Ensembl Genome Browser (https://www.ensembl.org/index.html) was used to annotate the exon limits of the investigated isoforms and to observe their location in comparison to the HR-DSCR limits. BLASTN software was used to align exon junction sequences (used as query sequence) with RNA-Seq experiments selected as explained below, in order to evaluate tissue isoform expression. For reads obtained by the bio-project “Illumina HiSeq 2500 of 4 human trisomy 21 and 4 human normal control blood cells” (PRJNA635385), recently annotated on Gene, we used local SRA-BLAST (https://trace.ncbi.nlm.nih.gov/Traces/sra/sra.cgi?view=software) to perform the alignments. We considered only alignments with at least 95% of query cover and 97% of identity as significative.

RNA-Seq experiments were selected on KCNJ6 (https://www.ncbi.nlm.nih.gov/gene/3763) and DSCR4 (https://www.ncbi.nlm.nih.gov/gene/10281) NCBI Gene schedule under the “Expression” heading. The three bio-projects selected were: “HPA RNA-seq normal tissues (PRJEB4337)”; “RNA sequencing of total RNA from 20 human tissues (PRJNA280600)”; “Illumina bodyMap2 transcriptome (PRJEB2445)”; we excluded “Tissue-specific circular RNA induction during human foetal development (PRJNA270632)” because it included experiments performed on circular RNA during the fetal phase. We included the bio-project “Illumina HiSeq 2500 of 4 human trisomy 21 and 4 human normal control blood cells” (PRJNA635385), in which RNA-Seq was performed on RNA samples extracted following the same procedures and conditions of RNA used in this work for in vitro analyses (Antonaros et al., 2021b).

We considered RNA samples obtained by tissues (both euploid and T21) involved in the DS phenotype and which were available in our laboratory to confirm computational data with molecular biology experiments.

The euploid tissues selected from the four bio-projects were: adrenal gland, brain, cerebellum, cerebral cortex, heart, liver, placenta, skeletal muscle, skin, testis, thymus, thyroid, white blood cells and blood cells. The T21 tissue selected was T21 blood cells (see Supplementary Table S1).

For the molecular biology experiments, we used commercially available RNA samples (Clontech, Mountain View, CA), derived from the following human euploid tissues: adrenal gland, brain, cerebellum, cerebral cortex, heart, liver, placenta, skeletal muscle, testis, thymus, thyroid. The main characteristics of RNA donors are listed in Supplementary Data Set S1. The results reported by the in vitro analyses were compared with those reported by the RNA-Seq experiments carried out on the same tissues and listed above.

It was possible to detect the expression of KCNJ6 and DSCR4 both in euploid and trisomic tissues only for RNA derived from fibroblast and blood cells obtained as follows.

Normal and T21 blood samples were collected at the Neonatology Unit of S. Orsola-Malpighi Hospital of Bologna, in the context of the “Genotype-phenotype correlation in trisomy 21 (Down syndrome)” project (for ethical approval, see “Declarations” section below). Blood samples were collected in ethylenediaminetetraacetic acid (EDTA)-coated blood collection tubes, and plasma fraction was isolated within 2 h from blood collection according to our previously published work (Antonaros et al., 2021a). Five mL of denaturing solution were added to the remaining blood fraction and it was stored at -20°C until RNA extraction. The main features of donors are listed in Supplementary Data Set S1. The results reported by the in vitro analyses were compared with those reported by the RNA-Seq experiments (PRJNA635385) performed on normal and T21 blood samples.

Normal and T21 primary fibroblast cell lines were provided by Galliera Genetic Bank (GGB), member of the “Network Telethon of Genetic Biobanks.” All the cell lines were tested for mycoplasma to exclude possible contamination. Furthermore, a karyotype analysis was carried out on T21 cell lines by GGB to confirm the cytogenetic diagnosis. Five mL of denaturing solution was added to the cell flasks and they were stored at −20°C until RNA extraction. The main features of donors are listed in Supplementary Data Set S1. The results reported by the in vitro analyses performed on euploid fibrobast cell lines were compared with those reported by the RNA-Seq experiments (PRJEB4337) performed on normal skin tissues.

RNA extraction from blood samples and fibroblast cell lines was performed with the method of Chomczynski and Sacchi (Chomczynski and Sacchi, 1987), and RNA quantity and quality were verified through electrophoresis on agarose gel (visualization and quantification with GelDoc 2000 and Quantity One software, Bio-Rad Laboratories, Hercules, CA, United States) and through Nanodrop spectrophotometer (ND-1000 spectrophotometer, Thermo Fisher Scientific, Waltham, MA, United States).

Primer pairs were designed with “Amplify 3” program (Engels, 1993) following the standard criteria described by Sharrocks (Sharrocks, 1994). Briefly, primers should have an annealing temperature (T_a) between 55°C and 65°C and the difference between the two temperatures should be less than 2°C. Guanines (Gs) and Cytosines (Cs) content should be around 40–60% and it is recommended to avoid secondary structures such as hairpins, self-dimers and cross-dimers which could be caused by self-complementarity of forward and reverse primers. The primers must have biological specificity to avoid non-specific amplifications within the genome (Ye et al., 2012). Primer BLAST software (https://www.ncbi.nlm.nih.gov/tools/primer-blast/) was used to evaluate the biological specificity by comparing the primer pair sequences versus the entire collection of nucleotides (nr/nt). The primer sequences and characteristics were reported in Supplementary Table S2.

Pools of RNA samples were created from RNA derived from T21 and normal control blood cells: one with three T21 blood cell RNA samples (330 ng of each) and the second with three normal control blood cell RNA samples (330 ng of each) (see Supplementary Data Set S1). Pools of RNA samples were created from RNA from T21 and normal control fibroblasts: one with three T21 fibroblast RNA samples (330 ng of each) and the second with three normal control fibroblast RNA samples (330 ng of each) (see Supplementary Data Set S1).

Reverse transcription (RT) was conducted on 1 µg of RNA (in a total volume of 20 µL). Complementary DNA (cDNA) was obtained using “SuperScript III First-strand Synthesis Supermix” kit (Invitrogen by Life technologies, Grand Island, NY, United States) according to manufacturer instructions: 10 min (minutes) at 25°C, 30 min at 50°C, 5 min at 85°C. cDNA obtained from RT was amplify through PCR to detect KCNJ6 and DSCR4 expression. Each 25 µL PCR reaction contained 2.5 µL of cDNA, 1 unit (U) of Platinum Taq DNA Polymerase (Invitrogen; Thermo Fisher Scientific), 2.5 µL of PCR buffer (10x), 25 mM of MgCl2 (final 1.5 mM), 5 mM dNTPs mix (final 0.2 mM), 0.2 µM of each primer. The mixture was denatured at 94°C for 5 min and the PCR was performed for 45 cycles under the following conditions: denaturation at 94°C for 30 seconds (s), annealing at 58–62°C (depending on primer features) for 30 s or 1 min (30 s for transcripts <1,000 bp, 1 min for transcripts >1,000 bp), and extension at 72°C for 1 min. The final extension cycle of 72°C was for 10 min. RT reactions were performed using MJ Research PTC-200 Thermal Cycler and PCR reactions were performed using GenePro TC-E-48D Thermal Cycler.

Amplification products were analyzed by electrophoresis on agarose gel and by Sanger sequencing. The RT-PCR products were purified with GenElute PCR Clean-up Kit (Sigma-Aldrich, Saint Louis, MO, United States) and prepared for Sanger sequencing. The mixture was prepared following BigDye Kit protocol (Applied Biosystems, Carlsbad, CA, United States): 10 ng of amplicon, 0.32 µM of forward or reverse primer, 2 µL of PCR buffer, 0.5–1 µL of BigDye (0.5 µL for transcripts <500 bp, 1 µL for transcripts ≥ 500 bp), and sterile water up to 10 µL of final volume. Automated sequencing was performed with Applied Biosystems ABI 3730 DNA.

Finally, we compared the results obtained by RT-PCR, SRA-BLAST alignments and data available on Ensembl to determine the structure of KCNJ6 and DSCR4 transcripts.

Figure 1 shows a graphical representation of KCNJ6-201 and KCNJ6-202 isoforms and Table 1 reports the exon coordinates of each transcript. The first and second exons of KCNJ6-202 are present only in this isoform and, in this work, they are named as Exon (E)1b and E1bis (see Figure 1 and Table 1). The third and the fourth exons of KCNJ6-202 correspond to the second and third exons of KCNJ6-201 (named as E2 and E3 in both isoforms; see Figure 1 and Table 1). The fifth exon of KCNJ6-202 (called E4b; see Figure 1 and Table 1) is identical, for the first part of the sequence, to the fourth exon of KCNJ6-201 (called E4; see Figure 1 and Table 1). KCNJ6-202 transcript includes the HR-DSCR in a long intron located between the second and third exons (E1bis and E2; see Figure 1 and Table 1).

Regarding DSCR4 gene isoforms to date, DSCR4-201, DSCR4-202 and DSCR4-203 are annotated in the genome databases. The three isoforms share the first and second exons of which the first is similar to the first exon of KCNJ6-202 (E1b; see Figure 1 and Table 1), and the second exon is totally shared with KCNJ6-202 (E1bis; see Figure 1 and Table 1). DSCR4-201, the only DSCR4 isoform registered on National Center for Biotechnology Information (NCBI) Gene database, was discovered in 1997 via Expressed Sequence Tag (EST) (Nakamura et al., 1997) and it has three exons (see Figure 1 and Table 1). DSCR4-202 transcript is described as long non-coding-RNA (lncRNA), and it has 4 exons of which the third and the fourth are included in the HR-DSCR (see Figure 1 and Table 1). The longest isoform of DSCR4 gene is DSCR4-203, which has 6 exons (see Figure 1 and Table 1) and it is described as lncRNA.

Finally, BLASTN software was used to align exon junction sequences (used as query sequence) with RNA-Seq experiments. The exon junction sequence of 50 bp was used as query sequence, and it was generated by the last 25 nucleotides (nt) of the previous exon (e.g., exon 1) and the first 25 nt of the next exon (e.g., exon 2). KCNJ6-201 and KCNJ6-202 shared 2 exon junctions, one generated by the last 25 nt of the E2 and the first 25 nt of the E3, and the second generated by the last 25 nt of the E3 and the first 25 nt of the E4 (see Figure 1 and Table 1). KCNJ6-202 and the three isoforms of DSCR4 shared 1 exon junction generated by the last 25 nt of the E1 (or E1b) and the first 25 nt of the E2 (or E1bis) (see Figure 1 and Table 1).

We analyzed 86 RNA-Seq experiments selected following the parameters mentioned above (Material and Methods Section 2.3). Exon junction sequences of each transcript were listed in the respective Supplementary Tables: KCNJ6-201 (see Supplementary Table S3); KCNJ6-202 (see Supplementary Table S4); DSCR4-201 (see Supplementary Table S5); DSCR4-202 (see Supplementary Table S6); DSCR4-203 (see Supplementary Table S7) were compared to RNA-Seq experiments using SRA-BLAST.

Reverse Transcription-Polymerase chain reaction (RT-PCR) was used to perform KCNJ6 and DSCR4 expression profiles in several tissues (see Supplementary Data Set S1).

The entire KCNJ6-201 transcript, amplified by primer pairs on E1-E4 (see Supplementary Table S2), was detected in adrenal gland, brain, cerebellum, cerebral cortex, heart, placenta, skeletal muscle, skin, thymus, and thyroid (see Figure 2). We obtained part of the transcript in several tissues (for more details see Supplementary Table S8a). Sanger sequencing in brain confirmed PCR results.

The entire KCNJ6-202 transcript, amplified by primer pairs on E1b-E4 (see Supplementary Table S2), was detected in brain, cerebellum, and placenta (see Figure 3). We obtained part of the transcript in a few tissues (for more details see Supplementary Table S8b). Sanger sequencing in placenta confirmed PCR results. The sequence exactly between primer pair has been deposited in NCBI GenBank under the accession number OK050111 (https://www.ncbi.nlm.nih.gov/nuccore/OK050111).

The entire DSCR4-201 transcript, amplified by primer pairs on E1-E3 (see Supplementary Table S2), was detected in brain, placenta, testis, and thyroid (Figure 4; Supplementary Table S8c). Sanger sequencing in thyroid confirmed PCR results.

The entire DSCR4-202 transcript was not detected in any tissues (see Supplementary Table S8d), but was detected the expression of E3-E4 (located in the HR-DSCR) in placenta and testis confirmed by Sanger sequencing in both tissues (see Figure 5). The sequence exactly between primer pair has been deposited in NCBI GenBank under the accession number OK318455 (https://www.ncbi.nlm.nih.gov/nuccore/OK318455).

The entire or partial DSCR4-203 transcript was not detected in any tissues (see Supplementary Table S8e).