The LINC00493 gene was predicted through the ENCODE project. The UCSC Genome Browser [28] shows that LINC00493 is located on human chromosome 20p11.23 and consists of two or three exons, according to mRNA and EST sequence data from RefSeq and Ensembl databases. In contrast to the protein-coding genes, the lncRNA gene annotations tend to have poorly defined boundaries, because of weak conservation, low and tissue-specific expression and lack of characteristic hallmarks of transcription initiation and termination [29]. Therefore, to define the exact structure of the LINC00493 transcript we performed reverse transcription (RT)-PCR and rapid amplification of cDNA ends (RACE) analysis on total RNA from HEK293T, HeLa cell lines and human primary skin fibroblasts. RT-PCR analysis revealed that the LINC00493 transcript consists of two exons and RACE showed the exact 5′ and 3′ cDNA ends (Figure 1A).

According to the Ensembl database, there are two LINC00493 isoforms, ENST00000411646.1 and ENST00000435844.3, which differ by three nucleotides at the beginning of the second exon. Our RT-PCR analysis confirmed that LINC00493 has two isoforms, expressed in all analyzed cell lines. The total length of the short and long isoforms was 500 and 497 bp, respectively. The nucleotide sequences of short and long isoforms were deposited into GenBank under accession numbers MW979249 and MW979250. The difference in the sequencing signal suggests that the major long isoform is expressed at an approximately 3-fold higher level than the minor one.

To identify LINC00493 expression profile, we provided analysis of the FANTOM5 and GTEx expression data. We observed that LINC00493 is highly expressed in most human cell lines and tissues. An expression profile of LINC00493 in 889 human samples from FANTOM5 classified this gene as a housekeeping gene [30]. We confirmed the high widespread expression level of this transcript using RT-qPCR analysis of 11 human cell lines, as well as human primary skin fibroblasts (Figure 1B). The highest expression level was observed in A549, MCF7 and HEK293T cell lines.

LncRNA subcellular localization is closely related to its biological function. Some lncRNAs play a role in a transcriptional regulation through their interaction with chromatin, while others are found in the cytoplasm and affect the post-transcriptional control of gene expression or could be translated into small peptides. We investigated the subcellular localization of the LINC00493 transcript using the soft lysis method. RNA was isolated from cytoplasmic, nuclear and chromatin-bound fractions of cells. To determine the level of the investigated transcript in each fraction we performed RT-qPCR. Our analysis revealed that the LINC00493 transcript is localized predominantly in cytoplasm (Figure 1C). This result suggests that the function of the LINC00493 transcript is not related to the transcription regulation and chromatin binding. The observed results highlight similarities between the investigated lncRNA and mRNAs, such as sequence length, high expression level and accumulation in the cytoplasm.

To determine the function of LINC00493, we analyzed previously published CRISPRi-based data for functional long noncoding RNA loci in human cells [13] and discovered that LINC00493 modified cell growth in a cell-type-specific manner. To confirm the cell-type-specific role of LINC00493, we performed knockdown experiments using RNA interference in three human cell lines: HEK293T, A375 and MDA-MB-231.

LINC00493 knockdown efficiency was about 60–70% (Figure 2A). After knockdown, cell proliferation was measured by MTT assay, and cell migration was examined using wound-healing assay. We revealed that LINC00493 knockdown reduced cell viability in HEK293T and A375 cell lines, while the opposite effect was observed in MDA-MB-231 (Figure 2C). Thus, knockdown experiments confirmed that downregulation of LINC00493 affected cell proliferation activity in a cell-type-specific manner. On the other hand, wound-healing assay revealed that LINC00493 knockdown did not affect cell migration (Figure 2B,D).

According to the Human Protein Atlas [31], LINC00493 contains a small open reading frame that could be translated to a 94/95-amino-acid protein—SMIM26. Using GWIPS-viz [32] and Trips-Viz [33], which provide ribosome profiling data, we confirmed that LINC00493 has a strong ribosome association in the predicted sORF region (Figure 3C). The predicted secondary and tertiary structures of SMIM26, constructed using an improved predictor of protein structure [34], showed that the protein contains two alpha helixes; one of them is supposed to be a transmembrane domain (Figure 3B,D). A number of tools (Phobius, SPOCTOPUS, MEMPACK) predict that SMIM26 is localized in the membrane. Despite this fact, the Human Protein Atlas immunocytochemistry analysis revealed that this protein is localized mainly in the nucleoplasm and partially in the cytosol [31]. However, the function of the protein remains unknown. Comparative analysis of amino-acid and nucleic sequences of LINC00493 reveals that the gene is evolutionarily conserved among mammals both at the RNA and protein level (Figure 3A). This fact supports a potential functional role for SMIM26, while the influence of this protein on cell migration and viability was not shown previously.

Data from knockdown experiments revealed that the LINC00493 gene is important for cell viability. However, it remains unclear whether the observed effect is associated with small protein or with the RNA itself. To investigate separately the effect of SMIM26 protein and the LINC00493 transcript on cell viability, we cloned a full-length LINC00493 cDNA into pcDNA3.1-GFP vector (pcDNA3.1) and obtained LINC00493_WT plasmid. Next, we mutated the SMIM26 start-codon by site-directed mutagenesis to prevent the translation of small protein and obtained the LINC00493_MUT plasmid. Wild-type and mutant constructs, along with an empty pcDNA3.1, were transfected into A375 and HEK293T cells and cell viability was measured using the MTT test. The expression of LINC00493 was increased about a thousand-fold times by qPCR (Figure 4A).

Interestingly, overexpression of WT LINC00493 did not influence the number of cells, while overexpression of LINC00493 with a mutated start-codon demonstrated increased cell viability (Figure 4B,C). This observation suggests that, even in the absence of SMIM26 protein translation, the LINC00493 transcript itself has an effect on cell viability. However, the difference between WT and MUT LINC00493 overexpression indicates that SMIM26 protein could also affect cell viability, but in the opposite manner.