Revisiting and corrections to the annotated SRSF3 (SRp20) gene structure and RefSeq sequences from the human and mouse genomes

Abstract

SRSF3 (SRp20) is the smallest member of the serine/arginine (SR)-rich protein family. We found the annotated human SRSF3 and mouse Srsf3 RefSeq sequences are much larger than the detected SRSF3/Srsf3 RNA size by Northern blot. Mapping of RNA-seq reads from various human and mouse cell lines to the annotated SRSF3/Srsf3 gene illustrated only a partial coverage of its terminal exon 7. By 5' RACE and 3' RACE, we determined that SRSF3 gene spanning over 8422 bases and Srsf3 gene spanning over 9423 bases. SRSF3/Srsf3 gene has seven exons with exon 7 bearing two alternative polyadenylation signals (PAS). Through alternative PAS selection and exon 4 exclusion/inclusion by alternative RNA splicing, SRSF3/Srsf3 gene expresses four RNA isoforms. The major SRSF3 mRNA isoform with exon 4 exclusion by using a favorable distal PAS to encode a full-length protein is 1411 nt long (not annotated 4228 nt) and the same major mouse Srsf3 mRNA isoform is only 1295 nt (not annotated 2585 nt). The difference from the redefined RNA size of SRSF3/Srsf3 to the corresponding RefSeq sequence is at the 3' UTR region. Collectively, the redefined SRSF3/Srsf3 gene structure and expression will allow better understanding of SRSF3 functions and its regulations in health and diseases.

Keywords: 3ʹ UTR; 5ʹ UTR; Gene expression; Genome structure; RNA isoforms; SRSF3.

Graphical abstract

Fig. 1

Detection of human SRSF3 and mouse Srsf3 transcripts by Northern blot analysis. Five micrograms of total RNA isolated from CaSki, SiHa, HeLa, W12–20863, C33A, BCBL-1, HEK293T cells and 10 μg of total RNA isolated from mouse epithelial keratinocytes (MEK) at passage 24 (P24) or passage 31 (P31) were separated by electrophoresis in 1% formaldehyde-denaturizing agarose gel together with the RNA Millennium Markers (Thermo Fisher Scientific). The ribosomal RNAs visualized by ethidium bromide staining (ETBR) were used as loading control. After transfer, the membrane was hybridized with ³²P-labeled oligo probes antisense to exon 3 of human SRSF3 (A) or exons 1–3 of mouse Srsf3 (B). See Table 1 for probe sequence details.

Fig. 2

Mapping of both human SRSF3 and mouse Srsf3 transcription start site (TSS). (A–C) Mapping of human SRSF3 TSS. A 5′ RACE was conducted with a human SRSF3-specific primer, oVM238 (located on SRSF3 exon 3), on total RNA isolated separately from three human cell lines, HeLa, CaSki, and SiHa (A). One major band from the 5ʹ RACE products in red rectangle (A) was gel purified and confirmed by Sanger sequencing (B). Human SRSF3 TSS was mapped to the hg38 chr6 nt 36594362 (in red) (B), which is 26 ​nt downstream of the TATA box (C). The entire sequence of the 5′ RACE region of human SRSF3 is shown along with a conserved TBP-binding site TATA box upstream of the mapped TSS position, the exon/exon junction positions, and the translation initiation codon ATG (red) (C). (D–F) Mapping of mouse Srsf3 TSS. A 5ʹ RACE was applied to mapping of mouse Srsf3 TSS on total RNA from mouse epithelial keratinocyte (MEK cells) (D) with a mouse Srsf3-specific primer oMA28 (located on exon 2) (F). Two major RACE products were gel purified (D) and sequenced by the Sanger sequencing (E). The products in the upper band (∗) turned to be a non-specific amplification. The lower band in red rectangle (D) is a specific 5ʹ RACE product with the mouse TSS (in red) being mapped at the mm10 chr17 nt 29032673 (E). The entire sequence of the 5ʹ RACE region of mouse Srsf3 is shown along with a conserved TBP-binding site TATA box 27 ​nt upstream of the mapped TSS position, exon 1/exon 2 junction positions, and the translation initiation codon ATG (red) (F). (G) The alignment showing the ∼70% homology between mapped SRSF3 and Srsf3 5ʹ UTRs.

Fig. 3

Mapping of human SRSF3 and mouse Srsf3 RNA polyadenylation cleavage sites. (A) Coverage and distribution of human SRSF3 RNA-seq reads from five different human cells along with human reference genome (hg38) by IGV visualization. An arrow to the right underneath the SRSF3 RefSeq is an oligo primer oJR56 from the terminal exon used for 3′ RACE in the panel B. (B) Mapping of human SRSF3 polyadenylation cleavage site by 3′ RACE and sequencing. A 3ʹ RACE was conducted with a human SRSF3-specific primer, oJR56, from the terminal exon (A), on total RNA isolated from three human cancer cell lines, HeLa, CaSki, SiHa and a human cervical keratinocyte cell line W12. Two major bands from 3ʹ RACE products in red rectangle were gel purified (left) and sequenced by the Sanger sequencing (right). The sequence reading on the right shows the boxed polyadenylation signal (PAS) and the mapped PA cleavage sites (CS) of human SRSF3 with the hg38 genome position labeled on the top. (C) Coverage and distribution of mouse Srsf3 RNA-seq reads from mouse skin tissues along with the mouse reference genome mm10. (D) Mapping of mouse Srsf3 polyadenylation cleavage site by 3′ RACE and sequencing. A 3ʹ RACE was performed using a mouse Srsf3-specific primer, oLLY531, from the terminal exon (arrow to the right in the panel C), on total RNA from mouse primary keratinocyte MEK cells. Similar to human cell 3ʹ RACE products, two major bands in red rectangle, a bigger band on the left panel with short exposure and a smaller band on the right panel with longer exposure, were gel purified and sequenced by the Sanger sequencing. The sequence reading in the middle panel shows the boxed PAS and mapped PA cleavage sites of mouse Srsf3 RNA with the labeled mm10 genome positions. (E) Identification of the core cis-elements upstream and downstream of the mapped SRSF3/Srsf3 polyadenylation signal and cleavage sites. Both human SRSF3 and mouse Srsf3 utilize AUUAAA as a major PAS for its RNA cleavage and polyadenylation, whereas a highly conserved AAUAAA hexamer pA signal further upstream only serves as a minor PAS for its RNA cleavage and polyadenylation. Both mapped cleavage sites are positioned 10–30 nt (highlighted in blue) downstream of PAS. The UGUA upstream of PAS (cyan) and G/U- or U-rich downstream sequence element (DSE) (green) located ∼10–30 nt downstream of the cleavage sites are also shown.

Fig. 4

Potential regulation of SRSF3/Srsf3 expression by 3ʹ UTR-targeting miRNAs. (A) The conserved miRNA binding sites (seed matches) based on TargetScan in the RefSeq annotated (grey) human SRSF3 3ʹ UTR in size of 3611 nt and mouse Srsf3 3ʹ UTR in size of 1961 nt or the mapped 3ʹ UTRs of human SRSF3 (red-794 nt long UTR, pink-304 nt short UTR) and mouse Srsf3 (blue −684 nt long UTR, light blue −306 nt short UTR). Vertical black bars above the human SRSF3 3ʹ UTR or below the mouse Srsf3 3ʹ UTR indicate the miRNA-binding sites predicted by TargetScan program (https://www.targetscan.org/vert_80/). (B) Alignment of the mapped long 3ʹ UTR of human SRSF3 and mouse Srsf3 in this study and their potential miRNA binding sites predicted by TargetScan. Human hsa-miRNA binding site are labeled in red color and mouse mmu-miRNA binding sites are in blue color.

Fig. 5

Revised mRNA structures of human SRSF3 and mouse Srsf3. (A) Revised gene structure of human SRSF3 and mouse Srsf3 from this report including genomic positions in the reference human (hg38) and mouse (mm10) genomes. (B and C) Revised RNA isoforms and their individual NCBI accession numbers (OP numbers) of human SRSF3 and mouse Srsf3. Both human SRSF3 and mouse Srsf3 contains 7 exons and 6 introns. Exon 4 is an alternative exon and contains a pre-mature termination (PMT) codon. Inclusion of exon 4 in spliced SRSF3 triggers nonsense-mediated decay of the exon 4-containing SRSF3 mRNA. Due to alternative exon 4 splicing and alternative PAS usage, four RNA isoforms of human or mouse SRSF3 will be produced, respectively, either from the human or mouse genome. The full-length SRSF3 mRNA without exon 4 is predominant and encodes SRSF3 protein of 164 amino acid residues (ORF in green), whereas the exon 4-bearing RNA, if not degraded, encodes a truncated protein of 124 aa residues (ORF in red). (D) The alignment of long (164 aa) and short (124 aa) SRSF3 or Srsf3 protein.