Cell culture and DNA isolation

The HEK-293 embryonic kidney cell line was cultivated following the guidelines of ATCC®. Next, total DNA was isolated after lysis of the cells in TRI Reagent® (Molecular Research Center, Inc. Cincinnati, OH, USA) and following the manufacturer’s instructions. The concentration and purity of the isolated DNA were assessed spectrophotometrically in a BioSpec-nano Micro-volume UV–Vis Spectrophotometer (Shimadzu, Kyoto, Japan).

tRF selection

We chose to investigate the role of 3′-tRF-CysGCA, a fragment of the mature tRNA-CysGCA, with the following sequence: 5′-UCCGGGUGCCCCCUCCA-3′. This 3′-tRF is recorded in two public repositories, namely tRFdb (ID: tRF-3003a) and MINTbase v2.0 (ID: tRF-17-8871K92) (Kumar et al. 2015; Pliatsika et al. 2018, 2016) and may derive from ten distinct tRNACysGCA genes: TRC-GCA2-1, TRC-GCA2-2, TRC-GCA2-3, TRC-GCA2-4, TRC-GCA4-1, TRC-GCA5-1, TRC-GCA7-1, TRC-GCA8-1, TRC-GCA11-1, and TRC-GCA14-1.

This 3′-tRF was selected among others for several reasons; firstly, it has been reported to interact with AGO proteins (Kumar et al. 2014), thus rendering it possible to have a regulatory role. Additionally, it is one of the most abundant tRFs, as it has been detected in various small RNA sequencing (RNA-seq), with high reads per million (RPM) values, as shown in tRFdb (Kumar et al. 2015). Additionally, this tRF is one of the two most abundant 3′-tRFs deriving from the aforementioned tRNAs, as shown in the MINTbase v2.0 data and presented in Fig. S1. Lastly, overexpression of TRC-GCA2-4 (also known as chr17.trna25), a tRNA gene located on chr17:39154491–39154562 (-), has already been shown to lead to the overexpression of the aforementioned 3′-tRF (Kuscu et al. 2018).

Primer designing and PCR

One primer pair was designed to amplify a gene region that included the TRC-GCA2-4 (tRNACysGCA) gene and its flanking sequences (200 nucleotides before its 5′ end and 192 nucleotides after its 3′ end). The coordinates of the amplified region are chr17:39154299–39154762 (-) (Homo sapiens genome assembly GRCh38.p14 (hg38)). The sense and antisense primers were designed to contain the recognition sequences of the restriction endonucleases SacI and XbaI, respectively. The total DNA isolated from the HEK-293 cell line was used as a template for PCR. The primer pair and the annealing temperature are shown in Table S1. More details about the reaction mixture and the thermal protocol can be found in Supplementary Methodology.

Agarose gel electrophoresis, gel clean-up, and Sanger sequencing

The PCR product generated through the amplification of the aforementioned genomic region was electrophorized in a 2% agarose gel. After visualization in UV light, the band was excised, and DNA was extracted using spin columns (Macherey–Nagel GmbH & Co. KG, Düren, Germany). The concentration of the extracted PCR product was evaluated using a Qubit 2.0 Fluorometer (Invitrogen™, Thermo Fisher Scientific Inc., Carlsbad, CA, USA). Subsequently, the sequence was confirmed by Sanger sequencing.

Plasmid construction, bacteria cell transformation, and plasmid purification

The TOPO™ TA Cloning™ kit (Invitrogen™) was used to ligate the pCR™II-TOPO™ vector with the purified PCR product, following the manufacturer’s guidelines. The ligation mixture was used to transform competent E. coli cells. After transformation, bacteria were spread in Luria Broth (LB) agar selection plates. Next, colony selection was done by blue-white screening. The selected colonies were inoculated in liquid bacteria cell culture. Plasmid purification followed, using spin columns (Macherey–Nagel GmbH & Co. KG, Düren, Germany). Next, the recombinant pCR™II-TOPO™ vector and the PCMV6-Neo (OriGene, Rockville, MD, USA) backbone were digested using SacI and XbaI restriction enzymes (New England Biolabs Ltd., Hitchin, UK). The restricted products were electrophoresed, and both the linearized PCMV6-Neo backbone and the insert for the recombinant pCR™II-TOPO™ vector were cleaned up using spin columns (Macherey–Nagel GmbH & Co. KG, Düren, Germany). The ligation procedure was repeated to ligate the linearized PCMV6-Neo backbone and the PCR product of the TRC-GCA2-4 gene and its flanking sequences. After the construction and clean-up of the recombinant PCMV6-Neo vector, the SmaI restriction enzyme (New England Biolabs Ltd.) was used to linearize the recombinant PCMV6-Neo vector. More details about the reaction mixtures and the thermal protocols can be found in Supplementary Methodology.

Cell transfection and clone selection

To transfect HEK-293 cells, 105 cells were seeded in two wells of a 24-well plate. Lipofectamine 2000 (Invitrogen™) was used to transfect 500 ng of the recombinant linearized PCMV6-Neo vector. To transfect cells, 2 µL of Lipofectamine® 2000 reagent (Invitrogen™) were diluted in 50 µL of Opti-MEM® medium (Invitrogen™). In parallel, 500 ng of the linear recombinant plasmid were diluted in 250 µL of Opti-MEM® medium. The two solutions were then mixed in a 1:1 ratio. Then, the DNA-lipofectamine mixture was incubated at room temperature for 5 min, and 50 µL of the mixture was added to the one well, while the other one served as negative control. Forty-eight (48) hours after transfection, the medium was replaced by a medium containing 600 µg/mL geneticin (G418; AppliChem GmbH, Darmstadt, Germany), to select those cells that had been transfected with the vector. The antibiotic concentration had previously been determined by an antibiotic kill curve, ranging from 0.2 to 0.8 mg/mL G418. After 14 days of selection, each cell that had incorporated the vector was transfected in a well of a 96-well plate. In this way, 3 HEK-293 clones were developed.

Nucleic acid extraction and PCR

Total DNA and RNA were extracted from the HEK-293 parental cell line and the three clones using the TRI Reagent® (Molecular Research Center, Inc.). To ensure the repeatability of the results, two genomic DNA and two total RNA extracts were isolated from each clone, while two genomic DNA and three total RNA extracts were isolated from the parental cell line. After determining the concentration and purity of the nucleic acids spectrophotometrically, the integrity of the RNA extracts was assessed after electrophoresis in a Bioanalyzer 2100 RNA nano chip (Agilent Technologies, Winooski, VT, USA). Next, the DNA extracts were used as template to conduct a PCR assay, to ensure the incorporation of the recombinant vector in the HEK-293 clone genome. The primer annealing temperature is shown in Table S1. More details about the reaction mixture and the thermal protocol can be found in Supplementary Methodology.

In vitro polyadenylation, cDNA synthesis, and real-time quantitative PCR (qPCR)

After ensuring the incorporation of the recombinant plasmid in the HEK-293 genome, each total RNA extract was subjected to in vitro polyadenylation, as previously described (Karousi et al. 2021). Next, the polyadenylated RNA extract was used as a template to conduct first-strand cDNA synthesis, using MML-V reverse transcriptase (Invitrogen™) and an oligo-dT adapter primer, following the manufacturer’s instructions, with the only exception being that the denaturation step was performed at 30 °C; we performed the denaturation at such low temperature in order to ensure that only tRFs and not tRNAs would be denatured and reversely transcribed. The oligo-dT adapter primer sequence was 5′-GCGAGCACAGAATTAATACGACTCACTATAGGTTTTTTTTTTTTVN-3′.

Specific forward primers were designed for 3′-tRF-CysGCA, SNORD43 and SNORD61 (C/D box-containing small nucleolar RNAs 43 and 61, respectively), which served as reference genes for 3′-tRF-CysGCA normalization. The reverse primers were complementary to the oligo-dT adaptor sequence used during reverse transcription (Table S1). Real-time qPCR assays were developed and optimized. For this purpose, a standard curve was generated for each amplicon, using serial cDNA dilutions. The expression levels of 3′-tRF-CysGCA were calculated using the comparative CT (2−∆∆CT) method (Livak and Schmittgen 2001; Schmittgen and Livak 2008). More details about the reaction mixtures and the thermal protocols can be found in Supplementary Methodology.

Poly(A)-RNA selection, library construction, and RNA-seq

Oligo-dT-based magnetic mRNA isolation from 5 µg of each RNA extract (two of each clone and three of the parental cell line) was conducted using NEBNext® Poly(A) mRNA Magnetic Isolation Module (New England Biolabs Ltd.). Each poly(A) enriched extract was used to generate a barcoded sequencing library, using the MGIEasy RNA Directional Library Prep Set (MGI Tech Co. Ltd., Shenzhen, China) and following the manufacturer’s instructions. RNA-seq was conducted in the DNBSEQ-G50 platform (MGI Tech Co. Ltd.), using a DNBSEQ-G50RS High-throughput Sequencing Set providing paired-end 100 chemistry.

Quantitative proteomics using data-independent acquisition (DIA)

Three biological replicates of 106 cells of each clone and the parental cell line were pelleted after washing with PBS and subjected to filter-aided sample preparation (FASP), in order to generate tryptic peptides (Wisniewski et al. 2009). The peptides were cleaned up according to the Sp3 strategy (Hughes et al. 2019), using a 1:1 mix of Sera-Mag™ SpeedBead Carboxylate-Modified [E3] and [E7] Magnetic Particles (Cytiva, Marlborough, MA, USA).

Equal amounts of peptide mixtures were analyzed by liquid chromatography with tandem mass spectrometry (LC–MS/MS) in two technical replicates for each of the three biological replicates. LC–MS/MS analysis was performed by injecting the peptidic eluate directly onto an analytical column (25 cm × 75 µm, 1.9 µm beads, C18 ReproSil AQ, Bruker GmbH, Mannheim, Germany), followed by gradient elution in an UltiMate™ 3000 RSLCnano system (Thermo Scientific™, Thermo Fisher Scientific Inc., Waltham, MA, USA). The separated peptides were ionized and sprayed into the Q Exactive™ HF-X Mass Spectrometer (Thermo Scientific™) through a PepSep Stainless Steel Emitter with Liquid Junction (Bruker GmbH). The mass spectrometer was operated in data-independent acquisition (DIA) mode, followed by data-independent analysis.

Data post-processing, biostatistics, and bioinformatics

Post-processing and bioinformatics analysis of the RNA-seq data was conducted using the Partek® Flow software (Partek Inc., Chesterfield, MO, USA). Specifically, quality control and trimming of low-quality bases were conducted, prior to the alignment of the reads; the reads were aligned to the human genome 38 (GRCh38) using the RNA STAR aligner (Dobin et al. 2013). In order to quantify gene expression, each read was assigned to one of the RefSeq transcripts and normalized. Differential expression analysis between the three clones and the parental cell line was conducted with the gene-specific analysis (GSA) module of Partek® Flow. The metrics of the RNA-seq experiment are shown in Table S2.

Regarding proteomics experiments, the Orbitrap raw data were analyzed in DIA-NN 1.8.1 (Data-Independent Acquisition by Neural Networks) (Demichev et al. 2020), through searching against the Human Proteome (downloaded from UniProt as one protein per gene, 20,583 protein entries, 11/2022) using the library free mode of the software, allowing up to two tryptic missed cleavages and a maximum of three variable modifications/peptide. The search was used with oxidation of methionine residues, N-terminal methionine excision, and acetylation of the protein N-termini set as variable modifications and carbamidomethylation of cysteine residues as fixed modification. A spectral library was created from the DIA runs and used to reanalyze them (double search mode). The match between runs feature was used for all analyses, and the output (precursor) was filtered at 0.01 FDR, and finally, the protein inference was performed on the level of genes using only proteotypic peptides.

The statistical analysis was performed within Perseus (v.1.6.15.0) (Tyanova et al. 2016). The four groups (clone 1, clone 2, clone 3, and parental cell line) were filtered for at least 70% valid (detected) values in at least one of the groups. The remaining missing values were imputed based on the Gaussian distribution. A two-sample Welch’s t-test with a P value of less than 0.05 was performed. The results obtained from these comparisons resulted in three lists, representing the differentially expressed genes between each clone and the parental cell line. The intersection (common proteins) between these three lists was used to create the final differentially expressed protein list. Enrichment analysis for both transcriptomics and proteomics data was conducted using the differentially expressed RNA and protein lists as input in Metascape (https://metascape.org) (Zhou et al. 2019).

Putative 3′-tRF-CysGCA targets were retrieved from the tRFtar, tRFtarget, tRFtars, and tRForest databases (Li et al. 2021; Parikh et al. 2022; Xiao et al. 2021; Zhou et al. 2021). The union of four target lists resulted in the final putative target list. The list of the putative targets, the differentially expressed RNAs list, and the differentially expressed protein list were compared and integrated.

Determination of cellular growth rate

To assess the proliferation rates of both parental and clonal cells, a fundamental technique involving cell counting in a hemacytometer chamber was utilized. In this routine procedure, each potential clone or control cell population was initially seeded in a 24-well plate at an initial concentration of 1 × 105 cells. At specific time intervals of 24, 48, 72, and 96 h, the cells were subjected to trypsinization, followed by washing, resuspension in trypan blue, and subsequent cell counting, using the Countess™ Automated Cell Counter (Invitrogen™).

The sulforhodamine B (SRB) assay was also used as an additional way to ascertain the cell density at every time point. The cells were initially plated into a 96-well plate in triplicate at a concentration of 1 × 105 cells/mL, with each well containing 100 µL of cell suspension. To determine cell density, the SRB assay was employed at four distinct time intervals subsequent to the initial seeding, specifically at 24, 48, 72, and 96 h, as previously described (Karousi et al. 2023). The cell counts and OD values were used to build growth curves for the clonal and parental cells, to assess the growth rate.

Quantification of putative 3′-tRF-CysGCA targets

Each total RNA extract was used as a template to conduct first-strand cDNA synthesis, using MML-V reverse transcriptase (Invitrogen™) and the aforementioned oligo-dT adapter primer, following the manufacturer’s instructions.

Specific primers were designed for thymopoietin (TMPO) transcript variant 1 (also known as LAP2α), endoplasmic reticulum-golgi intermediate compartment 1 (ERGIC1), and FTO alpha-ketoglutarate dependent dioxygenase (FTO), as well as for beta-2-microglobulin (B2M) and hypoxanthine phosphoribosyltransferase 1 (HPRT1), which served as reference genes (Table S1). Real-time qPCR assays were developed, as described above, by using the same reaction and thermal protocols. The expression levels of TMPO transcript variant 1, ERGIC1, and FTO were calculated using the comparative CT (2−∆∆CT) method (Livak and Schmittgen 2001; Schmittgen and Livak 2008).

Reporter plasmid construction, bacteria transformation, and plasmid purification

The psiCHECK-2 vector (Promega, Madison, WI, USA) was subjected to digestion using NotI and XhoI restriction enzymes (New England Biolabs Ltd). To do this, 1 µg of the vector was mixed with 20 units of each enzyme in rCutSmart buffer (New England Biolabs Ltd.). The mixture was then incubated at 37 °C for 60 min, followed by heat inactivation at 65 °C for 20 min. Afterwards, the reaction mixture was loaded onto an agarose gel, and the bands containing the linearized plasmid were carefully cut out and purified using the Monarch® DNA Gel Extraction Kit (New England Biolabs Ltd). The restricted plasmid was then ligated with the double-strand DNA sequence of part of the 3′-UTR of each mRNA of interest (Table S3), using the NEBuilder HiFi DNA Assembly Cloning Kit (New England Biolabs Ltd.). After the transformation of the NEB 5-alpha Competent E. coli cells (New England Biolabs Ltd.) with the recombinant plasmids, bacteria were spread in LB agar selection plates. Colonies were inoculated in liquid culture, and plasmid purification was conducted with ZymoPURE II Plasmid Midiprep Kit (Zymo Research Europe Gmbh, Freiburg, Germany).

Dual luciferase reporter assay

To conduct a dual luciferase assay, 9.5 × 104 ΗΕΚ-293 cells were seeded into individual wells of a 96-well plate. After a 24-h incubation period, 50 ng of each plasmid construct was transfected into their respective wells using the jetPRIME® transfection reagent (Polyplus, Illkirch, France). Following a 2-h incubation, 2 pmol of 3′-tRF-CysGCA mimic were introduced into each well using Lipofectamine RNAiMax (Invitrogen™). The sequence of the mimic was 5′-TCCGGGTGCCCCCTCCA-3′. Additionally, wells were prepared with plasmid constructs and a scrambled sequence for normalization purposes (sequence: 5′-AGCCTCTCGTCGCCCGC-3′), and control wells were set up with only the transfection reagents (mock control). Each reaction was performed in triplicate. Twenty-four (24) hours after the 3′-tRF- CysGCA transfection, luminescent signal quantification of Renilla and firefly luciferases was carried out. The Dual-Glo® Luciferase Assay System (Promega) was utilized for luminescence measurement, and a Cytation 5 Cell Imaging Multimode Reader (BioTek, Agilent Technologies) was used following the standard protocol. A schematic representation of the workflow followed in this work is shown in Fig. S2.

Western blot

The protein concentration for each protein extract deriving from clonal and parental cells was determined by Bradford assay. Four protein extracts (20 µg each) from each cell line were used. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a control. The selection of GAPDH as a reference protein was based on the fact that the levels of this protein did not show any significant alteration as a result of 3′-tRF-CysGCA overexpression, in contrast with actin beta (ACTB), which showed variations in the analyzed proteomics data and was hence excluded as an unsuitable reference in this study. In brief, two primary antibodies were added at a 1:1000 dilution: a mouse monoclonal anti-LAP2α and a horseradish peroxidase (HRP)-conjugated monoclonal anti-GAPDH. A secondary HRP-conjugated goat anti-mouse IgG was used to indirectly detect TMPO isoform alpha (LAP2α), whereas GAPDH was directly detected without using any secondary antibody. Detection of each targeted protein (TMPO isoform alpha and GAPDH) was performed by the enhanced chemiluminescence (ECL) detection system. Subsequently, the X-ray films were scanned, and image analysis was conducted.