A previously unknown Argonaute 2 variant positively modulates the viability of melanoma cells

In malignant melanoma, a highly aggressive form of skin cancer, many microRNAs are aberrantly expressed contributing to tumorigenesis and progression. Further, deregulation of microRNA processing enzymes, like the miRNA-binding protein Argonaute 2, significantly impacts microRNA function. This study characterizes a novel splice variant of Argonaut 2, AGO2-ex1/3. AGO2-ex1/3 is substantially expressed in different melanoma cell lines and patient-derived tissue samples. It is a mature mRNA, which is translated into an N-terminally truncated Argonaute 2 protein form. Molecular dynamics simulations show that the PAZ, MID, and PIWI domain largely retain their structure in AGO2-ex1/3 and that the truncation of the N-terminus leads to an increased interdomain flexibility. Expression of AGO2-ex1/3 provides a survival advantage for melanoma cells while the knockdown causes significantly reduced proliferation and increases apoptosis. RNA-sequencing revealed that in cells lacking AGO2-ex1/3 expression many miRNA target genes are deregulated, implicating a considerable role of AGO2-ex1/3 for miRNA function. This study inaugurates insights into an important role of a so far unknown splice variant of Argonaute 2 for the miRNA pathway as well as the mechanisms which drive growth and survival of melanoma cells. This knowledge provides the basis for potential new promising therapeutic targets focusing on small RNA-mediated gene regulation in melanoma. Supplementary Information The online version contains supplementary material available at 10.1007/s00018-022-04496-8.


XTT viability assay
Cell proliferation and viability was determined using the "Cell Proliferation Kit II (XTT)" (Roche Diagnostics GmbH, Mannheim, Germany).The assay was started one day after re-transfection with siRNA.1000 (Mel Im, Mel Wei) or 1600 (Mel Juso) transfected cells were seeded in triplicates into a 96-well plate in 100 µl medium without phenol red (PAN-Biotech GmbH, Aidenbach, Germany).The "XTT labeling mixture" was added at day 1, 4, 5, 6 and 7 to adherent cells according to the manufacturer's instructions and measured 4 h after adding the reagent at a "CLARIOStar" plate reader (BMG Labtech GmbH, Ortenberg, Germany) at 490 nm.

Clonogenic assay
For the clonogenic assay, 200 (Mel Wei) or 500 (Mel Im, Mel Juso) cells were seeded in a 6well plate.After seven days, the cells were fixed and stained with 300 μl of a crystal violet solution (23,8 % glutaraldehyde, 36,2 % crystal violet, 10 % PBS).Each condition in each experiment was performed in two replicate wells.The number of colonies relative to the number of seeded cells was determined for each well with the software "CellSens Dimension" (Olympus K.K., Shinjuku, Tokyo, Japan) using color intensity, hue, and saturation.Colony size was defined as cell clusters of at least 40 pixel² (corresponds to approx.50 cells).

Real time cell analysis (RTCA)
For RTCA, the "xCELLigence RTCA SP" real time cell analyzer (ACEA Biosciences Inc., San Diego, CA, USA) was used.For the analysis, 3000 (Mel Im, Mel Wei) or 4800 (Mel Juso) transfected cells in a volume of 100 μl medium were seeded in triplicates into an "E-Plate 96" (ACEA Biosciences Inc., San Diego, CA, USA).The measurement of the "Cell Index" (CI) was performed every 5 min over a period of 8 h and for further 70 h with measurements every 15 min.The CI at a measurement time point x is calculated from the difference of the electrical alternating current resistance (impedance) on the sensor surface at this time and the reference resistance (preincubated culture medium without cells), divided by 15 Ω.For determination of the "Delta Cell Index", the CI at a reference time is subtracted from the reference value 1 and this value is added to the respective CI for each measurement time point.The reference time point was determined individually for each cell line as time point where cell proliferation starts and is set to 0 h and "Delta Cell Index"=1 in the graphs.
Cellular senescence was determined using the "Senescence beta-Galactosidase Staining Kit" (Cell Signaling Technology, Danvers, MA, USA) according to the manufacturer's instructions.

Cell cycle analysis
For an analysis of the cell cycle, cells were fixed and permeabilized with cold 70 % methanol overnight at 4 °C.Subsequently, the cells were diluted in 0.2 % BSA in PBS, treated with 0.1 µg/µl RNAse A at 37 °C for 30 min and stained with propidium iodide (PI) (PromoKine, Pro-moCell GmbH, Heidelberg, Germany) corresponding to the cell number (for 500,000 cells 25 μl of a 1 mg/ml solution) on ice for 30 min.The intensity of the PI signal per cell was measured using the flow cytometer "FACSCalibur TM" (Becton, Dickinson and Company, Franklin Lakes, NJ, USA).The data were analyzed using "FlowJo_v10.8.0" (Becton, Dickinson and Company, Franklin Lakes, NJ, USA).Cell gates were set the same for one cell line.Single cell gates were set as appropriate and were kept the same for associated samples in each case.For analysis of cell cycle phases, the "Dean-Jett-Fox" model was used.Gating and analysis of SubG1 cells was performed using the "FACSCalibur TM " software (Becton, Dickinson and Company, Franklin Lakes, NJ, USA).

Analysis of apoptosis
Apoptosis in the melanoma cells was detected with annexin V-fluorescein isothiocyanate (FITC)-and PI-staining of living cells using the "Annexin V-FITC Detection Kit" (PromoKine, PromoCell GmbH, Heidelberg, Germany) according to the manufacturer's instructions.Analysis was done with the flow cytometer "FACSCaliburTM" (Becton, Dickinson and Company, Franklin Lakes, NJ, USA) and the software "FlowJo_v10.8.0" (Becton, Dickinson and Company, Franklin Lakes, NJ, USA).Cell gates were set the same for each cell line, quadrant gates were kept the same for associated samples.

CRISPR/Cas9-mediated knockout of AGO2 in melanoma cells
For the "clustered regularly interspaced short palindromic repeats" (CRISPR)/Cas9-mediated knockout of AGO2 two plasmids were used described previously by van Eijl et al. [1].The plasmids containing the guide RNAs and Cas9 were subcloned using EcoRV-HF and EcoRI-HF into the pU6-(BbsI)_CBh-Cas9-T2A-mCherry vector which was a gift from Ralf Kuehn (Addgene plasmid #64324; http://n2t.net/addgene:64324;RRID:Addgene_64324) [2].The two guide RNA sequences are complementary to an intronic region upstream of exon 2 of the AGO2 gene and to the beginning of exon 2 resulting in a 58 base pair genomic deletion preventing exon 2 from being incorporated into the AGO2 mRNA.Mel Im and Mel Juso cells were transfected with 2µg of each plasmid and after 24 h red fluorescent cells were sorted to single cells using a "FACS Aria II SORP" cell sorter (Becton, Dickinson and Company, Franklin Lakes, NJ, USA).To verify a successful homozygous knockdown, genomic DNA (gDNA) was isolated from the single cell derived clones using the "QIAamp DNA Mini Kit" (QIAGEN GmbH, Hilden, Germany) according to the manufacturer's instructions.GDNA was amplified via PCR and sequenced.

RNA isolation, DNase I or RNase R digestion and reverse transcription
For RNA isolation of melanoma cells or tissue samples, the "E.Z.N.A Total RNA Kit I" (Omega Bio-Tek, VWR, Germany) was used and the protocol was performed according to the manufacturer's instructions.For DNase digestion during RNA isolation, the "DNase I Digestion Set" (Omega Bio-Tek, VWR, Germany) was used.RNA concentration was determined on a "Nanodrop 200c" spectrophotometer (PeqLab, Thermo Fischer Scientific, Waltham, MA, USA).
Reverse transcription was performed with 1/5 volume of RNase R-digested RNA or 500 ng total RNA using the "SuperScript® II Reverse Transcriptase Kit" and a random p(dN)6 primer (Roche Diagnostics GmbH, Mannheim, Germany) according to the manufacturer's instructions.Remaining RNA was digested with 1 μl RNase A (10 µg/µl solution, Roche Diagnostics GmbH, Mannheim, Germany) for 30 min at 37 °C.For reverse transcription with a poly-adenine specific primer, 4 μg RNA and 2 μl of the primer (CCAGTGAGCAGAGTGACGAG-GACTCGAGCTCAAGCTTTTTTTTTTTTTTTTT, 25 μM) was used.
Cloning of pAGO2-ex1/3HAint and pAGO2-ex1/3GFP For cloning of an AGO2-ex1/3 expression plasmid, exon2 of the AGO2 sequence was removed from pIRESneo-FLAG/HA-AGO2 [4] using the "Phusion Site-Directed Mutagenesis Kit" (Thermo Fischer Scientific, Waltham, MA, USA) according to the manufacturer's instructions and the following 5'-phosphorylated primers: hAGO2-ex3-fwd P-GGAAATCGTG-GAACACATGGTCCAG, hAGO2-ex1-rev P-CGGGGCCGGCTCCC.Polymerase (2 U/μL)" (Thermo Fischer Scientific, Waltham, MA, USA) according to the manual with the primers A+B and C+D, respectively and the AGO2-ex1/3 plasmid described above as template.The resulting products were purified via an agarose gel.1µl of each eluted product was used as template for the next PCR using the primers A+D and a two-step PCR program was performed with 30s 98°C, 32 cycles with 15s 98°C and 100s 72 °C and a final elongation of 15min at 72°C.The resulting PCR product was digested using EcoRI-HF and XbaI and cloned into pcDNA3.1(+)(GeneArt, Thermo Fischer Scientific, Waltham, MA, USA).Cloning resulted in a plasmid containing the AGO2-ex1/3-sequence and an HA-tag after A735 (between E245/Q246) in the AGO2-CDS.
40 µg of protein lysate were separated on a 8.75-15 % SDS-polyacrylamide gel and blotted on an "Amersham™ Hybond™ 0,2 μm" polyvinylidene difluoride membrane (GE Healthcare Life Science Europe GmbH, Freiburg, Germany) in a semi dry process.The following antibodies were used for staining: AGO2 N-terminal epitope (clone 11A9 [6], hybridoma supernatant di- and developed using the "Clarity TM Western ECL Substrate Kit" (Bio-Rad, Hercules, CA, USA) for HRP or the "BCIP/NBT Kit" (Life Technologies, Frederick, MD, USA) for AP according to the manual.The HRP signal was detected using the "Intas ECL Chemocam" gel documentation system (Intas Science Imaging Instruments GmbH, Göttingen, Germany).

Molecular structure and dynamics simulations
A high-resolution X-ray structure of human Argonaute 2 bound to a target RNA (PDB code: 4z4d; [7]) was chosen as basis for the molecular dynamics (MD) simulations.Since the focus of the MD simulations was on the free protein, the RNA was removed from the simulation setup.Four missing loops (amino acid 121-126, 270-275, 296-304, and 822-835) of the protein structure were built with via Modeller 9.25 [8][9][10] using the ModLoop procedure.An N-terminal acetyl group was added to the resulting model, to account for the missing N-terminal residues 1-21 in the PDB file.
The structure of the truncated variant AGO2-ex1/3, which lacks the first N-terminal 77 residues, was prepared by deleting the respective residues from the full length model.For both models, AGO2 and AGO2-ex1/3, two slightly different conformations were created using the alternative side chain conformations present in the PDB file, to ensure independent simulation runs for a better sampling.Histidine protonation states were chosen according to propka [11] from the pdb2pqr server [12].All systems were solvated in a capped octahedral TIP3P [13] water box with at least 15 Å to the periodic bounding box border and electrically neutralized by the addition of chloride counter ions.
Following our established simulation procedure [14], a three-step geometry optimization was performed first, then the systems were heated to 310 K over the course of 2 ns, and subsequently simulated for 1.5 μs with a time step of 2 fs in an NPT ensemble with 1 bar.Amber 20 [15] with the parm14SB [16] parameter set was used for all simulations; all production runs were performed on GPUs [17,18].
To enhance conformational sampling, all four optimized systems were additionally investigated by Gaussian-accelerated MD (GaMD) [19].For this purpose, 10 ns of conventional MD were performed first, followed by 20 ns of boosted MD with statistics, and finally by 1.5 μs production runs.
Root mean square deviations (RMSD) of the protein backbone atoms with respect to the initial structure were calculated via cpptraj from AmberTools21 [20].Visualization and graphical representations were created with VMD [21], graphs from the simulations were created with gnuplot.
From the eight simulations conducted in total, the graphs and analyses are shown for the GaMD runs in the supplementary file 2. The four conventional MD simulations showed very similar results with a slightly smaller overall dynamics, which is due to the less intense sampling compared to the GaMD.
RNA-Seq library preparation, mapping, visualization and data analysis RNA-Seq analysis was performed with two two (Mel Im, Mel Wei) or three independent biological replicates (Mel Juso treated with siAGO2-ex1/3 or siCtrl respectively).RNA was isolated as described above.Library preparation was performed using the "TruSeq® Stranded Total RNA Library Prep Human/Mouse/Rat Kit" according to the manufacturer's instructions (Illumina Inc., San Diego, CA, USA).The resulting libraries were checked for size (200-500 bp) by TapeStation 4200 (Agilent, Santa Clara, CA, USA) using the High-Sensitivity DNA Kit (Agilent, Santa Clara, CA, USA) and concentration by the Qubit 4 Fluorometer (Thermo Fischer Scientific, Waltham, MA, USA).The samples were sequenced on a HiSeq4000 with a paired-end module (Illumina, Inc., San Diego, CA, USA) according to the paired-end RNA sequencing protocols from Illumina.Sequencing was performed from each side of a fragment about 75 bp long with a mean number of 20 million reads per sample.After quality check using FastQC (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/), paired-end reads were aligned to the human reference genome (hg38) using the STAR alignment software (v 2.5.2a)[22].
After mapping, only reads that mapped to a single unique location were considered for further analysis.The mapped reads were then used to generate a count table using the feature-counts software (v 1.4.6-p5)[23].The raw reads were filtered, normalized, and visualized by using by using R (version 4.0.1) or the Integrative Genomics Viewer (https://software.broadinstitute.org/software/igv/, [24]).
DESeq2 package version 1.26.0 [25] was used for logarithmic transformation of the data and for data exploration.Differential expression analysis was done using the DESeq2 standard approach.Adjusted p-values are calculated using the Benjamini-Hochberg method within DESeq2.Gene annotations were added to the result files using Ensemble data [26].

Table 2 :
Primer used for qRT-PCR