*4.4. Screening of the Natural Product Library and Luciferase Assay*

Natural compounds have been used to develop drugs for cancer and infectious diseases, since they are structurally optimized by evolution to serve particular biological functions, and their use in traditional medicine provides insights regarding efficacy and safety [58]. Thus, in this study, we used a natural product library to select natural compounds which can induce TTP expression in cancer cells. A library containing 1019 natural products was provided by the Korea Chemical Bank (http://www.chembank.org/. accessed on 26 January 2017) of the Korea Research Institute of Chemical Technology. MCF-7 cells in culture dishes (100 mm diameter) were co-transfected with the pGL3/TTPp-1343 luciferase reporter construct and pRL-SV40 Renilla luciferase construct using TurboFectTM in vitro transfection reagent (Fermentas, Waltham, MA, USA). After incubation for 24 h, cells were harvested and seeded in 96-well plates at a density of 4 × <sup>10</sup><sup>3</sup> cells per well in 100 μL and cultured with 30 μL of natural compounds diluted in fresh culture media. After further incubation for 24 h, cells were lysed with lysis buffer and mixed with luciferase assay reagent (Promega, Madison, WI, USA). Cells were also treated with the same volume of DMSO to detect luciferase activity induced by the native signal pathway. The chemiluminescent signal was measured using a SpectraMax L Microplate (Molecular Devices, Sunnyvale, CA, USA). Firefly luciferase was normalized to Renilla luciferase in each sample. All luciferase assays reported in this study represent at least three independent experiments, each consisting of three wells per transfection. We selected compounds that induced a greater than two-fold increase in luciferase activity.

#### *4.5. Quantitative Real-Time PCR and Semi-qRT-PCR*

DNase I–treated total RNA (3 μg) was reverse transcribed using oligo-dT and Superscript II reverse transcriptase (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. Quantitative real-time PCR (qRT-PCR) was performed by monitoring in real-time the increase in fluorescence of SYBR Green dye (QIAGEN, Hilden, Germany) using the StepOnePlusTM real-time PCR system (Applied Biosystems, Waltham, MA, USA). Semi-qRT-PCR was performed using Taq polymerase (Solgent, Daejeon, Korea) and the PCR primer pairs (Table 1).


**Table 1.** PCR primers used in this study.

#### *4.6. SDS-PAGE and Immunoblotting*

Proteins were resolved by SDS-PAGE and transferred onto Hybond-P membranes (Amersham Biosciences Inc., Amersham, UK). The membranes were blocked and then probed with appropriate dilutions of the following antibodies: rabbit anti-human TTP (T5327, Sigma, St. Louis, MO, USA) and anti-β-actin (A2228, Sigma, St. Louis, MO, USA). Immunoreactivity was detected using an ECL detection system (Amersham Biosciences Inc., Amersham, UK). Films were exposed at multiple time points to ensure that the images were not saturated.

#### *4.7. RNA Preparation and RNA-Seq*

We performed RNA-Seq on total RNA samples (RIN above 8.5) collected from MDA-MB-231 cells at 4 h after treatment with growth media control, 500 nM DXM-21-P or 500 nM BTM-21-P. Residual DNA from each sample was removed using the RNeasyMinElute-Cleanup Kit (Qiagen, Hilden, Germany). The cDNA library was prepared with 1.0 μg of total RNA using the TrueSeq RNA library Preparation Kit (Illumina, San Diego, CA, USA) following manufacturer's recommendations, followed by paired-end sequencing (2 × 100 bp) using the HiSeq1500 platform (Illumina, San Diego, CA, USA). cDNAs were amplified according to the RNAseq protocol provided by Illumina and sequenced using an Illumina HiSeq 2500 system to obtain 150-bp paired-end reads. The sequencing depth for each sample was >20 million reads. RNA-seq reads were mapped using STAR 2.7.9a [59] to the human genome GRCh38. Gene expression counts were measured using multicov implemented in bedtools [60]. Differentially expressed genes (DEGs) were obtained by comparing groups (Control, Beta, and Dexa) using EdgeR [41]. Genes with false discovery rate (FDR) <0.01 and log2fold change >0.3 were selected as DEGs. The DEGs were clustered using hierarchical clustering implemented in R. Ward's criterion. Pearson's correlation

coefficient was used as a distance measure. A clustering heatmap was drawn using a z-score scaled across samples for each gene. The enriched KEGG pathway terms were obtained from Enrichr software [61].
