and rinsed with xylene and ethanol to remove paraffin. The primary anti-HuR antibody was coated on the slides overnight at 4˚C in a moist chamber. The final concentration of the antibody was 1 µg/ml. Visualization was carried out using a secondary antibody conjugated to HRP and diaminobenzidine tetrahydrochloride as the chromogen.
Statistical Analysis. Survival curves were calculated using the Kaplan-Meier method, with the significance evaluated using the Mantel-Cox Log-rank test. The association between HuR staining scores and clinicopathological parameters including c-fms expression was calculated using the Chi-square test. The independent prognostic significance of the parameters was assessed using the Cox proportional hazards model.
Immunoprecipitation (IP) of Ribonucleoprotein Complexes and Reverse Transcription and Quantitative real-time PCR (qRT-PCR). Briefly, cytoplasmic and nuclear extracts were sonicated twice at 50% amplitude for 15 sec in ice. RNase inhibitor (Fermantas, Ontario, Canada) and RNase-free DNase (Promega, Madison, WI) were added to the cell lysates. Protein A/G PLUS-Agarose beads (50µl, sc-2003, Santa-Cruz Biotech, Santa Cruz, CA) were precoated with 5 µg of mouse IgG (Sigma, St. Louis, MO) or anti-HuR antibody (19F12, sc-56709, Santa-Cruz Biotech, Santa Cruz, CA) for 1 h at RT, washed four times with EMSA buffer (25mM HEPES, 3mM MgCl2, 150mM KCl, 1mM DTT, 5% Glycerol, 0.05% Tween-20), and incubated with cytoplasmic or nuclear extracts for 16 h at 4oC. After extensive washes with EMSA buffer, 170µl RNA-protein complexes were eluted with elution buffer (100mM Tris-HCl, pH 8.0, 10mM EDTA, 1% SDS). Twenty µl of eluates were taken for western blot analysis. The remaining 150ul eluates were incubated with 5M NaCl and 20ug Proteinase K for 1h at 420C, followed by 1 h at 650C to reverse the formaldehyde cross-links. RNA was extracted by using phenol and chloroform, and precipitated in the presence of glycogen. RNA was reverse transcribed by using oligo-dT18 primer and M-MuLV Reverse Transcriptase (Fermantas, Ontario, Canada). Real-time PCR (qPCR) was performed with iQ SYBR Green supermix (Bio-Rad, Hercules, CA) by using the following amplification conditions: 2 min at 950C, then 40 cycles of 950C for 30s, 550C for 15s, and 720C for 1 min, and finally 10 min at 720C.
Primers Used for qPCR. The oligomer pairs used for the amplification of PCR products are listed as follows.
Primer / Forward Sequence / Reverse SequenceNDUFV1 / GCCGAGATCGACTCCCTGTG / TTTTGGGCAGCACTCGCTTT
ProTα / CGCAGCCGTAGACACCAGCTC / GGAGAGCGCGTGTCATCTGCG
c-fms / GGAGTTGACGACAGGGAGT / AACACGAGGCCAACACCAT
HuR / GGATCCTCTGGCAGATGTTTG / CAAACACTTGTGAAAATTGGCGC
Production and Purification of HuR-GST and GST Proteins. Human HuR cDNA in vector pGEX2T and GST control plasmid pGEX4T, were a kind gift from Joan A. Steitz (YaleUniversity, New Haven, CT). Proteins were expressed in E. coli strain BL21(DE3) (Novagen, Gibbstown, NJ) and purified as described (Brennan et al., 2000).
c-fms Constructs and Riboprobe Synthesis. The plasmid pcfms 102 (Genentech, Inc., San Francisco, CA.) was used as the template for PCR amplification of the entire c-fms 3’UTR (776nt; 3217-3992nt) (Coussens et al., 1986) using the forward primer: (5’AAAAGCTT3217TGAGGAGTTGACGACAGGGAGT32383’), and reverse primer: (5’AAGGTACC3992GCATTAATGCTGTTAGTTT39743’). The PCR product was digested by Hind III and Kpn I, inserted into plasmid pGEM-3Z, and the construct sequenced. The same plasmid pcfms 102 was also used as the template for PCR amplification of the terminal 578nt (3415-3992) (Coussens et al., 1986) using the forward primer: (5’CGACCAACCTATGG3415GCCAGTCAGGGGC34273’), and reverse primer: (5’GCAGGATCC3992GCATTAATGCTGTTAG39773’). The 578bp fragment was then inserted into plasmid PCRII by TA Cloning (Invitrogen, Carlsbad, CA) and the construct sequenced. The constructs for c-fms 776nt and c-fms 578nt were linearized with Kpn I and Xho I, respectively, purified, and the 32P labeled riboprobe synthesized with SP6 RNA polymerase (Riboprobe in vitro Transcription Systems, Promega, Madison, WI) according to the manufacturer’s instructions using 32P-UTP (Amersham GE), (20 miCi/ml), 50 μCi per reaction, and gel-purified. To eliminate the terminal 3’ region to generate the 3415-3632nt c-fms riboprobe (218nt), the plasmid was linearized with Bgl II. We narrowed down the region responsible for protein binding by creating riboprobes from different regions within this 218nt sequence. To confirm that the 99nt c-fms 3’ UTR region (3488-3586nt) was a target for HuR protein binding, we generated a riboprobe specific for this sequence by PCR amplification of the 3’ UTR c-fms 3415-3992nt sequence described above using the following primers: forward: (5’-AATGAAGCTTTGCTATGCCAAC-3’), and reverse: (5’-ATTCTCGAGGTATCAGTGTAGC TCCT-3’). This 99nt sequence was cloned into pcDNA3.1(+) (Invitrogen, Carlsbad, CA), and sequenced. The 99nt riboprobe was synthesized as described above except the DNA was linearized with XhoI and T7 RNA polymerase was used in the transcription reaction.
The 3’UTR c-fms 69nt (3499-3567nt) deletion construct was created as follows. Theplasmid pcfms 102 was first used asthe template for PCR amplification of the terminal578nt (3415-3992nt). Deletion of the 69nt from these 578nt c-fms sequences was accomplished by PCR based on the work of Booth and colleagues (Booth et al., 1994) with modifications. Two sets of 5’-overlapping primers were generated: primer 1A (5’- CUACUACUACUAATGGGCCAGTCAGGGGCTG-3’), and primer 1B (5’- GCUCCUGGUUGGCAUAGCAAACACGA-3’); primer 2A (5’- CAACCAGGAGCUACACTGATA-3’), and primer 2B (5’- CAUCAUCAUCAUGCATTAATGCTGTTAGT-3’). The dTs in the overlapping regions were synthesized with dU. The overlaps were designed such that the subsequent joining of complementary fragments results in generation of the desired 509nt sequence. The resulting 509nt deletion mutant was cloned into pAMP1, and sequenced. To generate a riboprobe of 149nt containing the 69nt deletion, the plasmid was linearized with Bgl II, and in vitro transcription performed with T7 RNA polymerase.
The c-fms 3’UTR 69nt (3499-3567nt) wild type and mutants were synthesized as follows with a complementary T7 sequence at the 3’end (Mullis and Faloona, 1987). Mutations were created in each of 3 regions within the 69nt c-fms sequence, with the ‘U’s were mutated to ‘G’s as follows: for 69mt1, 3509CUUCUUU3515 was mutated to 3509CGGCGGG3515; for 69mt2, 3525CUUAUCUUCAU3535 was mutated to 3525CGGAGCGGCAG3535; for 69mt3, 3546UGACUUUAU3554 was mutated to 3546GGACGGGAG3554. For riboprobe generation, T7 sequence was annealed and T7 RNA polymerase (Ambion, Austin, TX) was used.
UV Crosslinking and Label Transfer of HuR with 3’UTR c-fms RNA. The reaction mixtures were digested by addition of 0.5 U RNase T1 (1000 U/l, Ambion, Austin, TX) at 30oC for 20 min. Fourlof 5x sample loading buffer (1 M Tris-HClpH 6.8, 10% SDS, 50% glycerol, 2% Bromophenol blue, 2% Xylene cyanol) was added and heated at 95oC for 2-3 min. To analyze protein-RNA cross-links, reaction mixtures were electrophoresed on 10% SDS-PAGE, and radioactive bands were detected by autoradiography for 2-24 h.
HuR overexpression. The full length HuR cDNA with Kozak sequence (GCCACC) just before the translation initiation site was cloned into pcDNA3.1 and its sequence was confirmed. BT20 and SKBR3 cells were transfected with empty expression plasmid or plasmid expressing HuR. After 72 h of serum starvation in the absence or presence of 100nM dex, cells were collected; total RNA was isolated and analyzed by qRT-PCR. The data (mean±SD), were averaged from several experiments and standardized to the NDUFV1 mRNA internal control. For protein analysis, cytoplasmic lysates were prepared using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Pierce, Rockford, IL).
siRNA Treatment. Transfection of siRNA into BT20 cells was performed in 6-well plates or 28 cm2 dishes. Cells were plated at a density of 8x104 cells/cm2. Transfection with the siRNA duplex HuSi-1 (Dharmacon, Lafayette, CO; sense=GCCUGUUCAGCAGCAUUGGdTdT, antisense=CCAAUGCUGCUGAACAGGCdTdT) which targets HuR mRNA (van der Giessen et al., 2003) or control RISC-Free siRNA (0.1 μM) (Dharmacon, Lafayette, CO) was performed when the cultures were 35-70% confluent. Rescue of changes in myogenic phenotype induced by HuSi-1 siRNA was previously demonstrated by introduction of HuR protein conjugated to a cell permeable peptide (van der Giessen et al., 2003). For all transfections, DharmaFECT-1 siRNA Transfection Reagent was used.
Reporter Constructs for Luciferase Analysis. Both wild type and mutant (described in Fig. 7A) 3’UTR c-fms 69nt (3499-3567nt) were synthesized and cloned into the 3’UTR of Luc-MutE in pcDNA3.1(-) (Baggett et al., 2004).
Cell Culture, Transfection, and Dual Luciferase Activity Analysis. BT20 cells were cultured in MEM as described above. BT20 cells were transfected with siRNA duplex HuSi-1 or control RISC-Free siRNA for 24 h prior to luciferase reporter plasmid transfection. For luciferase analysis, BT20 cells were transiently transfected using Fugene HD (Roche, Indianapolis, IN) with 10ng of reporter, and 5 ng of renilla control plasmid. Dual luciferase-activity assays were performed 48 h after transfection according to the manufacturer's directions (Promega, Madison, WI). The results are represented as mean ±SE from three independent experiments. Luciferase RNA levels were measured by qRT-PCR.
References
Baggett B, Roy R, Momen S, Morgan S, Tisi L, Morse D et al (2004). Thermostability of firefly luciferases affects efficiency of detection by in vivo bioluminescence. Mol Imaging3: 324-32.
Booth PM, Buchman GW, Rashtchian A (1994). Assembly and cloning of coding sequences for neurotrophic factors directly from genomic DNA using polymerase chain reaction and uracil DNA glycosylase. Gene146: 303-8.
Brennan CM, Gallouzi IE, Steitz JA (2000). Protein ligands to HuR modulate its interaction with target mRNAs in vivo. J Cell Biol151: 1-14.
Coussens L, Van Beveren C, Smith D, Chen E, Mitchell RL, Isacke CM et al (1986). Structural alteration of viral homologue of receptor proto-oncogene fms at carboxyl terminus. Nature320: 277-80.
Dignam JD, Lebovitz RM, Roeder RG (1983). Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucleic Acids Res11: 1475-89.
Mullis KB, Faloona FA (1987). Specific synthesis of DNA in vitro via a polymerase-catalyzed chain reaction. Methods Enzymol155: 335-50.
van der Giessen K, Di-Marco S, Clair E, Gallouzi IE (2003). RNAi-mediated HuR depletion leads to the inhibition of muscle cell differentiation. J Biol Chem278: 47119-28.